-
Multimodal Image-Text Matching Improves Retrieval-based Chest X-Ray Report Generation
Authors:
Jaehwan Jeong,
Katherine Tian,
Andrew Li,
Sina Hartung,
Fardad Behzadi,
Juan Calle,
David Osayande,
Michael Pohlen,
Subathra Adithan,
Pranav Rajpurkar
Abstract:
Automated generation of clinically accurate radiology reports can improve patient care. Previous report generation methods that rely on image captioning models often generate incoherent and incorrect text due to their lack of relevant domain knowledge, while retrieval-based attempts frequently retrieve reports that are irrelevant to the input image. In this work, we propose Contrastive X-Ray REpor…
▽ More
Automated generation of clinically accurate radiology reports can improve patient care. Previous report generation methods that rely on image captioning models often generate incoherent and incorrect text due to their lack of relevant domain knowledge, while retrieval-based attempts frequently retrieve reports that are irrelevant to the input image. In this work, we propose Contrastive X-Ray REport Match (X-REM), a novel retrieval-based radiology report generation module that uses an image-text matching score to measure the similarity of a chest X-ray image and radiology report for report retrieval. We observe that computing the image-text matching score with a language-image model can effectively capture the fine-grained interaction between image and text that is often lost when using cosine similarity. X-REM outperforms multiple prior radiology report generation modules in terms of both natural language and clinical metrics. Human evaluation of the generated reports suggests that X-REM increased the number of zero-error reports and decreased the average error severity compared to the baseline retrieval approach. Our code is available at: https://github.com/rajpurkarlab/X-REM
△ Less
Submitted 2 May, 2023; v1 submitted 29 March, 2023;
originally announced March 2023.
-
Comment on "Hysteretic transition between states of a filled hexagonal magnetic dipole cluster"
Authors:
Simeon Völkel,
Stefan Hartung,
Ingo Rehberg
Abstract:
In the paper "Andrew D.P. Smith, Peter T. Haugen, Boyd F. Edwards: Hysteretic transition between states of a filled hexagonal magnetic dipole cluster, Journal of Magnetism and Magnetic Materials 549 (2022): 168991" a hysteretic transition between two stable arrangements of a cluster of seven dipoles is presented. The relative strength of the center dipole in a hexagonal arrangement serves as the b…
▽ More
In the paper "Andrew D.P. Smith, Peter T. Haugen, Boyd F. Edwards: Hysteretic transition between states of a filled hexagonal magnetic dipole cluster, Journal of Magnetism and Magnetic Materials 549 (2022): 168991" a hysteretic transition between two stable arrangements of a cluster of seven dipoles is presented. The relative strength of the center dipole in a hexagonal arrangement serves as the bifurcation parameter. The authors clearly demonstrate the existence of two instabilities accompanied by discontinuous jumps of the dipole arrangement, but leave the question about the nature of these instabilities unanswered. This comment clarifies the nature of the two instabilities: the first one is a symmetry-breaking sub-critical bifurcation with parabolic scaling of the magnetic potential energy difference between the two branches, and the second one is a fold with its characteristic scaling in the form of a semi-cubic parabola.
△ Less
Submitted 7 June, 2022; v1 submitted 25 March, 2022;
originally announced March 2022.
-
Dynamics of a magnetic gear with two cogging-free operation modes
Authors:
Stefan Hartung,
Ingo Rehberg
Abstract:
The coupling of two rotating spherical magnets is investigated experimentally. For two specific angles between the input and output rotation axes, a cogging-free coupling is observed, where the driven magnet is phase-locked to the driving one. The striking difference between these two modes of operation is the reversed sense of rotation of the driven magnet. For other angles the experiments reveal…
▽ More
The coupling of two rotating spherical magnets is investigated experimentally. For two specific angles between the input and output rotation axes, a cogging-free coupling is observed, where the driven magnet is phase-locked to the driving one. The striking difference between these two modes of operation is the reversed sense of rotation of the driven magnet. For other angles the experiments reveal a more complex dynamical behavior, which is divided in three different classes. This is done by analysing the deviation from a periodic motion of the driven magnet, and by measuring the total harmonic distortion of this rotation. The experimental results can be understood by a mathematical model based on pure dipole-dipole interaction, with the addition of adequate friction terms.
△ Less
Submitted 13 August, 2020;
originally announced August 2020.
-
Graphical Magnetogranulometry of EMG909
Authors:
Ingo Rehberg,
Reinhard Richter,
Stefan Hartung
Abstract:
The magnetization curve of the ferrofluid EMG909 is measured. It can adequately be described by a superposition of four Langevin terms. The effective dipole strength of the magnetic particles in this fluid is subsequently obtained by a graphical rectification of the magnetization curve based on the inverse Langevin function. The method yields the arithmetic and the harmonic mean of the magnetic mo…
▽ More
The magnetization curve of the ferrofluid EMG909 is measured. It can adequately be described by a superposition of four Langevin terms. The effective dipole strength of the magnetic particles in this fluid is subsequently obtained by a graphical rectification of the magnetization curve based on the inverse Langevin function. The method yields the arithmetic and the harmonic mean of the magnetic moment distribution function, and a guess for the geometric mean and the relative standard deviation. It has the advantage that it does not require a prejudiced guess of the distribution function of the poly-disperse suspension of magnetic particles.
△ Less
Submitted 15 April, 2020; v1 submitted 30 July, 2019;
originally announced July 2019.
-
Measuring magnetic moments of polydisperse ferrofluids utilizing the inverse Langevin function
Authors:
Ingo Rehberg,
Reinhard Richter,
Stefan Hartung,
Niklas Lucht,
Birgit Hankiewicz,
Thomas Friedrich
Abstract:
The dipole strength of magnetic particles in a suspension is obtained by a graphical rectification of the magnetization curves based on the inverse Langevin function. The method yields the arithmetic and the harmonic mean of the particle distribution. It has an advantage compared to the fitting of magnetization curves to some appropriate mathematical model: It does not rely on assuming a particula…
▽ More
The dipole strength of magnetic particles in a suspension is obtained by a graphical rectification of the magnetization curves based on the inverse Langevin function. The method yields the arithmetic and the harmonic mean of the particle distribution. It has an advantage compared to the fitting of magnetization curves to some appropriate mathematical model: It does not rely on assuming a particular distribution function of the particles.
△ Less
Submitted 24 July, 2019;
originally announced July 2019.
-
Dotriacontapoles -- almost self-assembled
Authors:
Stefan Hartung,
Felix Sommer,
Simeon Völkel,
Johannes Schönke,
Ingo Rehberg
Abstract:
The magnetic field of a cuboidal cluster of eight magnetic spheres is measured. It decays with the inverse seventh power of the distance. This corresponds formally to a hitherto unheard-of multipole, namely a dotriacontapole. This strong decay is explained on the basis of dipole-dipole interaction and the symmetry of the ensuing ground state of the cuboidal cluster. A method to build such dotriaco…
▽ More
The magnetic field of a cuboidal cluster of eight magnetic spheres is measured. It decays with the inverse seventh power of the distance. This corresponds formally to a hitherto unheard-of multipole, namely a dotriacontapole. This strong decay is explained on the basis of dipole-dipole interaction and the symmetry of the ensuing ground state of the cuboidal cluster. A method to build such dotriacontapoles is provided.
△ Less
Submitted 11 September, 2018;
originally announced September 2018.
-
Fixed-Parameter Algorithms for DAG Partitioning
Authors:
René van Bevern,
Robert Bredereck,
Morgan Chopin,
Sepp Hartung,
Falk Hüffner,
André Nichterlein,
Ondřej Suchý
Abstract:
Finding the origin of short phrases propagating through the web has been formalized by Leskovec et al. [ACM SIGKDD 2009] as DAG Partitioning: given an arc-weighted directed acyclic graph on $n$ vertices and $m$ arcs, delete arcs with total weight at most $k$ such that each resulting weakly-connected component contains exactly one sink---a vertex without outgoing arcs. DAG Partitioning is NP-hard.…
▽ More
Finding the origin of short phrases propagating through the web has been formalized by Leskovec et al. [ACM SIGKDD 2009] as DAG Partitioning: given an arc-weighted directed acyclic graph on $n$ vertices and $m$ arcs, delete arcs with total weight at most $k$ such that each resulting weakly-connected component contains exactly one sink---a vertex without outgoing arcs. DAG Partitioning is NP-hard.
We show an algorithm to solve DAG Partitioning in $O(2^k \cdot (n+m))$ time, that is, in linear time for fixed $k$. We complement it with linear-time executable data reduction rules. Our experiments show that, in combination, they can optimally solve DAG Partitioning on simulated citation networks within five minutes for $k\leq190$ and $m$ being $10^7$ and larger. We use our obtained optimal solutions to evaluate the solution quality of Leskovec et al.'s heuristic.
We show that Leskovec et al.'s heuristic works optimally on trees and generalize this result by showing that DAG Partitioning is solvable in $2^{O(w^2)}\cdot n$ time if a width-$w$ tree decomposition of the input graph is given. Thus, we improve an algorithm and answer an open question of Alamdari and Mehrabian [WAW 2012].
We complement our algorithms by lower bounds on the running time of exact algorithms and on the effectivity of data reduction.
△ Less
Submitted 27 November, 2016;
originally announced November 2016.
-
Co-Clustering Under the Maximum Norm
Authors:
Laurent Bulteau,
Vincent Froese,
Sepp Hartung,
Rolf Niedermeier
Abstract:
Co-clustering, that is, partitioning a numerical matrix into homogeneous submatrices, has many applications ranging from bioinformatics to election analysis. Many interesting variants of co-clustering are NP-hard. We focus on the basic variant of co-clustering where the homogeneity of a submatrix is defined in terms of minimizing the maximum distance between two entries. In this context, we spot s…
▽ More
Co-clustering, that is, partitioning a numerical matrix into homogeneous submatrices, has many applications ranging from bioinformatics to election analysis. Many interesting variants of co-clustering are NP-hard. We focus on the basic variant of co-clustering where the homogeneity of a submatrix is defined in terms of minimizing the maximum distance between two entries. In this context, we spot several NP-hard as well as a number of relevant polynomial-time solvable special cases, thus charting the border of tractability for this challenging data clustering problem. For instance, we provide polynomial-time solvability when having to partition the rows and columns into two subsets each (meaning that one obtains four submatrices). When partitioning rows and columns into three subsets each, however, we encounter NP-hardness even for input matrices containing only values from {0, 1, 2}.
△ Less
Submitted 14 June, 2019; v1 submitted 17 December, 2015;
originally announced December 2015.
-
Constant-factor approximations for Capacitated Arc Routing without triangle inequality
Authors:
René van Bevern,
Sepp Hartung,
André Nichterlein,
Manuel Sorge
Abstract:
Given an undirected graph with edge costs and edge demands, the Capacitated Arc Routing problem (CARP) asks for minimum-cost routes for equal-capacity vehicles so as to satisfy all demands. Constant-factor polynomial-time approximation algorithms were proposed for CARP with triangle inequality, while CARP was claimed to be NP-hard to approximate within any constant factor in general. Correcting th…
▽ More
Given an undirected graph with edge costs and edge demands, the Capacitated Arc Routing problem (CARP) asks for minimum-cost routes for equal-capacity vehicles so as to satisfy all demands. Constant-factor polynomial-time approximation algorithms were proposed for CARP with triangle inequality, while CARP was claimed to be NP-hard to approximate within any constant factor in general. Correcting this claim, we show that any factor α approximation for CARP with triangle inequality yields a factor α approximation for the general CARP.
△ Less
Submitted 14 April, 2014;
originally announced April 2014.
-
Improved Upper and Lower Bound Heuristics for Degree Anonymization in Social Networks
Authors:
Sepp Hartung,
Clemens Hoffmann,
André Nichterlein
Abstract:
Motivated by a strongly growing interest in anonymizing social network data, we investigate the NP-hard Degree Anonymization problem: given an undirected graph, the task is to add a minimum number of edges such that the graph becomes k-anonymous. That is, for each vertex there have to be at least k-1 other vertices of exactly the same degree. The model of degree anonymization has been introduced b…
▽ More
Motivated by a strongly growing interest in anonymizing social network data, we investigate the NP-hard Degree Anonymization problem: given an undirected graph, the task is to add a minimum number of edges such that the graph becomes k-anonymous. That is, for each vertex there have to be at least k-1 other vertices of exactly the same degree. The model of degree anonymization has been introduced by Liu and Terzi [ACM SIGMOD'08], who also proposed and evaluated a two-phase heuristic. We present an enhancement of this heuristic, including new algorithms for each phase which significantly improve on the previously known theoretical and practical running times. Moreover, our algorithms are optimized for large-scale social networks and provide upper and lower bounds for the optimal solution. Notably, on about 26 % of the real-world data we provide (provably) optimal solutions; whereas in the other cases our upper bounds significantly improve on known heuristic solutions.
△ Less
Submitted 25 February, 2014;
originally announced February 2014.
-
On Structural Parameterizations for the 2-Club Problem
Authors:
Sepp Hartung,
Christian Komusiewicz,
André Nichterlein,
Ondrej Suchý
Abstract:
The NP-hard 2-Club problem is, given an undirected graph G=(V,E) and l\in N, to decide whether there is a vertex set S\subseteq V of size at least l such that the induced subgraph G[S] has diameter at most two. We make progress towards a systematic classification of the complexity of 2-Club with respect to a hierarchy of prominent structural graph parameters. First, we present the following tight…
▽ More
The NP-hard 2-Club problem is, given an undirected graph G=(V,E) and l\in N, to decide whether there is a vertex set S\subseteq V of size at least l such that the induced subgraph G[S] has diameter at most two. We make progress towards a systematic classification of the complexity of 2-Club with respect to a hierarchy of prominent structural graph parameters. First, we present the following tight NP-hardness results: 2-Club is NP-hard on graphs that become bipartite by deleting one vertex, on graphs that can be covered by three cliques, and on graphs with domination number two and diameter three. Then, we consider the parameter h-index of the input graph. This parameter is motivated by real-world instances and the fact that 2-Club is fixed-parameter tractable with respect to the larger parameter maximum degree. We present an algorithm that solves 2-Club in |V|^{f(k)} time with k being the h-index. By showing W[1]-hardness for this parameter, we provide evidence that the above algorithm cannot be improved to a fixed-parameter algorithm. Furthermore, the reduction used for this hardness result can be modified to show that 2-Club is NP-hard if the input graph has constant degeneracy. Finally, we show that 2-Club is fixed-parameter tractable with respect to distance to cographs.
△ Less
Submitted 16 May, 2013;
originally announced May 2013.
-
Image Subtraction Noise Reduction Using Point Spread Function Cross-correlation
Authors:
Steven Hartung
Abstract:
Image subtraction in astronomy is a tool for transient object discovery and characterization, particularly useful in wide fields, and is well suited for moving or photometrically varying objects such as asteroids, extra-solar planets and supernovae. A convolution technique is used to match point spread functions (PSFs) between images of the same field taken at different times prior to pixel-by-pix…
▽ More
Image subtraction in astronomy is a tool for transient object discovery and characterization, particularly useful in wide fields, and is well suited for moving or photometrically varying objects such as asteroids, extra-solar planets and supernovae. A convolution technique is used to match point spread functions (PSFs) between images of the same field taken at different times prior to pixel-by-pixel subtraction. Particularly suitable for large-scale images is a spatially-varying kernel, where the convolution is allowed to adapt to PSF changes as a function of position within the images. The most versatile basis for fitting the spatially-varying kernel is the Dirac delta function. However, the convolution kernel based on the delta function does not discriminate between pixel scale noise variations and the intended stellar point spread function signals. The situation can frequently lead to reduced signal to noise ratios for variable objects detectable in the resulting subtraction. This work presents a cross-correlation method for reducing noise effects on the delta function derived convolution kernels, thus yielding significantly improved signal to noise in the resulting subtraction.
△ Less
Submitted 8 January, 2013;
originally announced January 2013.
-
On the Parameterized and Approximation Hardness of Metric Dimension
Authors:
Sepp Hartung,
André Nichterlein
Abstract:
The NP-hard Metric Dimension problem is to decide for a given graph G and a positive integer k whether there is a vertex subset of size at most k that separates all vertex pairs in G. Herein, a vertex v separates a pair {u,w} if the distance (length of a shortest path) between v and u is different from the distance of v and w. We give a polynomial-time computable reduction from the Bipartite Domin…
▽ More
The NP-hard Metric Dimension problem is to decide for a given graph G and a positive integer k whether there is a vertex subset of size at most k that separates all vertex pairs in G. Herein, a vertex v separates a pair {u,w} if the distance (length of a shortest path) between v and u is different from the distance of v and w. We give a polynomial-time computable reduction from the Bipartite Dominating Set problem to Metric Dimension on maximum degree three graphs such that there is a one-to-one correspondence between the solution sets of both problems. There are two main consequences of this: First, it proves that Metric Dimension on maximum degree three graphs is W[2]-complete with respect to the parameter k. This answers an open question concerning the parameterized complexity of Metric Dimension posed by Díaz et al. [ESA'12] and already mentioned by Lokshtanov [Dagstuhl seminar, 2009]. Additionally, it implies that Metric Dimension cannot be solved in n^{o(k)} time, unless the assumption FPT \neq W[1] fails. This proves that a trivial n^{O(k)} algorithm is probably asymptotically optimal.
Second, as Bipartite Dominating Set is inapproximable within o(log n), it follows that Metric Dimension on maximum degree three graphs is also inapproximable by a factor of o(log n), unless NP=P. This strengthens the result of Hauptmann et al. [JDA 2012] who proved APX-hardness on bounded-degree graphs.
△ Less
Submitted 7 November, 2012;
originally announced November 2012.
-
GPU Acceleration of Image Convolution using Spatially-varying Kernel
Authors:
Steven Hartung,
Hemant Shukla,
J. Patrick Miller,
Carlton Pennypacker
Abstract:
Image subtraction in astronomy is a tool for transient object discovery such as asteroids, extra-solar planets and supernovae. To match point spread functions (PSFs) between images of the same field taken at different times a convolution technique is used. Particularly suitable for large-scale images is a computationally intensive spatially-varying kernel. The underlying algorithm is inherently ma…
▽ More
Image subtraction in astronomy is a tool for transient object discovery such as asteroids, extra-solar planets and supernovae. To match point spread functions (PSFs) between images of the same field taken at different times a convolution technique is used. Particularly suitable for large-scale images is a computationally intensive spatially-varying kernel. The underlying algorithm is inherently massively parallel due to unique kernel generation at every pixel location. The spatially-varying kernel cannot be efficiently computed through the Convolution Theorem, and thus does not lend itself to acceleration by Fast Fourier Transform (FFT). This work presents results of accelerated implementation of the spatially-varying kernel image convolution in multi-cores with OpenMP and graphic processing units (GPUs). Typical speedups over ANSI-C were a factor of 50 and a factor of 1000 over the initial IDL implementation, demonstrating that the techniques are a practical and high impact path to terabyte-per-night image pipelines and petascale processing.
△ Less
Submitted 25 September, 2012;
originally announced September 2012.
-
NP-Hardness and Fixed-Parameter Tractability of Realizing Degree Sequences with Directed Acyclic Graphs
Authors:
Sepp Hartung,
André Nichterlein
Abstract:
In graph realization problems one is given a degree sequence and the task is to decide whether there is a graph whose vertex degrees match to the given sequence. This realization problem is known to be polynomial-time solvable when the graph is directed or undirected. In contrary, we show NP-completeness for the problem of realizing a given sequence of pairs of positive integers (representing inde…
▽ More
In graph realization problems one is given a degree sequence and the task is to decide whether there is a graph whose vertex degrees match to the given sequence. This realization problem is known to be polynomial-time solvable when the graph is directed or undirected. In contrary, we show NP-completeness for the problem of realizing a given sequence of pairs of positive integers (representing indegrees and outdegrees) with a directed acyclic graph, answering an open question of Berger and Müller-Hannemann [FCT 2011]. Furthermore, we classify the problem as fixed-parameter tractable with respect to the parameter "maximum degree".
△ Less
Submitted 17 January, 2012; v1 submitted 7 October, 2011;
originally announced October 2011.