-
DiPT: Enhancing LLM reasoning through diversified perspective-taking
Authors:
Hoang Anh Just,
Mahavir Dabas,
Lifu Huang,
Ming Jin,
Ruoxi Jia
Abstract:
Existing work on improving language model reasoning typically explores a single solution path, which can be prone to errors. Inspired by perspective-taking in social studies, this paper introduces DiPT, a novel approach that complements current reasoning methods by explicitly incorporating diversified viewpoints. This approach allows the model to gain a deeper understanding of the problem's contex…
▽ More
Existing work on improving language model reasoning typically explores a single solution path, which can be prone to errors. Inspired by perspective-taking in social studies, this paper introduces DiPT, a novel approach that complements current reasoning methods by explicitly incorporating diversified viewpoints. This approach allows the model to gain a deeper understanding of the problem's context and identify the most effective solution path during the inference stage. Additionally, it provides a general data-centric AI recipe for augmenting existing data to improve their quality for fine-tuning.
Our empirical results demonstrate that DiPT can be flexibly integrated into existing methods that focus on a single reasoning approach, enhancing their reasoning performance and stability when presented with paraphrased problems. Furthermore, we illustrate improved context understanding by maintaining the model's safe outputs against "jailbreaking" prompts intentionally designed to bypass safeguards built into deployed models. Lastly, we show that fine-tuning with data enriched with diverse perspectives can boost the reasoning capabilities of the model compared to fine-tuning with raw data alone.
△ Less
Submitted 10 September, 2024;
originally announced September 2024.
-
Data-Centric Human Preference Optimization with Rationales
Authors:
Hoang Anh Just,
Ming Jin,
Anit Sahu,
Huy Phan,
Ruoxi Jia
Abstract:
Reinforcement learning from human feedback plays a crucial role in aligning language models towards human preferences, traditionally represented through comparisons between pairs or sets of responses within a given context. While many studies have enhanced algorithmic techniques to optimize learning from such data, this work shifts focus to improving preference learning through a data-centric appr…
▽ More
Reinforcement learning from human feedback plays a crucial role in aligning language models towards human preferences, traditionally represented through comparisons between pairs or sets of responses within a given context. While many studies have enhanced algorithmic techniques to optimize learning from such data, this work shifts focus to improving preference learning through a data-centric approach. Specifically, we propose enriching existing preference datasets with machine-generated rationales that explain the reasons behind choices. We develop a simple and principled framework to augment current preference learning methods with rationale information. Our comprehensive analysis highlights how rationales enhance learning efficiency. Extensive experiments reveal that rationale-enriched preference learning offers multiple advantages: it improves data efficiency, accelerates convergence to higher-performing models, and reduces verbosity bias and hallucination. Furthermore, this framework is versatile enough to integrate with various preference optimization algorithms. Overall, our findings highlight the potential of re-imagining data design for preference learning, demonstrating that even freely available machine-generated rationales can significantly boost performance across multiple dimensions. The code repository is available at https: //github.com/reds-lab/preference-learning-with-rationales
△ Less
Submitted 3 August, 2024; v1 submitted 19 July, 2024;
originally announced July 2024.
-
Get more for less: Principled Data Selection for Warming Up Fine-Tuning in LLMs
Authors:
Feiyang Kang,
Hoang Anh Just,
Yifan Sun,
Himanshu Jahagirdar,
Yuanzhi Zhang,
Rongxing Du,
Anit Kumar Sahu,
Ruoxi Jia
Abstract:
This work focuses on leveraging and selecting from vast, unlabeled, open data to pre-fine-tune a pre-trained language model. The goal is to minimize the need for costly domain-specific data for subsequent fine-tuning while achieving desired performance levels. While many data selection algorithms have been designed for small-scale applications, rendering them unsuitable for our context, some emerg…
▽ More
This work focuses on leveraging and selecting from vast, unlabeled, open data to pre-fine-tune a pre-trained language model. The goal is to minimize the need for costly domain-specific data for subsequent fine-tuning while achieving desired performance levels. While many data selection algorithms have been designed for small-scale applications, rendering them unsuitable for our context, some emerging methods do cater to language data scales. However, they often prioritize data that aligns with the target distribution. While this strategy may be effective when training a model from scratch, it can yield limited results when the model has already been pre-trained on a different distribution. Differing from prior work, our key idea is to select data that nudges the pre-training distribution closer to the target distribution. We show the optimality of this approach for fine-tuning tasks under certain conditions. We demonstrate the efficacy of our methodology across a diverse array of tasks (NLU, NLG, zero-shot) with models up to 2.7B, showing that it consistently surpasses other selection methods. Moreover, our proposed method is significantly faster than existing techniques, scaling to millions of samples within a single GPU hour. Our code is open-sourced (Code repository: https://anonymous.4open.science/r/DV4LLM-D761/ ). While fine-tuning offers significant potential for enhancing performance across diverse tasks, its associated costs often limit its widespread adoption; with this work, we hope to lay the groundwork for cost-effective fine-tuning, making its benefits more accessible.
△ Less
Submitted 4 May, 2024;
originally announced May 2024.
-
Star-by-star dynamical evolution of the physical pair of the Collinder 135 and UBC 7 open clusters
Authors:
Maryna Ishchenko,
Dana A. Kovaleva,
Peter Berczik,
Nina V. Kharchenko,
Anatoly E. Piskunov,
Evgeny Polyachenko,
Ekaterina Postnikova,
Andreas Just,
Olga Borodina,
Chingis Omarov,
Olexandr Sobodar
Abstract:
In a previous paper using Gaia DR2 data, we demonstrated that the two closely situated open clusters Collinder 135 and UBC 7 might have formed together about 50 Myr ago. In this work, we performed star-by-star dynamical modelling of the evolution of the open clusters Collinder 135 and UBC 7 from their supposed initial state to their present-day state, reproducing observational distributions of mem…
▽ More
In a previous paper using Gaia DR2 data, we demonstrated that the two closely situated open clusters Collinder 135 and UBC 7 might have formed together about 50 Myr ago. In this work, we performed star-by-star dynamical modelling of the evolution of the open clusters Collinder 135 and UBC 7 from their supposed initial state to their present-day state, reproducing observational distributions of members. Modelling of the Collinder 135 and UBC 7 dynamical evolution was done using the high-order parallel N-body code φ-GPU with up-to-date stellar evolution. Membership and characteristics of the clusters were acquired based on Gaia DR3 data. The comparison of the present-day radial cumulative star count obtained from the N-body simulations with the current observational data gave us full consistency of the model with observational data, especially in the central 8 pc, where 80% of the stars reside. The proper motion velocity components obtained from the N-body simulations of the stars are also quite consistent with the observed distributions and error bars. These results show that our numerical modelling is able to reproduce the open clusters' current complex 6D observed phase-space distributions with a high level of confidence. Thus, the model demonstrates that the hypothesis of a common origin of Collinder 135 and UBC 7 complies with present-day observational data.
△ Less
Submitted 15 May, 2024; v1 submitted 18 April, 2024;
originally announced April 2024.
-
Identification of new nearby white dwarfs using Gaia DR3
Authors:
Alex Golovin,
Sabine Reffert,
Akash Vani,
Ulrich Bastian,
Stefan Jordan,
Andreas Just
Abstract:
Based on the astrometry and photometry in Gaia DR3, we identified new nearby white dwarfs and validated those that had been missed from recent white dwarf catalogues despite being previously documented. To ensure the reliability of their astrometric solutions, we used a cut on just two parameters from Gaia DR3: the amplitude of the image parameter determination goodness-of-fit and the parallax-ove…
▽ More
Based on the astrometry and photometry in Gaia DR3, we identified new nearby white dwarfs and validated those that had been missed from recent white dwarf catalogues despite being previously documented. To ensure the reliability of their astrometric solutions, we used a cut on just two parameters from Gaia DR3: the amplitude of the image parameter determination goodness-of-fit and the parallax-over-error ratio. In addition, we imposed photometric signal-to-noise requirements to ensure the reliable identification of white dwarfs when using the colour-magnitude diagram.
We have identified nine previously unreported white dwarfs within the local population of 50 pc, and validated 21 previously reported white dwarfs missing from the GCWD21 (Gentile Fusillo et al. 2021) and other recent volume-limited white dwarf samples. A few of these objects belong to the rare class of ultra-cool white dwarfs. Four white dwarfs in our sample have an effective temperature of $T_{eff}\leq4000$ K within the $1σ$ interval, and two of them have an absolute magnitude of $M_G > 16.0$ mag. The identified white dwarfs are predominantly located in crowded fields, such as near the Galactic plane or in the foreground of the Large Magellanic Cloud. We also find that 19 of these white dwarfs have common proper motion companions with angular separations ranging from $1.1''$ to $7.1''$ and brightness differences between the components of up to 9.8 magnitudes. One of these systems is a triple system consisting of a white dwarf and two K dwarfs, while another is a double white dwarf system. We have identified 103 contaminants among the 2338 high-confidence white dwarfs in the 50 pc subsample of the GCWD21 and have found that their astrometric solutions in Gaia DR3 are spurious, improving the purity by 4.4%.
△ Less
Submitted 3 January, 2024; v1 submitted 18 December, 2023;
originally announced December 2023.
-
Self-Supervised Pre-Training Boosts Semantic Scene Segmentation on LiDAR Data
Authors:
Mariona Carós,
Ariadna Just,
Santi Seguí,
Jordi Vitrià
Abstract:
Airborne LiDAR systems have the capability to capture the Earth's surface by generating extensive point cloud data comprised of points mainly defined by 3D coordinates. However, labeling such points for supervised learning tasks is time-consuming. As a result, there is a need to investigate techniques that can learn from unlabeled data to significantly reduce the number of annotated samples. In th…
▽ More
Airborne LiDAR systems have the capability to capture the Earth's surface by generating extensive point cloud data comprised of points mainly defined by 3D coordinates. However, labeling such points for supervised learning tasks is time-consuming. As a result, there is a need to investigate techniques that can learn from unlabeled data to significantly reduce the number of annotated samples. In this work, we propose to train a self-supervised encoder with Barlow Twins and use it as a pre-trained network in the task of semantic scene segmentation. The experimental results demonstrate that our unsupervised pre-training boosts performance once fine-tuned on the supervised task, especially for under-represented categories.
△ Less
Submitted 22 December, 2023; v1 submitted 5 September, 2023;
originally announced September 2023.
-
Dynamics of supermassive black hole triples in the ROMULUS25 cosmological simulation
Authors:
Hauke Koehn,
Andreas Just,
Peter Berczik,
Michael Tremmel
Abstract:
For a pair of supermassive black holes (SMBHs) in the remnant of a dual galaxy merger, well-known models exist to describe their dynamical evolution until the final coalescence accompanied by the emission of a low-frequency gravitational wave (GW) signal. In this article, we investigate the dynamical evolution of three SMBH triple systems recovered from the ROMULUS25 cosmological simulation to exp…
▽ More
For a pair of supermassive black holes (SMBHs) in the remnant of a dual galaxy merger, well-known models exist to describe their dynamical evolution until the final coalescence accompanied by the emission of a low-frequency gravitational wave (GW) signal. In this article, we investigate the dynamical evolution of three SMBH triple systems recovered from the ROMULUS25 cosmological simulation to explore common dynamical evolution patterns and assess typical coalescence times. For this purpose, we construct initial conditions from the ROMULUS25 data and perform high-resolution gravitodynamical \N-body simulations. We track the orbital evolution from the galactic inspiral to the formation of hard binaries at sub-parsec separation and use the observed hardening rates to project the time of coalescence. In all cases, the two heaviest black holes form an efficiently hardening binary that merges within fractions of the Hubble time. The lightest SMBH either gets ejected, forms a stable hierarchical triple system with the heavier binary, forms a hardening binary with the previously merged binary's remnant, or remains on a wide galactic orbit. The coalescence times of the lighter black holes are thus significantly longer than for the heavier binary, as they experience lower dynamical friction and stellar hardening rates. We observe the formation of hierarchical triples when the density profile of the galactic nucleus is sufficiently steep.
△ Less
Submitted 21 August, 2023;
originally announced August 2023.
-
Performance Scaling via Optimal Transport: Enabling Data Selection from Partially Revealed Sources
Authors:
Feiyang Kang,
Hoang Anh Just,
Anit Kumar Sahu,
Ruoxi Jia
Abstract:
Traditionally, data selection has been studied in settings where all samples from prospective sources are fully revealed to a machine learning developer. However, in practical data exchange scenarios, data providers often reveal only a limited subset of samples before an acquisition decision is made. Recently, there have been efforts to fit scaling laws that predict model performance at any size a…
▽ More
Traditionally, data selection has been studied in settings where all samples from prospective sources are fully revealed to a machine learning developer. However, in practical data exchange scenarios, data providers often reveal only a limited subset of samples before an acquisition decision is made. Recently, there have been efforts to fit scaling laws that predict model performance at any size and data source composition using the limited available samples. However, these scaling functions are black-box, computationally expensive to fit, highly susceptible to overfitting, or/and difficult to optimize for data selection. This paper proposes a framework called <projektor>, which predicts model performance and supports data selection decisions based on partial samples of prospective data sources. Our approach distinguishes itself from existing work by introducing a novel *two-stage* performance inference process. In the first stage, we leverage the Optimal Transport distance to predict the model's performance for any data mixture ratio within the range of disclosed data sizes. In the second stage, we extrapolate the performance to larger undisclosed data sizes based on a novel parameter-free mapping technique inspired by neural scaling laws. We further derive an efficient gradient-based method to select data sources based on the projected model performance. Evaluation over a diverse range of applications demonstrates that <projektor> significantly improves existing performance scaling approaches in terms of both the accuracy of performance inference and the computation costs associated with constructing the performance predictor. Also, <projektor> outperforms by a wide margin in data selection effectiveness compared to a range of other off-the-shelf solutions.
△ Less
Submitted 5 July, 2023;
originally announced July 2023.
-
2D-Shapley: A Framework for Fragmented Data Valuation
Authors:
Zhihong Liu,
Hoang Anh Just,
Xiangyu Chang,
Xi Chen,
Ruoxi Jia
Abstract:
Data valuation -- quantifying the contribution of individual data sources to certain predictive behaviors of a model -- is of great importance to enhancing the transparency of machine learning and designing incentive systems for data sharing. Existing work has focused on evaluating data sources with the shared feature or sample space. How to valuate fragmented data sources of which each only conta…
▽ More
Data valuation -- quantifying the contribution of individual data sources to certain predictive behaviors of a model -- is of great importance to enhancing the transparency of machine learning and designing incentive systems for data sharing. Existing work has focused on evaluating data sources with the shared feature or sample space. How to valuate fragmented data sources of which each only contains partial features and samples remains an open question. We start by presenting a method to calculate the counterfactual of removing a fragment from the aggregated data matrix. Based on the counterfactual calculation, we further propose 2D-Shapley, a theoretical framework for fragmented data valuation that uniquely satisfies some appealing axioms in the fragmented data context. 2D-Shapley empowers a range of new use cases, such as selecting useful data fragments, providing interpretation for sample-wise data values, and fine-grained data issue diagnosis.
△ Less
Submitted 26 July, 2023; v1 submitted 18 June, 2023;
originally announced June 2023.
-
LAVA: Data Valuation without Pre-Specified Learning Algorithms
Authors:
Hoang Anh Just,
Feiyang Kang,
Jiachen T. Wang,
Yi Zeng,
Myeongseob Ko,
Ming Jin,
Ruoxi Jia
Abstract:
Traditionally, data valuation (DV) is posed as a problem of equitably splitting the validation performance of a learning algorithm among the training data. As a result, the calculated data values depend on many design choices of the underlying learning algorithm. However, this dependence is undesirable for many DV use cases, such as setting priorities over different data sources in a data acquisit…
▽ More
Traditionally, data valuation (DV) is posed as a problem of equitably splitting the validation performance of a learning algorithm among the training data. As a result, the calculated data values depend on many design choices of the underlying learning algorithm. However, this dependence is undesirable for many DV use cases, such as setting priorities over different data sources in a data acquisition process and informing pricing mechanisms in a data marketplace. In these scenarios, data needs to be valued before the actual analysis and the choice of the learning algorithm is still undetermined then. Another side-effect of the dependence is that to assess the value of individual points, one needs to re-run the learning algorithm with and without a point, which incurs a large computation burden. This work leapfrogs over the current limits of data valuation methods by introducing a new framework that can value training data in a way that is oblivious to the downstream learning algorithm. Our main results are as follows. (1) We develop a proxy for the validation performance associated with a training set based on a non-conventional class-wise Wasserstein distance between training and validation sets. We show that the distance characterizes the upper bound of the validation performance for any given model under certain Lipschitz conditions. (2) We develop a novel method to value individual data based on the sensitivity analysis of the class-wise Wasserstein distance. Importantly, these values can be directly obtained for free from the output of off-the-shelf optimization solvers when computing the distance. (3) We evaluate our new data valuation framework over various use cases related to detecting low-quality data and show that, surprisingly, the learning-agnostic feature of our framework enables a significant improvement over SOTA performance while being orders of magnitude faster.
△ Less
Submitted 19 December, 2023; v1 submitted 28 April, 2023;
originally announced May 2023.
-
Evolution of open clusters with or without black holes
Authors:
B. Shukirgaliyev,
P. Berczik,
A. Otebay,
M. Kalambay,
A. Kamlah,
Y. Tleukhanov,
E. Abdikamalov,
S. Banerjee,
A. Just
Abstract:
Binary black holes (BHs) can be formed dynamically in the centers of star clusters. The high natal kicks for stellar-mass BHs used in previous works made it hard to retain BHs in star clusters. Recent studies of massive star evolution and supernovae (SN) propose kick velocities that are lower due to the fallback of the SN ejecta. We study the impact of these updates by performing $N$-body simulati…
▽ More
Binary black holes (BHs) can be formed dynamically in the centers of star clusters. The high natal kicks for stellar-mass BHs used in previous works made it hard to retain BHs in star clusters. Recent studies of massive star evolution and supernovae (SN) propose kick velocities that are lower due to the fallback of the SN ejecta. We study the impact of these updates by performing $N$-body simulations following instantaneous gas expulsion. For comparison, we simulate two additional model sets with the previous treatment of stars: one with high kicks and another with artificial removal of the kicks. Our model clusters initially consist of about one hundred thousand stars, formed with centrally-peaked efficiency. We find that the updated treatment of stars, due to the fallback-scaled lower natal kicks, allows clusters to retain SN remnants after violent relaxation. The mass contribution of the retained remnants does not exceed a few percent of the total bound cluster mass during the early evolution. For this reason, the first giga year of evolution is not affected significantly by this effect. Nevertheless, during the subsequent long-term evolution, the retained BHs accelerate mass segregation, leading to the faster dissolution of the clusters.
△ Less
Submitted 24 April, 2023;
originally announced April 2023.
-
Global survey of star clusters in the Milky Way -- VII. Tidal parameters and mass function
Authors:
Andreas Just,
Anatoly E. Piskunov,
Jonathan H. Klos,
Dana A. Kovaleva,
Evgeny V. Polyachenko
Abstract:
We built Galactic open star cluster mass functions (CMFs) for different age sub-samples in the wider solar neighbourhood. We present a simple cluster formation and evolution model to reproduce the main features of the CMFs. We used an unbiased sample of 2227 clusters of the Milky Way Star Cluster (MWSC) catalogue, which occupy the heliocentric cylinders with magnitude-dependent completeness radii…
▽ More
We built Galactic open star cluster mass functions (CMFs) for different age sub-samples in the wider solar neighbourhood. We present a simple cluster formation and evolution model to reproduce the main features of the CMFs. We used an unbiased sample of 2227 clusters of the Milky Way Star Cluster (MWSC) catalogue, which occupy the heliocentric cylinders with magnitude-dependent completeness radii of 1-5 kpc. We derived tidal masses of clusters with an accuracy of 70%. Our cluster formation and evolution model is based on the cluster initial mass function, the cluster formation rate, cluster mass loss due to stellar evolution and the clusters' dynamical evolution in the Galactic tidal field. The obtained tidal masses have been added to the MWSC catalogue. A general CMF (GCMF), built for all cluster ages around the Sun, extends over four decades in mass. The high-mass slope is +1.14. The CMFs for different age groups show the same high-mass slopes, while the low-mass slope is flat for the youngest sub-sample and about -0.7 for the others. The sub-samples inside and outside the solar Galactocentric radius are consistent with the GCMF, once the exponential decline of the Galactic disc density is taken into account. The model suggests star formation with low efficiency of 15%, where 10% of stars remain bound in a cluster after gas expulsion and violent relaxation. The cluster formation rate required to reproduce the observed age-mass-distribution is 0.4 solar masses per square pc and Gyr. The obtained high-mass slope of the GCMF for the wide solar neighbourhood is similar to slopes determined in nearby galaxies. The MWSC catalogue supports models with low star-formation efficiency, where 90% of stars are lost quickly after gas expulsion. The cluster formation rate corresponds to open clusters' contribution to the stellar content of the thin disc of 30%.
△ Less
Submitted 21 February, 2023;
originally announced February 2023.
-
On Perfect Bases in Finite Abelian Groups
Authors:
Bela Bajnok,
Connor Berson,
Hoang Anh Just
Abstract:
Let $G$ be a finite abelian group and $s$ be a positive integer. A subset $A$ of $G$ is called a {\em perfect $s$-basis of $G$} if each element of $G$ can be written uniquely as the sum of at most $s$ (not-necessarily-distinct) elements of $A$; similarly, we say that $A$ is a {\em perfect restricted $s$-basis of $G$} if each element of $G$ can be written uniquely as the sum of at most $s$ distinct…
▽ More
Let $G$ be a finite abelian group and $s$ be a positive integer. A subset $A$ of $G$ is called a {\em perfect $s$-basis of $G$} if each element of $G$ can be written uniquely as the sum of at most $s$ (not-necessarily-distinct) elements of $A$; similarly, we say that $A$ is a {\em perfect restricted $s$-basis of $G$} if each element of $G$ can be written uniquely as the sum of at most $s$ distinct elements of $A$. We prove that perfect $s$-bases exist only in the trivial cases of $s=1$ or $|A|=1$. The situation is different with restricted addition where perfection is more frequent; here we treat the case of $s=2$ and prove that $G$ has a perfect restricted $2$-basis if, and only if, it is isomorphic to $\mathbb{Z}_2$, $\mathbb{Z}_4$, $\mathbb{Z}_7$, $\mathbb{Z}_2^2$, $\mathbb{Z}_2^4$, or $\mathbb{Z}_2^2 \times \mathbb{Z}_4$.
△ Less
Submitted 24 November, 2022;
originally announced November 2022.
-
The Fifth Catalogue of Nearby Stars (CNS5)
Authors:
Alex Golovin,
Sabine Reffert,
Andreas Just,
Stefan Jordan,
Akash Vani,
Hartmut Jahreiß
Abstract:
We present the compilation of the Fifth Catalogue of Nearby Stars (CNS5), based on astrometric and photometric data from Gaia EDR3 and Hipparcos, and supplemented with parallaxes from ground-based astrometric surveys carried out in the infrared.
The aim of the CNS5 is to provide the most complete sample of objects in the solar neighbourhood. For all known stars and brown dwarfs in the 25 pc sphe…
▽ More
We present the compilation of the Fifth Catalogue of Nearby Stars (CNS5), based on astrometric and photometric data from Gaia EDR3 and Hipparcos, and supplemented with parallaxes from ground-based astrometric surveys carried out in the infrared.
The aim of the CNS5 is to provide the most complete sample of objects in the solar neighbourhood. For all known stars and brown dwarfs in the 25 pc sphere around the Sun, basic astrometric and photometric parameters are given. Furthermore, we provide the colour-magnitude diagram and various luminosity functions of the stellar content in the solar neighbourhood, and characterise the completeness of the CNS5 catalogue.
We compile a sample of stars and brown dwarfs which most likely are located within 25 pc of the Sun, taking space-based parallaxes from Gaia EDR3 and Hipparcos as well as ground-based parallaxes from Best et al. (2021), Kirkpatrick et al. (2021), and from the CNS4 into account. We develop a set of selection criteria to clean the sample from spurious sources. Furthermore, we show that effects of blending in the Gaia photometry, which affect mainly the faint and red sources in Gaia, can be mitigated, to reliably place those objects in a colour-magnitude diagram. We also assess the completeness of the CNS5 using a Kolmogorov-Smirnov test and derive observational optical and mid-infrared luminosity functions for the main-sequence stars and white dwarfs in the solar neighbourhood. The CNS5 contains 5931 objects, including 5230 stars (4946 main-sequence stars, 20 red giants and 264 white dwarfs) and 701 brown dwarfs.
We find that the CNS5 catalogue is statistically complete down to 19.7 mag in G-band and 11.8 mag in W1-band absolute magnitudes, corresponding to a spectral type of L8.
△ Less
Submitted 2 November, 2022;
originally announced November 2022.
-
Object Segmentation of Cluttered Airborne LiDAR Point Clouds
Authors:
Mariona Caros,
Ariadna Just,
Santi Segui,
Jordi Vitria
Abstract:
Airborne topographic LiDAR is an active remote sensing technology that emits near-infrared light to map objects on the Earth's surface. Derived products of LiDAR are suitable to service a wide range of applications because of their rich three-dimensional spatial information and their capacity to obtain multiple returns. However, processing point cloud data still requires a significant effort in ma…
▽ More
Airborne topographic LiDAR is an active remote sensing technology that emits near-infrared light to map objects on the Earth's surface. Derived products of LiDAR are suitable to service a wide range of applications because of their rich three-dimensional spatial information and their capacity to obtain multiple returns. However, processing point cloud data still requires a significant effort in manual editing. Certain human-made objects are difficult to detect because of their variety of shapes, irregularly-distributed point clouds, and low number of class samples. In this work, we propose an efficient end-to-end deep learning framework to automatize the detection and segmentation of objects defined by an arbitrary number of LiDAR points surrounded by clutter. Our method is based on a light version of PointNet that achieves good performance on both object recognition and segmentation tasks. The results are tested against manually delineated power transmission towers and show promising accuracy.
△ Less
Submitted 6 February, 2023; v1 submitted 28 October, 2022;
originally announced October 2022.
-
Towards a fully consistent Milky Way disk model V. The disk model for 4 $-$ 14 kpc
Authors:
K. Sysoliatina,
A. Just
Abstract:
We present a generalised version of the semi-analytic Just-Jahreiss (JJ) model of the Galactic disk that incorporates our findings for the solar neighbourhood and is applicable to a wide range of galactocentric distances, 4 kpc $\lesssim R \lesssim $ 14 kpc. The JJ model is a flexible tool for stellar population synthesis with a fine age resolution of 25 Myr. It includes six Milky-Way (MW) compone…
▽ More
We present a generalised version of the semi-analytic Just-Jahreiss (JJ) model of the Galactic disk that incorporates our findings for the solar neighbourhood and is applicable to a wide range of galactocentric distances, 4 kpc $\lesssim R \lesssim $ 14 kpc. The JJ model is a flexible tool for stellar population synthesis with a fine age resolution of 25 Myr. It includes six Milky-Way (MW) components: the four flattened (exponential thin and thick disk, atomic and molecular gas) and two spheroidal (spherical stellar halo and a cored isothermal dark matter sphere). The overall thin-disk thickness is assumed to be constant at all radii, though flaring can also be tested. The adopted radial variation in the thin-disk star-formation rate (SFR) reflects the inside-out disk growth scenario. We allow a smooth power-law SFR continuum to be modified by an arbitrary number of Gaussian peaks. We present a public code of the JJ model complemented by the three sets of isochrones (PARSEC, MIST, and BaSTI). Using metallicity distributions of the red clump giants from APOGEE, we constrain the radial variation of the JJ-model age-metallicity relation (AMR) and propose a new analytic form for the AMR function. The generalised JJ model is a publicly available tool for studying different stellar populations across the MW disk. With its fine age resolution and flexibility, it can be particularly useful for reconstructing the thin-disk SFR, as a variety of different SFR shapes can be constructed within its framework.
△ Less
Submitted 26 August, 2022; v1 submitted 25 July, 2022;
originally announced July 2022.
-
Narcissus: A Practical Clean-Label Backdoor Attack with Limited Information
Authors:
Yi Zeng,
Minzhou Pan,
Hoang Anh Just,
Lingjuan Lyu,
Meikang Qiu,
Ruoxi Jia
Abstract:
Backdoor attacks insert malicious data into a training set so that, during inference time, it misclassifies inputs that have been patched with a backdoor trigger as the malware specified label. For backdoor attacks to bypass human inspection, it is essential that the injected data appear to be correctly labeled. The attacks with such property are often referred to as "clean-label attacks." Existin…
▽ More
Backdoor attacks insert malicious data into a training set so that, during inference time, it misclassifies inputs that have been patched with a backdoor trigger as the malware specified label. For backdoor attacks to bypass human inspection, it is essential that the injected data appear to be correctly labeled. The attacks with such property are often referred to as "clean-label attacks." Existing clean-label backdoor attacks require knowledge of the entire training set to be effective. Obtaining such knowledge is difficult or impossible because training data are often gathered from multiple sources (e.g., face images from different users). It remains a question whether backdoor attacks still present a real threat.
This paper provides an affirmative answer to this question by designing an algorithm to mount clean-label backdoor attacks based only on the knowledge of representative examples from the target class. With poisoning equal to or less than 0.5% of the target-class data and 0.05% of the training set, we can train a model to classify test examples from arbitrary classes into the target class when the examples are patched with a backdoor trigger. Our attack works well across datasets and models, even when the trigger presents in the physical world.
We explore the space of defenses and find that, surprisingly, our attack can evade the latest state-of-the-art defenses in their vanilla form, or after a simple twist, we can adapt to the downstream defenses. We study the cause of the intriguing effectiveness and find that because the trigger synthesized by our attack contains features as persistent as the original semantic features of the target class, any attempt to remove such triggers would inevitably hurt the model accuracy first.
△ Less
Submitted 15 April, 2022; v1 submitted 11 April, 2022;
originally announced April 2022.
-
Label-Only Model Inversion Attacks via Boundary Repulsion
Authors:
Mostafa Kahla,
Si Chen,
Hoang Anh Just,
Ruoxi Jia
Abstract:
Recent studies show that the state-of-the-art deep neural networks are vulnerable to model inversion attacks, in which access to a model is abused to reconstruct private training data of any given target class. Existing attacks rely on having access to either the complete target model (whitebox) or the model's soft-labels (blackbox). However, no prior work has been done in the harder but more prac…
▽ More
Recent studies show that the state-of-the-art deep neural networks are vulnerable to model inversion attacks, in which access to a model is abused to reconstruct private training data of any given target class. Existing attacks rely on having access to either the complete target model (whitebox) or the model's soft-labels (blackbox). However, no prior work has been done in the harder but more practical scenario, in which the attacker only has access to the model's predicted label, without a confidence measure. In this paper, we introduce an algorithm, Boundary-Repelling Model Inversion (BREP-MI), to invert private training data using only the target model's predicted labels. The key idea of our algorithm is to evaluate the model's predicted labels over a sphere and then estimate the direction to reach the target class's centroid. Using the example of face recognition, we show that the images reconstructed by BREP-MI successfully reproduce the semantics of the private training data for various datasets and target model architectures. We compare BREP-MI with the state-of-the-art whitebox and blackbox model inversion attacks and the results show that despite assuming less knowledge about the target model, BREP-MI outperforms the blackbox attack and achieves comparable results to the whitebox attack.
△ Less
Submitted 3 March, 2022;
originally announced March 2022.
-
ModelPred: A Framework for Predicting Trained Model from Training Data
Authors:
Yingyan Zeng,
Jiachen T. Wang,
Si Chen,
Hoang Anh Just,
Ran Jin,
Ruoxi Jia
Abstract:
In this work, we propose ModelPred, a framework that helps to understand the impact of changes in training data on a trained model. This is critical for building trust in various stages of a machine learning pipeline: from cleaning poor-quality samples and tracking important ones to be collected during data preparation, to calibrating uncertainty of model prediction, to interpreting why certain be…
▽ More
In this work, we propose ModelPred, a framework that helps to understand the impact of changes in training data on a trained model. This is critical for building trust in various stages of a machine learning pipeline: from cleaning poor-quality samples and tracking important ones to be collected during data preparation, to calibrating uncertainty of model prediction, to interpreting why certain behaviors of a model emerge during deployment. Specifically, ModelPred learns a parameterized function that takes a dataset $S$ as the input and predicts the model obtained by training on $S$. Our work differs from the recent work of Datamodels [1] as we aim for predicting the trained model parameters directly instead of the trained model behaviors. We demonstrate that a neural network-based set function class is capable of learning the complex relationships between the training data and model parameters. We introduce novel global and local regularization techniques to prevent overfitting and we rigorously characterize the expressive power of neural networks (NN) in approximating the end-to-end training process. Through extensive empirical investigations, we show that ModelPred enables a variety of applications that boost the interpretability and accountability of machine learning (ML), such as data valuation, data selection, memorization quantification, and model calibration.
△ Less
Submitted 23 December, 2022; v1 submitted 24 November, 2021;
originally announced November 2021.
-
The Neuroscience of Advanced Scientific Concepts
Authors:
Robert A. Mason,
Reinhard A. Schumacher,
Marcel A. Just
Abstract:
Cognitive neuroscience methods can identify the fMRI-measured neural representation of familiar individual concepts, such as apple, and decompose them into meaningful neural and semantic components. This approach was applied here to determine the neural representations and underlying dimensions of representation of far more abstract physics concepts related to matter and energy, such as fermion an…
▽ More
Cognitive neuroscience methods can identify the fMRI-measured neural representation of familiar individual concepts, such as apple, and decompose them into meaningful neural and semantic components. This approach was applied here to determine the neural representations and underlying dimensions of representation of far more abstract physics concepts related to matter and energy, such as fermion and dark matter, in the brains of 10 Carnegie Mellon physics faculty members who thought about the main properties of each of the concepts. One novel dimension coded the measurability vs. immeasurability of a concept. Another novel dimension of representation evoked particularly by post-classical concepts was associated with four types of cognitive processes, each linked to particular brain regions: (1) Reasoning about intangibles, taking into account their separation from direct experience and observability; (2) Assessing consilience with other, firmer knowledge; (3) Causal reasoning about relations that are not apparent or observable; and (4) Knowledge management of a large knowledge organization consisting of a multi-level structure of other concepts. Two other underlying dimensions, previously found in physics students, periodicity, and mathematical formulation, were also present in this faculty sample. The data were analyzed using factor analysis of stably responding voxels, a Gaussian-naïve Bayes machine-learning classification of the activation patterns associated with each concept, and a regression model that predicted activation patterns associated with each concept based on independent ratings of the dimensions of the concepts. The findings indicate that the human brain systematically organizes novel scientific concepts in terms of new dimensions of neural representation.
△ Less
Submitted 12 October, 2021;
originally announced October 2021.
-
The impact of Post-Newtonian effects on massive black hole binary evolution at $\sim 1000\,R_{sch}$ separations
Authors:
Branislav Avramov,
Peter Berczik,
Andreas Just
Abstract:
Aims: We study the impact of Post-Newtonian correction terms on the energetic interaction between a gravitational wave (GW)-emitting supermassive black hole (SMBH) binary system and incoming stars via three-body scattering experiments. Methods: We use the AR-chain code to simulate with high accuracy the interactions between stars and an SMBH binary at separations of$\sim 1000\,R_{sch}$. For all of…
▽ More
Aims: We study the impact of Post-Newtonian correction terms on the energetic interaction between a gravitational wave (GW)-emitting supermassive black hole (SMBH) binary system and incoming stars via three-body scattering experiments. Methods: We use the AR-chain code to simulate with high accuracy the interactions between stars and an SMBH binary at separations of$\sim 1000\,R_{sch}$. For all of the interactions, we investigate in detail the energy balance of the three-body systems, using both Newtonian and Post-Newtonian expressions for the SMBH binary orbital energy, taking into account the GW emission by the binary. Results: We find that at these separations, purely Newtonian treatment of the binary orbital energy is insufficient to properly account for the SMBH binary orbital evolution. Instead, along with GW emission, even terms in the PN-corrections must be included in order to describe the energy change of the binary during the stellar interaction.es, 2
△ Less
Submitted 17 September, 2021;
originally announced September 2021.
-
The bound mass of Dehnen models with centrally peaked star formation efficiency
Authors:
B. Shukirgaliyev,
A. Otebay,
M. Sobolenko,
M. Ishchenko,
O. Borodina,
T. Panamarev,
S. Myrzakul,
M. Kalambay,
A. Naurzbayeva,
E. Abdikamalov,
E. Polyachenko,
S. Banerjee,
P. Berczik,
R. Spurzem,
A. Just
Abstract:
Understanding the formation of star clusters with low star-formation efficiency (SFE) is very important to know about the star-formation history. In N-body models of star cluster evolution after gas expulsion, the Plummer model with outer power law density profile has been used massively. We study the impact of the density profile slopes on the survivability of the low-SFE star clusters after inst…
▽ More
Understanding the formation of star clusters with low star-formation efficiency (SFE) is very important to know about the star-formation history. In N-body models of star cluster evolution after gas expulsion, the Plummer model with outer power law density profile has been used massively. We study the impact of the density profile slopes on the survivability of the low-SFE star clusters after instantaneous gas expulsion. We compare cases when stellar cluster has Plummer profile and Dehnen profiles with cusp of different slopes at the time of formation. We determine the corresponding density profile of the residual gas for a given global SFE, assuming that our model clusters formed with a constant efficiency per free-fall time and hence have shallower density profile of gas than that of stars. We perform direct $N$-body simulations of evolution of clusters initially in virial equilibrium within gas potential after gas removal. We find that the violent relaxation lasts no longer than 20~Myr independently of the density profile power law slopes. Dehnen model clusters survive after violent relaxation with significantly lower SFEs when the global SFE measured within the Jacobi radius or within a half-mass radius. Dehnen $γ=0$ model clusters show similar final bound fraction with the Plummer model clusters if global SFE is measured within 10 scale radii. The final bound fraction increases with $γ$ values for a given global SFE. We conclude that Dehnen clusters better resist the consequences of the violent relaxation followed the instantaneous gas expulsion than the Plummer clusters. Thus the shallower the outer density slope of the low-SFE clusters, the better for their survivability after gas expulsion. Among Dehnen clusters we find that the steeper the inner slope (cusp) the higher the bound mass fraction is retained after violent relaxation for a given global SFE.
△ Less
Submitted 23 July, 2021; v1 submitted 20 May, 2021;
originally announced May 2021.
-
Towards a fully consistent Milky Way disk model -- IV. The impact of Gaia DR2 and APOGEE
Authors:
K. Sysoliatina,
A. Just
Abstract:
We present an updated version of the semi-analytic Just-Jahreiß (JJ) model of the Galactic disk and constrain its parameters in the Solar neighbourhood. The new features of the JJ model include a simple two-component gaseous disk, a star-formation rate (SFR) function of the thick disk that has been extended in time, and a correlation between the kinematics of molecular gas and thin-disk population…
▽ More
We present an updated version of the semi-analytic Just-Jahreiß (JJ) model of the Galactic disk and constrain its parameters in the Solar neighbourhood. The new features of the JJ model include a simple two-component gaseous disk, a star-formation rate (SFR) function of the thick disk that has been extended in time, and a correlation between the kinematics of molecular gas and thin-disk populations. Here, we study the vertical number density profiles and W-velocity distributions determined from ~2 million local stars of the Gaia DR2. We also investigate an apparent Hess diagram of the Gaia stars selected in a conic volume towards the Galactic poles. Using a stellar evolution library, we synthesise stellar populations with a four-slope broken power-law initial mass function (IMF), the SFR, and an age-metallicity relation (AMR). The latter is consistently derived with the observed metallicity distribution of the local Red Clump (RC) giants from the APOGEE. Working within a Bayesian approach, we sample the posterior probability distribution in a multidimensional parameter space using the MCMC method. We find that the spatial distribution and motion of the Gaia stars imply two recent SF bursts centered at ages of ~0.5 Gyr and ~3 Gyr and characterised by a ~30% and ~55% SF enhancement, respectively, relative to a monotonously declining SFR continuum. The stellar populations associated with this SF excess are found to be dynamically hot for their age: they have W-velocity dispersions of ~12.5 km/s and ~26 km/s. The new JJ model is able to reproduce the local star counts with an accuracy of ~5 %. Using Gaia DR2 data, we self-consistently constrained 22 parameters of the updated JJ model. Our optimised model predicts two SF bursts within the last ~4 Gyr, which may point to recent episodes of gas infall.
△ Less
Submitted 18 February, 2021;
originally announced February 2021.
-
Properties of loss cone stars in a cosmological galaxy merger remnant
Authors:
Branislav Avramov,
Peter Berczik,
Yohai Meiron,
Anshuman Acharya,
Andreas Just
Abstract:
Aims: We investigate the orbital and phase space properties of loss cone stars that interact strongly with a hard, high-redshift binary supermassive black hole (SMBH) system formed in a cosmological scenario. Methods: We use a novel hybrid integration approach that combines the direct N-body code $\varphi$-GRAPE with ETICS, a collisionless code that employs the self-consistent field method for for…
▽ More
Aims: We investigate the orbital and phase space properties of loss cone stars that interact strongly with a hard, high-redshift binary supermassive black hole (SMBH) system formed in a cosmological scenario. Methods: We use a novel hybrid integration approach that combines the direct N-body code $\varphi$-GRAPE with ETICS, a collisionless code that employs the self-consistent field method for force calculation. The hybrid approach shows considerable speed-up over direct summation for particle numbers $> 10^6$, while retaining accuracy of direct N-body for a subset of particles. During the SMBH binary evolution we monitor individual stellar interactions with the binary in order to identify stars that noticeably contribute to the SMBH binary hardening. Results: We successfully identify and analyze in detail the properties of stars that extract energy from the binary. We find that the summed energy changes seen in these stars match very well with the overall binary energy change, demonstrating that stellar interactions are the primary drivers of SMBH binary hardening in triaxial, gas-poor systems. The slight triaxiality of our system results in efficient loss cone refilling, avoiding the final parsec problem. We distinguish three different populations of interactions based on their apocenter. We find a clear prevalence of interactions co-rotating with the binary. Nevertheless, retrograde interactions are the most energetic, contributing only slightly less than the prograde population to the overall energy exchange. The most energetic interactions are also likely to result in a change of sign in the angular momentum of the star. We estimate the merger timescale of the binary to be $\approx 20$ $\mathrm{Myr}$, a value larger by a factor of two than the timescale reported in a previous study.
△ Less
Submitted 2 March, 2021; v1 submitted 16 November, 2020;
originally announced November 2020.
-
Collinder 135 and UBC 7: A Physical Pair of Open Clusters
Authors:
Dana Kovaleva,
Marina Ishchenko,
Ekaterina Postnikova,
Peter Berczik,
Anatoly E. Piskunov,
Nina V. Kharchenko,
Evgeny Polyachenko,
Sabine Reffert,
Kseniia Sysoliatina,
Andreas Just
Abstract:
Given the closeness of the two open clusters Cr 135 and UBC 7 on the sky, we investigate the possibility of the two clusters to be physically related. We aim to recover the present-day stellar membership in the open clusters Collinder 135 and UBC 7 (300 pc from the Sun), to constrain their kinematic parameters, ages and masses, and to restore their primordial phase space configuration. The most re…
▽ More
Given the closeness of the two open clusters Cr 135 and UBC 7 on the sky, we investigate the possibility of the two clusters to be physically related. We aim to recover the present-day stellar membership in the open clusters Collinder 135 and UBC 7 (300 pc from the Sun), to constrain their kinematic parameters, ages and masses, and to restore their primordial phase space configuration. The most reliable cluster members are selected with our traditional method modified for the use of Gaia DR2 data. Numerical simulations use the integration of cluster trajectories backwards in time with our original high order Hermite4 code \PGRAPE. We constrain the age, spatial coordinates and velocities, radii and masses of the clusters. We estimate the actual separation of the cluster centres equal to 24 pc. The orbital integration shows that the clusters were much closer in the past if their current line-of-sight velocities are very similar and the total mass is more than 7 times larger the mass of the determined most reliable members. We conclude that the two clusters Cr 135 and UBC 7 might very well have formed a physial pair, based on the observational evidence as well as numerical simulations. The probability of a chance coincidence is only about $2\%$.
△ Less
Submitted 4 September, 2020;
originally announced September 2020.
-
Addressing via N-body simulations the distribution of the satellite tidal debris in the Milky Way environment
Authors:
Matteo Mazzarini,
Andreas Just,
Andrea V. Macciò,
Reza Moetazedian
Abstract:
We study the distribution of the Milky Way satellites stellar and dark matter debris. For the first time we address the question of the tidal disruption of satellites in simulations by utilising simultaneously a) a realistic set of orbits extracted from cosmological simulations, b) a three component host galaxy with live halo, disc and bulge components, and c) satellites from hydrodynamical simula…
▽ More
We study the distribution of the Milky Way satellites stellar and dark matter debris. For the first time we address the question of the tidal disruption of satellites in simulations by utilising simultaneously a) a realistic set of orbits extracted from cosmological simulations, b) a three component host galaxy with live halo, disc and bulge components, and c) satellites from hydrodynamical simulations. We analyse the statistical properties of the satellite debris of all massive galaxies reaching the inner Milky Way on a timescale of 2 Gyr. Up to 80$\%$ of the dark matter is stripped from the satellites, while this happens for up to 30$\%$ of their stars. The stellar debris ends mostly in the inner Milky Way halo, whereas the dark matter debris shows a flat mass distribution over the full main halo. The dark matter debris follows a density profile with inner power law index $α_{\rm DM}=-0.66$ and outer index $β_{\rm DM}=2.94$, while for stars $α_{*}=-0.44$ and $β_{*}=6.17$. In the inner 25 kpc, the distribution of the stellar debris is flatter than that of the dark matter debris and the orientations of their short axes differ significantly. Changing the orientation of the stellar disc by 90$^{\rm{o}}$ has only a minor impact on the distribution of the satellite debris. Our results indicate that the dark matter is more easily stripped than stars from the Milky Way satellites. The structure of the debris is dominated by the satellite orbital properties. The radial profiles, the flattening and the orientation of the stellar and dark matter debris are significantly different, which prevents the prediction of the dark matter distribution from the observed stellar component.
△ Less
Submitted 20 March, 2020;
originally announced March 2020.
-
The Sixth Data Release of the Radial Velocity Experiment (RAVE) -- II: Stellar Atmospheric Parameters, Chemical Abundances and Distances
Authors:
Matthias Steinmetz,
Guillaume Guiglion,
Paul J. McMillan,
Gal Matijevic,
Harry Enke,
Georges Kordopatis,
Tomaz Zwitter,
Marica Valentini,
Cristina Chiappini,
Luca Casagrande,
Jennifer Wojno,
Borja Anguiano,
Olivier Bienayme,
Albert Bijaoui,
James Binney,
Donna Burton,
Paul Cass,
Patrick de Laverny,
Kristin Fiegert,
Kenneth Freeman,
Jon P. Fulbright,
Brad K. Gibson,
Gerard Gilmore,
Eva K. Grebel,
Amina Helmi
, et al. (36 additional authors not shown)
Abstract:
We present part 2 of the 6th and final Data Release (DR6 or FDR) of the Radial Velocity Experiment (RAVE), a magnitude-limited (9<I<12) spectroscopic survey of Galactic stars randomly selected in the southern hemisphere. The RAVE medium-resolution spectra (R~7500) cover the Ca-triplet region (8410-8795A) and span the complete time frame from the start of RAVE observations on 12 April 2003 to their…
▽ More
We present part 2 of the 6th and final Data Release (DR6 or FDR) of the Radial Velocity Experiment (RAVE), a magnitude-limited (9<I<12) spectroscopic survey of Galactic stars randomly selected in the southern hemisphere. The RAVE medium-resolution spectra (R~7500) cover the Ca-triplet region (8410-8795A) and span the complete time frame from the start of RAVE observations on 12 April 2003 to their completion on 4 April 2013. In the second of two publications, we present the data products derived from 518387 observations of 451783 unique stars using a suite of advanced reduction pipelines focussing on stellar atmospheric parameters, in particular purely spectroscopically derived stellar atmospheric parameters (Teff, log(g), and the overall metallicity), enhanced stellar atmospheric parameters inferred via a Bayesian pipeline using Gaia DR2 astrometric priors, and asteroseismically calibrated stellar atmospheric parameters for giant stars based on asteroseismic observations for 699 K2 stars. In addition, we provide abundances of the elements Fe, Al, and Ni, as well as an overall [alpha/Fe] ratio obtained using a new pipeline based on the GAUGUIN optimization method that is able to deal with variable signal-to-noise ratios. The RAVE DR6 catalogs are cross matched with relevant astrometric and photometric catalogs, and are complemented by orbital parameters and effective temperatures based on the infrared flux method. The data can be accessed via the RAVE Web site (http://rave-survey.org) or the Vizier database.
△ Less
Submitted 9 June, 2020; v1 submitted 11 February, 2020;
originally announced February 2020.
-
The Sixth Data Release of the Radial Velocity Experiment (RAVE) -- I: Survey Description, Spectra and Radial Velocities
Authors:
Matthias Steinmetz,
Gal Matijevic,
Harry Enke,
Tomaz Zwitter,
Guillaume Guiglion,
Paul J. McMillan,
Georges Kordopatis,
Marica Valentini,
Cristina Chiappini,
Luca Casagrande,
Jennifer Wojno,
Borja Anguiano,
Olivier Bienayme,
Albert Bijaoui,
James Binney,
Donna Burton,
Paul Cass,
Patrick de Laverny,
Kristin Fiegert,
Kenneth Freeman,
Jon P. Fulbright,
Brad K. Gibson,
Gerard Gilmore,
Eva K. Grebel,
Amina Helmi
, et al. (37 additional authors not shown)
Abstract:
The Radial Velocity Experiment (RAVE) is a magnitude-limited (9<I<12) spectroscopic survey of Galactic stars randomly selected in the southern hemisphere. The RAVE medium-resolution spectra (R~7500) cover the Ca-triplet region (8410-8795A). The 6th and final data release (DR6 or FDR) is based on 518387 observations of 451783 unique stars. RAVE observations were taken between 12 April 2003 and 4 Ap…
▽ More
The Radial Velocity Experiment (RAVE) is a magnitude-limited (9<I<12) spectroscopic survey of Galactic stars randomly selected in the southern hemisphere. The RAVE medium-resolution spectra (R~7500) cover the Ca-triplet region (8410-8795A). The 6th and final data release (DR6 or FDR) is based on 518387 observations of 451783 unique stars. RAVE observations were taken between 12 April 2003 and 4 April 2013. Here we present the genesis, setup and data reduction of RAVE as well as wavelength-calibrated and flux-normalized spectra and error spectra for all observations in RAVE DR6. Furthermore, we present derived spectral classification and radial velocities for the RAVE targets, complemented by cross matches with Gaia DR2 and other relevant catalogs. A comparison between internal error estimates, variances derived from stars with more than one observing epoch and a comparison with radial velocities of Gaia DR2 reveals consistently that 68% of the objects have a velocity accuracy better than 1.4 km/s, while 95% of the objects have radial velocities better than 4.0 km/s. Stellar atmospheric parameters, abundances and distances are presented in subsequent publication. The data can be accessed via the RAVE Web (http://rave-survey.org) or the Vizier database.
△ Less
Submitted 9 June, 2020; v1 submitted 11 February, 2020;
originally announced February 2020.
-
Simulation of the loss-cone instability in spherical systems. II. Dominating Keplerian potential
Authors:
E. V. Polyachenko,
P. Berczik,
A. Just,
I. G. Shukhman
Abstract:
A new so-called `gravitational loss-cone instability' in stellar systems has recently been investigated theoretically in the framework of linear perturbation theory and proved to be potentially important in understanding the physical processes in centres of galaxies, star clusters, and the Oort comet cloud. Using N-body simulations, we confirm previous findings and go beyond the linear theory. Unl…
▽ More
A new so-called `gravitational loss-cone instability' in stellar systems has recently been investigated theoretically in the framework of linear perturbation theory and proved to be potentially important in understanding the physical processes in centres of galaxies, star clusters, and the Oort comet cloud. Using N-body simulations, we confirm previous findings and go beyond the linear theory. Unlike the well-known instabilities, the new one shows no notable change in spherical geometry of the cluster, but it significantly accelerates the speed of diffusion of particles in phase space leading to a repopulation of the loss cone and early instability saturation.
△ Less
Submitted 17 December, 2019;
originally announced December 2019.
-
Simulation of the loss-cone instability in spherical systems. I. Dominating harmonic potential
Authors:
E. V. Polyachenko,
P. Berczik,
A. Just,
I. G. Shukhman
Abstract:
A new so-called `gravitational loss-cone instability' in stellar systems has recently been investigated theoretically in the framework of linear perturbation theory and proved to be potentially important in understanding the physical processes in centres of galaxies, star clusters, and the Oort comet cloud. Using N-body simulations of a toy model, we confirm previous findings for the harmonic domi…
▽ More
A new so-called `gravitational loss-cone instability' in stellar systems has recently been investigated theoretically in the framework of linear perturbation theory and proved to be potentially important in understanding the physical processes in centres of galaxies, star clusters, and the Oort comet cloud. Using N-body simulations of a toy model, we confirm previous findings for the harmonic dominating potential and go beyond the linear theory. Unlike the well-known instabilities, the new one shows no notable change in the spherical geometry of the cluster, but it significantly accelerates the speed of diffusion of particles in phase space leading to an early instability saturation.
△ Less
Submitted 17 December, 2019;
originally announced December 2019.
-
Star Clusters in the Galactic tidal field, from birth to dissolution
Authors:
Bekdaulet Shukirgaliyev,
Genevieve Parmentier,
Peter Berczik,
Andreas Just
Abstract:
We study the evolution of star clusters in the Galactic tidal field starting from their birth in molecular clumps. Our model clusters form according to the local-density-driven cluster formation model in which the stellar density profile is steeper than that of gas. As a result, clusters resist the gas expulsion better than predicted by earlier models.
We vary the impact of the Galactic tidal fi…
▽ More
We study the evolution of star clusters in the Galactic tidal field starting from their birth in molecular clumps. Our model clusters form according to the local-density-driven cluster formation model in which the stellar density profile is steeper than that of gas. As a result, clusters resist the gas expulsion better than predicted by earlier models.
We vary the impact of the Galactic tidal field λ, considering different Galactocentric distances (3-18 kpc), as well as different cluster sizes. Our model clusters survive the gas expulsion independent of λ.
We investigated the relation between the cluster mass at the onset of secular evolution and their dissolution time. The model clusters formed with a high star-formation efficiency (SFE) follow a tight mass-dependent dissolution relation, in agreement with previous theoretical studies. However, the low-SFE models present a shallower mass-dependent relation than high-SFE clusters, and most dissolve before reaching 1 Gyr (cluster teenage mortality).
△ Less
Submitted 30 July, 2019;
originally announced July 2019.
-
Peculiar motions of the gas at the centre of the barred galaxy UGC 4056
Authors:
I. A. Zinchenko,
L. S. Pilyugin,
F. Sakhibov,
E. K. Grebel,
A. Just,
P. Berczik,
Y. A. Nefedyev,
J. M. Vilchez,
V. M. Shulga
Abstract:
We derive the circular velocity curves of the gaseous and stellar discs of UGC 4056, a giant barred galaxy with an active galactic nucleus (AGN). We analyse UGC 4056 using the 2D spectroscopy obtained within the framework of the Mapping Nearby Galaxies at APO (MaNGA) survey. Using images and the colour index g-r from the Sloan Digital Sky Survey (SDSS), we determined the tilt of the galaxy, which…
▽ More
We derive the circular velocity curves of the gaseous and stellar discs of UGC 4056, a giant barred galaxy with an active galactic nucleus (AGN). We analyse UGC 4056 using the 2D spectroscopy obtained within the framework of the Mapping Nearby Galaxies at APO (MaNGA) survey. Using images and the colour index g-r from the Sloan Digital Sky Survey (SDSS), we determined the tilt of the galaxy, which allows us to conclude that the galaxy rotates clockwise with trailing spiral arms. We found that the gas motion at the central part of the UGC 4056 shows peculiar features. The rotation velocity of the gaseous disc shows a bump within around three kiloparsecs while the rotation velocity of the stellar disc falls smoothly to zero with decreasing galactocentric distance. We demonstrate that the peculiar radial velocities in the central part of the galaxy may be caused by the inflow of the gas towards the nucleus of the galaxy. The unusual motion of the gas takes place at the region with the AGN-like radiation and can be explained by the gas response to the bar potential.
△ Less
Submitted 18 July, 2019;
originally announced July 2019.
-
The star cluster survivability after gas expulsion is independent of the impact of the Galactic tidal field
Authors:
B. Shukirgaliyev,
G. Parmentier,
P. Berczik,
A. Just
Abstract:
We study the impact of the tidal field on the survivability of star clusters following instantaneous gas expulsion. Our model clusters are formed with a centrally-peaked star-formation efficiency profile as a result of star-formation taking place with a constant efficiency per free-fall time. We define the impact of the tidal field as the ratio of the cluster half-mass radius to its Jacobi radius…
▽ More
We study the impact of the tidal field on the survivability of star clusters following instantaneous gas expulsion. Our model clusters are formed with a centrally-peaked star-formation efficiency profile as a result of star-formation taking place with a constant efficiency per free-fall time. We define the impact of the tidal field as the ratio of the cluster half-mass radius to its Jacobi radius immediately after gas expulsion, $λ= r_{h}/R_{J}$. We vary $λ$ by varying either the Galactocentric distance, or the size (hence volume density) of star clusters.
We propose a new method to measure the violent relaxation duration, in which we compare the total mass-loss rate of star clusters with their stellar evolutionary mass-loss rate. That way, we can robustly estimate the bound mass fraction of our model clusters at the end of violent relaxation. The duration of violent relaxation correlates linearly with the Jacobi radius, when considering identical clusters at different Galactocentric distances. In contrast, it is nearly constant for the solar neighbourhood clusters, slightly decreasing with $λ$. The violent relaxation does not last longer than 50 Myr in our simulations.
Identical model clusters placed at different Galactocentric distances have the same final bound fraction, despite experiencing different impacts of the tidal field. The solar neighbourhood clusters with different densities experience only limited variations of their final bound fraction.
In general, we conclude that the cluster survivability after instantaneous gas expulsion, as measured by their bound mass fraction at the end of violent relaxation, $F_{bound}$, is independent of the impact of the tidal field, $λ$.
△ Less
Submitted 22 March, 2019;
originally announced March 2019.
-
Gaussian Markov Random Fields versus Linear Mixed Models for satellite-based PM2.5 assessment: Evidence from the Northeastern USA
Authors:
Ron Sarafian,
Itai Kloog,
Allan C. Just,
Johnathan D. Rosenblatt
Abstract:
Studying the effects of air-pollution on health is a key area in environmental epidemiology. An accurate estimation of air-pollution effects requires spatio-temporally resolved datasets of air-pollution, especially, Fine Particulate Matter (PM). Satellite-based technology has greatly enhanced the ability to provide PM assessments in locations where direct measurement is impossible.
Indirect PM m…
▽ More
Studying the effects of air-pollution on health is a key area in environmental epidemiology. An accurate estimation of air-pollution effects requires spatio-temporally resolved datasets of air-pollution, especially, Fine Particulate Matter (PM). Satellite-based technology has greatly enhanced the ability to provide PM assessments in locations where direct measurement is impossible.
Indirect PM measurement is a statistical prediction problem. The spatio-temporal statistical literature offer various predictive models: Gaussian Random Fields (GRF) and Linear Mixed Models (LMM), in particular. GRF emphasize the spatio-temporal structure in the data, but are computationally demanding to fit. LMMs are computationally easier to fit, but require some tampering to deal with space and time.
Recent advances in the spatio-temporal statistical literature propose to alleviate the computation burden of GRFs by approximating them with Gaussian Markov Random Fields (GMRFs). Since LMMs and GMRFs are both computationally feasible, the question arises: which is statistically better? We show that despite the great popularity of LMMs in environmental monitoring and pollution assessment, LMMs are statistically inferior to GMRF for measuring PM in the Northeastern USA.
△ Less
Submitted 22 February, 2019;
originally announced February 2019.
-
Discovery of Tidal Tails in Disrupting Open Clusters: Coma Berenices and a Neighbor Stellar Group
Authors:
Shih-Yun Tang,
Xiaoying Pang,
Zhen Yuan,
W. P. Chen,
Jongsuk Hong,
Bertrand Goldman,
Andreas Just,
Bekdaulet Shukirgaliyev,
Chien-Cheng Lin
Abstract:
We report the discovery of tidal structures around the intermediate-aged ($\sim$ 700--800~Myr), nearby ($\sim85$~pc) star cluster Coma Berenices. The spatial and kinematic grouping of stars is determined with the {\it Gaia} DR2 parallax and proper motion data, by a clustering analysis tool, \textsc{StarGO}, to map 5D parameters ($X, Y, Z$, $μ_α\cosδ, μ_δ$) onto a 2D neural network. A leading and a…
▽ More
We report the discovery of tidal structures around the intermediate-aged ($\sim$ 700--800~Myr), nearby ($\sim85$~pc) star cluster Coma Berenices. The spatial and kinematic grouping of stars is determined with the {\it Gaia} DR2 parallax and proper motion data, by a clustering analysis tool, \textsc{StarGO}, to map 5D parameters ($X, Y, Z$, $μ_α\cosδ, μ_δ$) onto a 2D neural network. A leading and a trailing tails, each with an extension of $\sim50$~pc are revealed for the first time around this disrupting star cluster. The cluster members, totaling to $\sim115^{+5}_{-3}\,\rm {M_\odot}$, are clearly mass segregated, and exhibit a flat mass function with $α\sim 0.79\pm0.16$, in the sense of $dN/dm \propto m^{-α}$, where $N$ is the number of member stars and $m$ is stellar mass, in the mass range of $m=0.25$--$2.51~{\rm M_\odot}$. Within the tidal radius of $\sim$6.9~pc, there are 77 member candidates with an average position, i.e., as the cluster center, of R.A.= 186.8110~deg, and decl.= 25.8112~deg, and an average distance of 85.8~pc. Additional 120 member candidates reside in the tidal structures, i.e., outnumbering those in the cluster core. The expansion of escaping members lead to an anisotropy in the velocity field of the tidal tails. Our analysis also serendipitously uncovers an adjacent stellar group, part of which has been cataloged in the literature. We identify 218 member candidates, 10 times more than previously known. This star group is some 65~pc away from, and $\sim400$~Myr younger than, Coma Ber, but is already at the final stage of disruption.
△ Less
Submitted 25 March, 2019; v1 submitted 4 February, 2019;
originally announced February 2019.
-
Characterizing the radial oxygen abundance distribution in disk galaxies
Authors:
I. A. Zinchenko,
A. Just,
L. S. Pilyugin,
M. A. Lara-Lopez
Abstract:
We examine the possible dependence of the radial oxygen abundance distribution on non-axisymmetrical structures (bar/spirals) and other macroscopic parameters such as the mass, the optical radius R25, the color g-r, and the surface brightness of the galaxy. A sample of disk galaxies from the CALIFA DR3 is considered. We adopted the Fourier amplitude A2 of the surface brightness as a quantitative c…
▽ More
We examine the possible dependence of the radial oxygen abundance distribution on non-axisymmetrical structures (bar/spirals) and other macroscopic parameters such as the mass, the optical radius R25, the color g-r, and the surface brightness of the galaxy. A sample of disk galaxies from the CALIFA DR3 is considered. We adopted the Fourier amplitude A2 of the surface brightness as a quantitative characteristic of the strength of non-axisymmetric structures in a galactic disk, in addition to the commonly used morphologic division for A, AB, and B types based on the Hubble classification. To distinguish changes in local oxygen abundance caused by the non-axisymmetrical structures, the multiparametric mass--metallicity relation was constructed as a function of parameters such as the bar/spiral pattern strength, the disk size, color index g-r in the SDSS bands, and central surface brightness of the disk. The gas-phase oxygen abundance gradient is determined by using the R calibration. We find that there is no significant impact of the non-axisymmetric structures such as a bar and/or spiral patterns on the local oxygen abundance and radial oxygen abundance gradient of disk galaxies. Galaxies with higher mass, however, exhibit flatter oxygen abundance gradients in units of dex/kpc, but this effect is significantly less prominent for the oxygen abundance gradients in units of dex/R25 and almost disappears when the inner parts are avoided. We show that the oxygen abundance in the central part of the galaxy depends neither on the optical radius R25 nor on the color g-r or the surface brightness of the galaxy. Instead, outside the central part of the galaxy, the oxygen abundance increases with g-r value and central surface brightness of the disk.
△ Less
Submitted 31 January, 2019; v1 submitted 18 October, 2018;
originally announced October 2018.
-
Local disc model in view of Gaia DR1 and RAVE data
Authors:
K. Sysoliatina,
A. Just,
I. Koutsouridou,
E. K. Grebel,
G. Kordopatis,
M. Steinmetz,
O. Bienaymé,
B. K. Gibson,
J. Navarro,
W. Reid,
G. Seabroke
Abstract:
We test the performance of the semi-analytic self-consistent Just-Jahreiß disc model (JJ model) with the astrometric data from the Tycho-Gaia Astrometric Solution (TGAS) sub-catalogue of the first Gaia data release (Gaia DR1), as well as the radial velocities from the fifth data release of the Radial Velocity Experiment survey (RAVE DR5). We use a sample of 19,746 thin disc stars from the TGAS…
▽ More
We test the performance of the semi-analytic self-consistent Just-Jahreiß disc model (JJ model) with the astrometric data from the Tycho-Gaia Astrometric Solution (TGAS) sub-catalogue of the first Gaia data release (Gaia DR1), as well as the radial velocities from the fifth data release of the Radial Velocity Experiment survey (RAVE DR5). We use a sample of 19,746 thin disc stars from the TGAS$\times$RAVE cross-match selected in the local solar cylinder of 300 pc radius and 1 kpc height below the Galactic plane and simulate this sample via the forward modelling technique. First, we convert the predicted vertical density laws of the thin disc populations into a mock sample. Then the obtained mock populations are reddened with a 3D dust map and are subjected to the selection criteria corresponding to the RAVE and TGAS observational limitations as well as to additional cuts applied to the data sample. We calculate the quantities of interest separately at different heights above the Galactic plane taking into account the distance error effects separately in horizontal and vertical directions. We investigate the simulated sample in terms of the vertical number density profiles, Hess diagrams and velocity distribution functions. Basing on a good agreement of our simulations with the data, we conclude that our fiducial disc model confidently reproduces the vertical trends in the thin disc stellar population properties. Thus, it can serve as a starting point for the future extension of the JJ model to other Galactocentric distances.
△ Less
Submitted 1 October, 2018;
originally announced October 2018.
-
The long-term evolution of star clusters formed with a centrally-peaked star-formation-efficiency profile
Authors:
Bekdaulet Shukirgaliyev,
Genevieve Parmentier,
Andreas Just,
Peter Berczik
Abstract:
We have studied the long-term evolution of star clusters of the solar neighborhood, starting from their birth in gaseous clumps until their complete dissolution in the Galactic tidal field. We have combined the "local-density-driven cluster formation model" of Parmentier & Pfalzner (2013) with direct N-body simulations of clusters following instantaneous gas expulsion.
We have studied the relati…
▽ More
We have studied the long-term evolution of star clusters of the solar neighborhood, starting from their birth in gaseous clumps until their complete dissolution in the Galactic tidal field. We have combined the "local-density-driven cluster formation model" of Parmentier & Pfalzner (2013) with direct N-body simulations of clusters following instantaneous gas expulsion.
We have studied the relation between cluster dissolution time, $t_{dis}$, and cluster "initial" mass, $M_{init}$, defined as the cluster {mass at the end of the dynamical response to gas expulsion (i.e. violent relaxation), when the cluster age is 20-30 Myr}. We consider the "initial" mass to be consistent with other works which neglect violent relaxation. The model clusters formed with a high star formation efficiency (SFE -- i.e. gas mass fraction converted into stars) follow a tight mass-dependent relation, in agreement with previous theoretical studies. However, the low-SFE models present a large scatter in both the "initial" mass and the dissolution time, and a shallower mass-dependent relation than high-SFE clusters. Both groups differ in their structural properties on the average. Combining two populations of clusters, high- and low-SFE ones, with domination of the latter, yields a cluster dissolution time for the solar neighborhood in agreement with that inferred from observations, without any additional destructive processes such as giant molecular cloud encounters.
An apparent mass-independent relation may emerge for our low-SFE clusters when we neglect low-mass clusters (as expected for extra-galactic observations), although more simulations are needed to investigate this aspect.
△ Less
Submitted 12 July, 2018;
originally announced July 2018.
-
Direct N-body simulation of the Galactic centre
Authors:
Taras Panamarev,
Andreas Just,
Rainer Spurzem,
Peter Berczik,
Long Wang,
Manuel Arca Sedda
Abstract:
We study the dynamics and evolution of the Milky Way nuclear star cluster performing a high resolution direct one-million-body simulation. Focusing on the interactions between such stellar system and the central supermassive black hole, we find that different stellar components adapt their overall distribution differently. After 5 Gyr, stellar mass black holes are characterized by a spatial distri…
▽ More
We study the dynamics and evolution of the Milky Way nuclear star cluster performing a high resolution direct one-million-body simulation. Focusing on the interactions between such stellar system and the central supermassive black hole, we find that different stellar components adapt their overall distribution differently. After 5 Gyr, stellar mass black holes are characterized by a spatial distribution with power-slope $-1.75$, fully consistent with the prediction of Bahcall-Wolf pioneering work. Using the vast amount of data available, we infer the rate for tidal disruption events, being $4 \times 10^{-6}$ per yr, and estimate the number of objects that emit gravitational waves during the phases preceding the accretion onto the super-massive black hole, $\sim 270$ per Gyr. We show that some of these sources could form extreme mass-ratio inspirals. We follow the evolution of binary stars population, showing that the initial binary fraction of $5\%$ drops down to $2.5\%$ inside the inner parsec. Also, we explored the possible formation of binary systems containing a compact object, discussing the implications for millisecond pulsars formation and the development of Ia Supernovae.
△ Less
Submitted 18 January, 2019; v1 submitted 6 May, 2018;
originally announced May 2018.
-
Gravitational Wave Driven Mergers and Coalescence Time of Supermassive Black Holes
Authors:
Fazeel Mahmood Khan,
Peter Berczik,
Andreas Just
Abstract:
The evolution of Supermassive Black Holes (SMBHs) initially embedded in the centres of merging galaxies realised with a stellar mass function (SMF) is studied from the onset of galaxy mergers till coalescence. We performed a large set of direct N-body simulations with three different slopes of the central stellar cusp and different random seeds. Post Newtonian terms up to order 3.5 are used to dri…
▽ More
The evolution of Supermassive Black Holes (SMBHs) initially embedded in the centres of merging galaxies realised with a stellar mass function (SMF) is studied from the onset of galaxy mergers till coalescence. We performed a large set of direct N-body simulations with three different slopes of the central stellar cusp and different random seeds. Post Newtonian terms up to order 3.5 are used to drive the SMBH binary evolution in the relativistic regime. The impact of a SMF on the hardening rate and the coalescence time is investigated. We find that SMBH binaries coalesce well within one billion years when our models are scaled to galaxies with a steep cusp at low redshift. Here higher central densities provide larger supply of stars to efficiently extract energy from the SMBH binary orbit and shrink it to the phase where gravitational wave (GW) emission becomes dominant leading to the coalescence of the SMBHs. Mergers of models with shallow cusps that are representative for giant elliptical galaxies having central cores result in less efficient extraction of binary orbital energy due to the lower stellar densities in the centre. However, high values of eccentricity witnessed for SMBH binaries in such galaxy mergers ensure that the GW emission dominated phase sets in earlier at larger values of the semi-major axis. This helps to compensate for the less efficient energy extraction during the phase dominated by stellar encounters resulting in mergers of SMBHs in about one Gyr after the formation of the binary. Additionally, we witness mass segregation in the merger remnant resulting in enhanced SMBH binary hardening rates. We show that at least the final phase of the merger in cuspy low mass galaxies would be observable with the GW detector eLISA.
△ Less
Submitted 30 March, 2018;
originally announced March 2018.
-
The local rotation curve of the Milky Way based on SEGUE and RAVE data
Authors:
K. Sysoliatina,
A. Just,
O. Golubov,
Q. A. Parker,
E. K. Grebel,
G. Kordopatis,
T. Zwitter,
J. Bland-Hawthorn,
B. K. Gibson,
A. Kunder,
U. Munari,
J. Navarro,
W. Reid,
G. Seabroke,
M. Steinmetz,
F. Watson
Abstract:
We construct the rotation curve of the Milky Way in the extended solar neighbourhood using a sample of SEGUE (Sloan Extension for Galactic Understanding and Exploration) G-dwarfs. We investigate the rotation curve shape for the presence of any peculiarities just outside the solar radius as has been reported by some authors. We approach the problem in a framework of classical Jeans analysis. Using…
▽ More
We construct the rotation curve of the Milky Way in the extended solar neighbourhood using a sample of SEGUE (Sloan Extension for Galactic Understanding and Exploration) G-dwarfs. We investigate the rotation curve shape for the presence of any peculiarities just outside the solar radius as has been reported by some authors. We approach the problem in a framework of classical Jeans analysis. Using the most recent data from RAVE (RAdial Velocity Experiment), we determine the solar peculiar velocity and the radial scalelengths for the three populations of different metallicities representing the Galactic thin disc. Then with the same binning in metallicity for the SEGUE G-dwarfs, we construct the rotation curve in the range of Galactocentric distances 7-10 kpc. We derive the circular velocity by correcting the mean tangential velocity for the asymmetric drift in each distance bin. With SEGUE data we also calculate the radial scalelength of the thick disc taking as known the derived peculiar motion of the Sun and the slope of the rotation curve. The rotation curve constructed through SEGUE G-dwarfs appears to be smooth in the selected radial range. The local kinematics of the thin disc rotation as determined in the framework of our new careful analysis does not favour the presence of a massive overdensity ring just outside the solar radius.
△ Less
Submitted 21 February, 2018;
originally announced February 2018.
-
Global survey of star clusters in the Milky Way VI. Age distribution and cluster formation history
Authors:
A. E. Piskunov,
A. Just,
N. V. Kharchenko,
P. Berczik,
R. -D. Scholz,
S. Reffert,
S. X. Yen
Abstract:
The all-sky Milky Way Star Clusters (MWSC) survey provides uniform and precise ages and other parameters for a variety of clusters in the Solar Neighbourhood. We construct the cluster age distribution, investigate its spatial variations, and discuss constraints on cluster formation scenarios of the Galactic disk during the last 5 Gyrs. Due to the spatial extent of the MWSC, we consider spatial var…
▽ More
The all-sky Milky Way Star Clusters (MWSC) survey provides uniform and precise ages and other parameters for a variety of clusters in the Solar Neighbourhood. We construct the cluster age distribution, investigate its spatial variations, and discuss constraints on cluster formation scenarios of the Galactic disk during the last 5 Gyrs. Due to the spatial extent of the MWSC, we consider spatial variations of the age distribution along galactocentric radius $R_G$, and along $Z$-axis. For the analysis of the age distribution we use 2242 clusters, which all lie within roughly 2.5 kpc of the Sun. To connect the observed age distribution to the cluster formation history we build an analytical model based on simple assumptions on the cluster initial mass function and on the cluster mass-lifetime relation, fit it to the observations, and determine the parameters of the cluster formation law. Comparison with the literature shows that earlier results strongly underestimated the number of evolved clusters with ages $t\gtrsim 100$ Myr. Recent studies based on all-sky catalogues agree better with our data, but still lack the oldest clusters with ages $t\gtrsim 1$ Gyr. We do not observe a strong variation in the age distribution along $R_G$, though we find an enhanced fraction of older clusters ($t>1$ Gyr) in the inner disk. In contrast, the distribution strongly varies along $Z$. The high altitude distribution practically does not contain clusters with $t<1$ Gyr. With simple assumptions on the cluster formation history, cluster initial mass function and cluster lifetime we can reproduce the observations. Cluster formation rate and cluster lifetime are strongly degenerate, which does not allow us to disentangle different formation scenarios. In all cases the cluster formation rate is strongly declining with time, and the cluster initial mass function is very shallow at the high mass end. (abridged)
△ Less
Submitted 19 February, 2018;
originally announced February 2018.
-
Star-disc interaction in galactic nuclei: formation of a central stellar disc
Authors:
Taras Panamarev,
Bekdaulet Shukirgaliyev,
Yohai Meiron,
Peter Berczik,
Andreas Just,
Rainer Spurzem,
Chingis Omarov,
Emmanuil Vilkoviskij
Abstract:
We perform high resolution direct $N$-body simulations to study the effect of an accretion disc on stellar dynamics in an active galactic nucleus (AGN). We show that the interaction of the nuclear stellar cluster (NSC) with the gaseous disc (AD) leads to formation of a stellar disc in the central part of the NSC. The accretion of stars from the stellar disc onto the super-massive black hole is bal…
▽ More
We perform high resolution direct $N$-body simulations to study the effect of an accretion disc on stellar dynamics in an active galactic nucleus (AGN). We show that the interaction of the nuclear stellar cluster (NSC) with the gaseous disc (AD) leads to formation of a stellar disc in the central part of the NSC. The accretion of stars from the stellar disc onto the super-massive black hole is balanced by the capture of stars from the NSC into the stellar disc, yielding a stationary density profile. We derive the migration time through the AD to be 3\% of the half-mass relaxation time of the NSC. The mass and size of the stellar disc are 0.7\% of the mass and 5\% of the influence radius of the super-massive black hole. An AD lifetime shorter than the migration time would result in a less massive nuclear stellar disc. The detection of such a stellar disc could point to past activity of the hosting galactic nucleus.
△ Less
Submitted 8 February, 2018;
originally announced February 2018.
-
Spiral arms in CALIFA galaxies traced by non-circular velocities, abundances, and extinctions
Authors:
F. Sakhibov,
I. A. Zinchenko,
L. S. Pilyugin,
E. K. Grebel,
A. Just,
J. M. Vilchez
Abstract:
We derive maps of the observed velocity of ionized gas, the oxygen abundance, and the extinction (Balmer decrement) across the area of the four spiral galaxies NGC36, NGC180, NGC6063, and NGC7653 from integral field spectroscopy obtained by the Calar Alto Legacy Integral Field Area (CALIFA) survey. We searched for spiral arms through Fourier analysis of the spatial distribution of three tracers (n…
▽ More
We derive maps of the observed velocity of ionized gas, the oxygen abundance, and the extinction (Balmer decrement) across the area of the four spiral galaxies NGC36, NGC180, NGC6063, and NGC7653 from integral field spectroscopy obtained by the Calar Alto Legacy Integral Field Area (CALIFA) survey. We searched for spiral arms through Fourier analysis of the spatial distribution of three tracers (non-circular motion, enhancement of the oxygen abundance, and of the extinction) in the discs of our target galaxies. The spiral arms (two-armed logarithmic spirals in the deprojected map) are shown in each target galaxy for each tracer considered. The pitch angles of the spiral arms in a given galaxy obtained with the three different tracers are close to each other. The enhancement of the oxygen abundance in the spiral arms as compared to the abundance in the interarm regions at a given galactocentric distance is small; within a few per cent. We identified a metallicity gradient in our target galaxies. Both barred galaxies in our sample show flatter gradients than the two galaxies without bars. Galactic inclination, position angle of the major axis, and the rotation curve were also obtained for each target galaxy using the Fourier analysis of the two-dimensional velocity map.
△ Less
Submitted 1 November, 2017; v1 submitted 27 October, 2017;
originally announced October 2017.
-
Is the Milky Way still breathing? RAVE-Gaia streaming motions
Authors:
I. Carrillo,
I. Minchev,
G. Kordopatis,
M. Steinmetz,
J. Binney,
F. Anders,
O. Bienaymé,
J. Bland-Hawthorn,
B. Famaey,
K. C. Freeman,
G. Gilmore,
B. K. Gibson,
E. K. Grebel,
A. Helmi,
A. Just,
A. Kunder,
P. McMillan,
G. Monari,
U. Munari,
J. Navarro,
Q. A. Parker,
W. Reid,
G. Seabroke,
S. Sharma,
A. Siebert
, et al. (4 additional authors not shown)
Abstract:
We use data from the Radial Velocity Experiment (RAVE) and the Tycho-Gaia astrometric solution catalogue (TGAS) to compute the velocity fields yielded by the radial (VR), azimuthal (Vphi) and vertical (Vz) components of associated Galactocentric velocity. We search in particular for variation in all three velocity components with distance above and below the disc midplane, as well as how each comp…
▽ More
We use data from the Radial Velocity Experiment (RAVE) and the Tycho-Gaia astrometric solution catalogue (TGAS) to compute the velocity fields yielded by the radial (VR), azimuthal (Vphi) and vertical (Vz) components of associated Galactocentric velocity. We search in particular for variation in all three velocity components with distance above and below the disc midplane, as well as how each component of Vz (line-of-sight and tangential velocity projections) modifies the obtained vertical structure. To study the dependence of velocity on proper motion and distance we use two main samples: a RAVE sample including proper motions from the Tycho-2, PPMXL and UCAC4 catalogues, and a RAVE-TGAS sample with inferred distances and proper motions from the TGAS and UCAC5 catalogues. In both samples, we identify asymmetries in VR and Vz. Below the plane we find the largest radial gradient to be dVR / dR = -7.01+- 0.61 km\s kpc, in agreement with recent studies. Above the plane we find a similar gradient with dVR / dR= -9.42+- 1.77 km\s kpc. By comparing our results with previous studies, we find that the structure in Vz is strongly dependent on the adopted proper motions. Using the Galaxia Milky Way model, we demonstrate that distance uncertainties can create artificial wave-like patterns. In contrast to previous suggestions of a breathing mode seen in RAVE data, our results support a combination of bending and breathing modes, likely generated by a combination of external or internal and external mechanisms.
△ Less
Submitted 15 February, 2018; v1 submitted 10 October, 2017;
originally announced October 2017.
-
Impact of a star formation efficiency profile on the evolution of open clusters
Authors:
Bekdaulet Shukirgaliyev,
Genevieve Parmentier,
Peter Berczik,
Andreas Just
Abstract:
We study the effect of the instantaneous gas expulsion on star clusters wherein the residual gas has a density profile shallower than that of the embedded cluster. This is expected if star formation proceeds with a given SFE per free-fall time in a centrally-concentrated molecular clump. We perform direct N-body simulations whose initial conditions are generated by the program "mkhalo" "falcON" ad…
▽ More
We study the effect of the instantaneous gas expulsion on star clusters wherein the residual gas has a density profile shallower than that of the embedded cluster. This is expected if star formation proceeds with a given SFE per free-fall time in a centrally-concentrated molecular clump. We perform direct N-body simulations whose initial conditions are generated by the program "mkhalo" "falcON" adapted for our models. Our model clusters initially have a Plummer profile and are in virial equilibrium with the gravitational potential of the cluster-forming clump. The residual gas contribution is computed based on the model of Parmentier&Pfalzner(2013). Our simulations include mass loss by stellar evolution and the tidal field of the Galaxy. We find that a star cluster with a minimum global SFE of 15% is able to survive instantaneous gas expulsion and to produce a bound cluster. Its violent relaxation lasts no longer than 20 Myr, independently of its global SFE and initial stellar mass. At the end of violent relaxation the bound fractions of surviving clusters with the same global SFEs are similar regardless of their initial stellar mass. Their subsequent lifetime in the gravitational field of the Galaxy depends on their bound stellar masses. We therefore conclude that the critical SFE needed to produce a bound cluster is 15%, which is twice smaller than earlier estimates of 33%. Thus we have improved the survival likelihood of young clusters after instantaneous gas expulsion. Those can now survive instantaneous gas expulsion with global SFEs as low as those observed for embedded clusters of Solar Neighbourhood (15-30%). This is the consequence of the star cluster having a density profile steeper than that of the residual gas. However, in terms of the effective SFE, measured by the virial ratio of the cluster at gas expulsion, our results are in agreement with previous studies.
△ Less
Submitted 21 June, 2017; v1 submitted 10 June, 2017;
originally announced June 2017.
-
Effects of galaxy--satellite interactions on bar formation
Authors:
R. Moetazedian,
E. V. Polyachenko,
P. Berczik,
A. Just
Abstract:
Aims. We aim to show how encounters with low-mass satellite galaxies may alter the bar formation in a Milky Way-like disc galaxy. Methods. We use high-resolution N-body simulations of a disc galaxy prone to mild bar instability. For realistic initial conditions of satellites, we take advantage of cosmological simulations of Milky Way-like dark matter haloes. Results. The satellites may have a sign…
▽ More
Aims. We aim to show how encounters with low-mass satellite galaxies may alter the bar formation in a Milky Way-like disc galaxy. Methods. We use high-resolution N-body simulations of a disc galaxy prone to mild bar instability. For realistic initial conditions of satellites, we take advantage of cosmological simulations of Milky Way-like dark matter haloes. Results. The satellites may have a significant impact on the time of bar formation. Some runs with satellites demonstrate a delay, while others show an advancement in bar formation compared to the isolated run, with such time differences reaching $\sim$ 1 Gyr. Meanwhile, the final bar configuration, including its very appearance and the bar characteristics such as the pattern speed and the exponential growth rate of its amplitude are independent of the number of encounters and their orbits. The contribution of satellites with masses below $10^9 M_{\odot}$ is insignificant, unless their pericentre distances are small. We suggest that the encounters act indirectly via inducing perturbations across the disc that evolve to delayed waves in the central part and interfere with an emerging seed bar. The predicted effect for the present-day host galaxy is expected to be even more significant at redshifts $z \gtrsim 0.5$.
△ Less
Submitted 17 March, 2017;
originally announced March 2017.
-
$\mathit{Chempy}$: A flexible chemical evolution model for abundance fitting - Do the Sun's abundances alone constrain chemical evolution models?
Authors:
Jan Rybizki,
Andreas Just,
Hans-Walter Rix
Abstract:
Elemental abundances of stars are the result of the complex enrichment history of their galaxy. Interpretation of observed abundances requires flexible modeling tools to explore and quantify the information about Galactic chemical evolution (GCE) stored in such data. Here we present Chempy, a newly developed code for GCE modeling, representing a parametrized open one-zone model within a Bayesian f…
▽ More
Elemental abundances of stars are the result of the complex enrichment history of their galaxy. Interpretation of observed abundances requires flexible modeling tools to explore and quantify the information about Galactic chemical evolution (GCE) stored in such data. Here we present Chempy, a newly developed code for GCE modeling, representing a parametrized open one-zone model within a Bayesian framework. A Chempy model is specified by a set of 5-10 parameters that describe the effective galaxy evolution along with the stellar and star-formation physics: e.g. the star-formation history, the feedback efficiency, the stellar initial mass function (IMF) and the incidence of supernova type Ia (SN Ia). Unlike established approaches, Chempy can sample the posterior probability distribution in the full model parameter space and test data-model matches for different nucleosynthetic yield sets. We extend Chempy to a multi-zone scheme. As an illustrative application, we show that interesting parameter constraints result from only the ages and elemental abundances of Sun, Arcturus and the present-day interstellar medium (ISM). For the first time, we use such information to infer IMF parameter via GCE modeling, where we properly marginalize over nuisance parameters and account for different yield sets. We find that of the IMF $11.6_{-1.6}^{+2.1}$ % explodes as core-collapse SN, compatible with Salpeter 1955. We also constrain the incidence of SN Ia per 10^3 Msun to 0.5-1.4. At the same time, this Chempy application shows persistent discrepancies between predicted and observed abundances for some elements, irrespective of the chosen yield set. These cannot be remedied by any variations of Chempy's parameters and could be an indication for missing nucleosynthetic channels. Chempy should be a powerful tool to confront predictions from stellar nucleosynthesis with far more complex abundance data sets.
△ Less
Submitted 24 April, 2017; v1 submitted 28 February, 2017;
originally announced February 2017.
-
Bar formation in the Milky Way type galaxies
Authors:
E. V. Polyachenko,
P. Berczik,
A. Just
Abstract:
Many barred galaxies, possibly including the Milky Way, have cusps in the centres. There is a widespread belief, however, that usual bar instability taking place in bulgeless galaxy models is impossible for the cuspy models, because of the presence of the inner Lindblad resonance for any pattern speed. At the same time there are numerical evidences that the bar instability can form a bar. We analy…
▽ More
Many barred galaxies, possibly including the Milky Way, have cusps in the centres. There is a widespread belief, however, that usual bar instability taking place in bulgeless galaxy models is impossible for the cuspy models, because of the presence of the inner Lindblad resonance for any pattern speed. At the same time there are numerical evidences that the bar instability can form a bar. We analyse this discrepancy, by accurate and diverse N-body simulations and using the calculation of normal modes. We show that bar formation in cuspy galaxies can be explained by taking into account the disc thickness. The exponential growth time is moderate for typical current disc masses (about 250 Myr), but considerably increases (factor 2 or more) upon substitution of the live halo and bulge with a rigid halo/bulge potential; meanwhile pattern speeds remain almost the same. Normal mode analysis with different disc mass favours a young bar hypothesis, according to which the bar instability saturated only recently.
△ Less
Submitted 6 February, 2017;
originally announced February 2017.
-
Star-disc interaction in galactic nuclei: orbits and rates of accreted stars
Authors:
Gareth F. Kennedy,
Yohai Meiron,
Bekdaulet Shukirgaliyev,
Taras Panamarev,
Peter Berczik,
Andreas Just,
Rainer Spurzem
Abstract:
We examine the effect of an accretion disc on the orbits of stars in the central star cluster surrounding a central massive black hole by performing a suite of 39 high-accuracy direct N-body simulations using state-of-the art software and accelerator hardware, with particle numbers up to 128k. The primary focus is on the accretion rate of stars by the black hole (equivalent to their tidal disrupti…
▽ More
We examine the effect of an accretion disc on the orbits of stars in the central star cluster surrounding a central massive black hole by performing a suite of 39 high-accuracy direct N-body simulations using state-of-the art software and accelerator hardware, with particle numbers up to 128k. The primary focus is on the accretion rate of stars by the black hole (equivalent to their tidal disruption rate for black holes in the small to medium mass range) and the eccentricity distribution of these stars. Our simulations vary not only the particle number, but disc model (two models examined), spatial resolution at the centre (characterised by the numerical accretion radius) and softening length. The large parameter range and physically realistic modelling allow us for the first time to confidently extrapolate these results to real galactic centres. While in a real galactic centre both particle number and accretion radius differ by a few orders of magnitude from our models, which are constrained by numerical capability, we find that the stellar accretion rate converges for models with N > 32k. The eccentricity distribution of accreted stars, however, does not converge. We find that there are two competing effects at work when improving the resolution: larger particle number leads to a smaller fraction of stars accreted on nearly-circular orbits, while higher spatial resolution increases this fraction. We scale our simulations to some nearby galaxies and find that the expected boost in stellar accretion (or tidal disruption, which could be observed as X-ray flares) in the presence of a gas disc is about a factor of 10. Even with this boost, the accretion of mass from stars is still a factor of ~ 100 slower than the accretion of gas from the disc. Thus, it seems accretion of stars is not a major contributor to black hole mass growth.
△ Less
Submitted 18 April, 2016;
originally announced April 2016.