@article{7421, abstract = {X and Y chromosomes can diverge when rearrangements block recombination between them. Here we present the first genomic view of a reciprocal translocation that causes two physically unconnected pairs of chromosomes to be coinherited as sex chromosomes. In a population of the common frog (Rana temporaria), both pairs of X and Y chromosomes show extensive sequence differentiation, but not degeneration of the Y chromosomes. A new method based on gene trees shows both chromosomes are sex‐linked. Furthermore, the gene trees from the two Y chromosomes have identical topologies, showing they have been coinherited since the reciprocal translocation occurred. Reciprocal translocations can thus reshape sex linkage on a much greater scale compared with inversions, the type of rearrangement that is much better known in sex chromosome evolution, and they can greatly amplify the power of sexually antagonistic selection to drive genomic rearrangement. Two more populations show evidence of other rearrangements, suggesting that this species has unprecedented structural polymorphism in its sex chromosomes.}, author = {Toups, Melissa A and Rodrigues, Nicolas and Perrin, Nicolas and Kirkpatrick, Mark}, issn = {1365-294X}, journal = {Molecular Ecology}, number = {8}, pages = {1877--1889}, publisher = {Wiley}, title = {{A reciprocal translocation radically reshapes sex‐linked inheritance in the common frog}}, doi = {10.1111/mec.14990}, volume = {28}, year = {2019}, } @article{7422, abstract = {Biochemical reactions often occur at low copy numbers but at once in crowded and diverse environments. Space and stochasticity therefore play an essential role in biochemical networks. Spatial-stochastic simulations have become a prominent tool for understanding how stochasticity at the microscopic level influences the macroscopic behavior of such systems. While particle-based models guarantee the level of detail necessary to accurately describe the microscopic dynamics at very low copy numbers, the algorithms used to simulate them typically imply trade-offs between computational efficiency and biochemical accuracy. eGFRD (enhanced Green’s Function Reaction Dynamics) is an exact algorithm that evades such trade-offs by partitioning the N-particle system into M ≤ N analytically tractable one- and two-particle systems; the analytical solutions (Green’s functions) then are used to implement an event-driven particle-based scheme that allows particles to make large jumps in time and space while retaining access to their state variables at arbitrary simulation times. Here we present “eGFRD2,” a new eGFRD version that implements the principle of eGFRD in all dimensions, thus enabling efficient particle-based simulation of biochemical reaction-diffusion processes in the 3D cytoplasm, on 2D planes representing membranes, and on 1D elongated cylinders representative of, e.g., cytoskeletal tracks or DNA; in 1D, it also incorporates convective motion used to model active transport. We find that, for low particle densities, eGFRD2 is up to 6 orders of magnitude faster than conventional Brownian dynamics. We exemplify the capabilities of eGFRD2 by simulating an idealized model of Pom1 gradient formation, which involves 3D diffusion, active transport on microtubules, and autophosphorylation on the membrane, confirming recent experimental and theoretical results on this system to hold under genuinely stochastic conditions.}, author = {Sokolowski, Thomas R and Paijmans, Joris and Bossen, Laurens and Miedema, Thomas and Wehrens, Martijn and Becker, Nils B. and Kaizu, Kazunari and Takahashi, Koichi and Dogterom, Marileen and ten Wolde, Pieter Rein}, issn = {1089-7690}, journal = {The Journal of Chemical Physics}, number = {5}, publisher = {AIP Publishing}, title = {{eGFRD in all dimensions}}, doi = {10.1063/1.5064867}, volume = {150}, year = {2019}, } @article{7423, abstract = {We compare finite rank perturbations of the following three ensembles of complex rectangular random matrices: First, a generalised Wishart ensemble with one random and two fixed correlation matrices introduced by Borodin and Péché, second, the product of two independent random matrices where one has correlated entries, and third, the case when the two random matrices become also coupled through a fixed matrix. The singular value statistics of all three ensembles is shown to be determinantal and we derive double contour integral representations for their respective kernels. Three different kernels are found in the limit of infinite matrix dimension at the origin of the spectrum. They depend on finite rank perturbations of the correlation and coupling matrices and are shown to be integrable. The first kernel (I) is found for two independent matrices from the second, and two weakly coupled matrices from the third ensemble. It generalises the Meijer G-kernel for two independent and uncorrelated matrices. The third kernel (III) is obtained for the generalised Wishart ensemble and for two strongly coupled matrices. It further generalises the perturbed Bessel kernel of Desrosiers and Forrester. Finally, kernel (II), found for the ensemble of two coupled matrices, provides an interpolation between the kernels (I) and (III), generalising previous findings of part of the authors.}, author = {Akemann, Gernot and Checinski, Tomasz and Liu, Dangzheng and Strahov, Eugene}, issn = {0246-0203}, journal = {Annales de l'Institut Henri Poincaré, Probabilités et Statistiques}, number = {1}, pages = {441--479}, publisher = {Institute of Mathematical Statistics}, title = {{Finite rank perturbations in products of coupled random matrices: From one correlated to two Wishart ensembles}}, doi = {10.1214/18-aihp888}, volume = {55}, year = {2019}, } @article{7436, abstract = {For an ordinary K3 surface over an algebraically closed field of positive characteristic we show that every automorphism lifts to characteristic zero. Moreover, we show that the Fourier-Mukai partners of an ordinary K3 surface are in one-to-one correspondence with the Fourier-Mukai partners of the geometric generic fiber of its canonical lift. We also prove that the explicit counting formula for Fourier-Mukai partners of the K3 surfaces with Picard rank two and with discriminant equal to minus of a prime number, in terms of the class number of the prime, holds over a field of positive characteristic as well. We show that the image of the derived autoequivalence group of a K3 surface of finite height in the group of isometries of its crystalline cohomology has index at least two. Moreover, we provide a conditional upper bound on the kernel of this natural cohomological descent map. Further, we give an extended remark in the appendix on the possibility of an F-crystal structure on the crystalline cohomology of a K3 surface over an algebraically closed field of positive characteristic and show that the naive F-crystal structure fails in being compatible with inner product. }, author = {Srivastava, Tanya K}, issn = {1431-0643}, journal = {Documenta Mathematica}, pages = {1135--1177}, publisher = {EMS Press}, title = {{On derived equivalences of k3 surfaces in positive characteristic}}, doi = {10.25537/dm.2019v24.1135-1177}, volume = {24}, year = {2019}, } @inproceedings{7437, abstract = {Most of today's distributed machine learning systems assume reliable networks: whenever two machines exchange information (e.g., gradients or models), the network should guarantee the delivery of the message. At the same time, recent work exhibits the impressive tolerance of machine learning algorithms to errors or noise arising from relaxed communication or synchronization. In this paper, we connect these two trends, and consider the following question: Can we design machine learning systems that are tolerant to network unreliability during training? With this motivation, we focus on a theoretical problem of independent interest-given a standard distributed parameter server architecture, if every communication between the worker and the server has a non-zero probability p of being dropped, does there exist an algorithm that still converges, and at what speed? The technical contribution of this paper is a novel theoretical analysis proving that distributed learning over unreliable network can achieve comparable convergence rate to centralized or distributed learning over reliable networks. Further, we prove that the influence of the packet drop rate diminishes with the growth of the number of parameter servers. We map this theoretical result onto a real-world scenario, training deep neural networks over an unreliable network layer, and conduct network simulation to validate the system improvement by allowing the networks to be unreliable.}, author = {Yu, Chen and Tang, Hanlin and Renggli, Cedric and Kassing, Simon and Singla, Ankit and Alistarh, Dan-Adrian and Zhang, Ce and Liu, Ji}, booktitle = {36th International Conference on Machine Learning, ICML 2019}, isbn = {9781510886988}, location = {Long Beach, CA, United States}, pages = {12481--12512}, publisher = {IMLS}, title = {{Distributed learning over unreliable networks}}, volume = {2019-June}, year = {2019}, } @article{7451, abstract = {We prove that the observable telegraph signal accompanying the bistability in the photon-blockade-breakdown regime of the driven and lossy Jaynes–Cummings model is the finite-size precursor of what in the thermodynamic limit is a genuine first-order phase transition. We construct a finite-size scaling of the system parameters to a well-defined thermodynamic limit, in which the system remains the same microscopic system, but the telegraph signal becomes macroscopic both in its timescale and intensity. The existence of such a finite-size scaling completes and justifies the classification of the photon-blockade-breakdown effect as a first-order dissipative quantum phase transition.}, author = {Vukics, A. and Dombi, A. and Fink, Johannes M and Domokos, P.}, issn = {2521-327X}, journal = {Quantum}, publisher = {Verein zur Förderung des Open Access Publizierens in den Quantenwissenschaften}, title = {{Finite-size scaling of the photon-blockade breakdown dissipative quantum phase transition}}, doi = {10.22331/q-2019-06-03-150}, volume = {3}, year = {2019}, } @inbook{7453, abstract = {We illustrate the ingredients of the state-of-the-art of model-based approach for the formal design and verification of cyber-physical systems. To capture the interaction between a discrete controller and its continuously evolving environment, we use the formal models of timed and hybrid automata. We explain the steps of modeling and verification in the tools Uppaal and SpaceEx using a case study based on a dual-chamber implantable pacemaker monitoring a human heart. We show how to design a model as a composition of components, how to construct models at varying levels of detail, how to establish that one model is an abstraction of another, how to specify correctness requirements using temporal logic, and how to verify that a model satisfies a logical requirement.}, author = {Alur, Rajeev and Giacobbe, Mirco and Henzinger, Thomas A and Larsen, Kim G. and Mikučionis, Marius}, booktitle = {Computing and Software Science}, editor = {Steffen, Bernhard and Woeginger, Gerhard}, isbn = {9783319919072}, issn = {0302-9743}, pages = {452--477}, publisher = {Springer Nature}, title = {{Continuous-time models for system design and analysis}}, doi = {10.1007/978-3-319-91908-9_22}, volume = {10000}, year = {2019}, } @inproceedings{7468, abstract = {We present a new proximal bundle method for Maximum-A-Posteriori (MAP) inference in structured energy minimization problems. The method optimizes a Lagrangean relaxation of the original energy minimization problem using a multi plane block-coordinate Frank-Wolfe method that takes advantage of the specific structure of the Lagrangean decomposition. We show empirically that our method outperforms state-of-the-art Lagrangean decomposition based algorithms on some challenging Markov Random Field, multi-label discrete tomography and graph matching problems.}, author = {Swoboda, Paul and Kolmogorov, Vladimir}, booktitle = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition}, isbn = {9781728132938}, issn = {10636919}, location = {Long Beach, CA, United States}, publisher = {IEEE}, title = {{Map inference via block-coordinate Frank-Wolfe algorithm}}, doi = {10.1109/CVPR.2019.01140}, volume = {2019-June}, year = {2019}, } @inproceedings{7479, abstract = {Multi-exit architectures, in which a stack of processing layers is interleaved with early output layers, allow the processing of a test example to stop early and thus save computation time and/or energy. In this work, we propose a new training procedure for multi-exit architectures based on the principle of knowledge distillation. The method encourage searly exits to mimic later, more accurate exits, by matching their output probabilities. Experiments on CIFAR100 and ImageNet show that distillation-based training significantly improves the accuracy of early exits while maintaining state-of-the-art accuracy for late ones. The method is particularly beneficial when training data is limited and it allows a straightforward extension to semi-supervised learning,i.e. making use of unlabeled data at training time. Moreover, it takes only afew lines to implement and incurs almost no computational overhead at training time, and none at all at test time.}, author = {Bui Thi Mai, Phuong and Lampert, Christoph}, booktitle = {IEEE International Conference on Computer Vision}, isbn = {9781728148038}, issn = {15505499}, location = {Seoul, Korea}, pages = {1355--1364}, publisher = {IEEE}, title = {{Distillation-based training for multi-exit architectures}}, doi = {10.1109/ICCV.2019.00144}, volume = {2019-October}, year = {2019}, } @inbook{7513, abstract = {Social insects (i.e., ants, termites and the social bees and wasps) protect their colonies from disease using a combination of individual immunity and collectively performed defenses, termed social immunity. The first line of social immune defense is sanitary care, which is performed by colony members to protect their pathogen-exposed nestmates from developing an infection. If sanitary care fails and an infection becomes established, a second line of social immune defense is deployed to stop disease transmission within the colony and to protect the valuable queens, which together with the males are the reproductive individuals of the colony. Insect colonies are separated into these reproductive individuals and the sterile worker force, forming a superorganismal reproductive unit reminiscent of the differentiated germline and soma in a multicellular organism. Ultimately, the social immune response preserves the germline of the superorganism insect colony and increases overall fitness of the colony in case of disease. }, author = {Cremer, Sylvia and Kutzer, Megan}, booktitle = {Encyclopedia of Animal Behavior}, editor = {Choe, Jae}, isbn = {9780128132517}, pages = {747--755}, publisher = {Elsevier}, title = {{Social immunity}}, doi = {10.1016/B978-0-12-809633-8.90721-0}, year = {2019}, } @unpublished{7524, abstract = {We prove a lower bound for the free energy (per unit volume) of the two-dimensional Bose gas in the thermodynamic limit. We show that the free energy at density $\rho$ and inverse temperature $\beta$ differs from the one of the non-interacting system by the correction term $4 \pi \rho^2 |\ln a^2 \rho|^{-1} (2 - [1 - \beta_{\mathrm{c}}/\beta]_+^2)$. Here $a$ is the scattering length of the interaction potential, $[\cdot]_+ = \max\{ 0, \cdot \}$ and $\beta_{\mathrm{c}}$ is the inverse Berezinskii--Kosterlitz--Thouless critical temperature for superfluidity. The result is valid in the dilute limit $a^2\rho \ll 1$ and if $\beta \rho \gtrsim 1$.}, author = {Deuchert, Andreas and Mayer, Simon and Seiringer, Robert}, booktitle = {arXiv:1910.03372}, pages = {61}, publisher = {ArXiv}, title = {{The free energy of the two-dimensional dilute Bose gas. I. Lower bound}}, year = {2019}, } @inproceedings{7542, abstract = {We present a novel class of convolutional neural networks (CNNs) for set functions,i.e., data indexed with the powerset of a finite set. The convolutions are derivedas linear, shift-equivariant functions for various notions of shifts on set functions.The framework is fundamentally different from graph convolutions based on theLaplacian, as it provides not one but several basic shifts, one for each element inthe ground set. Prototypical experiments with several set function classificationtasks on synthetic datasets and on datasets derived from real-world hypergraphsdemonstrate the potential of our new powerset CNNs.}, author = {Wendler, Chris and Alistarh, Dan-Adrian and Püschel, Markus}, issn = {1049-5258}, location = {Vancouver, Canada}, pages = {927--938}, publisher = {Neural Information Processing Systems Foundation}, title = {{Powerset convolutional neural networks}}, volume = {32}, year = {2019}, } @article{7550, abstract = {We consider an optimal control problem for an abstract nonlinear dissipative evolution equation. The differential constraint is penalized by augmenting the target functional by a nonnegative global-in-time functional which is null-minimized in the evolution equation is satisfied. Different variational settings are presented, leading to the convergence of the penalization method for gradient flows, noncyclic and semimonotone flows, doubly nonlinear evolutions, and GENERIC systems. }, author = {Portinale, Lorenzo and Stefanelli, Ulisse}, issn = {1343-4373}, journal = {Advances in Mathematical Sciences and Applications}, number = {2}, pages = {425--447}, publisher = {Gakko Tosho}, title = {{Penalization via global functionals of optimal-control problems for dissipative evolution}}, volume = {28}, year = {2019}, } @unpublished{7552, abstract = {There is increasing evidence that protein binding to specific sites along DNA can activate the reading out of genetic information without coming into direct physical contact with the gene. There also is evidence that these distant but interacting sites are embedded in a liquid droplet of proteins which condenses out of the surrounding solution. We argue that droplet-mediated interactions can account for crucial features of gene regulation only if the droplet is poised at a non-generic point in its phase diagram. We explore a minimal model that embodies this idea, show that this model has a natural mechanism for self-tuning, and suggest direct experimental tests. }, author = {Bialek, William and Gregor, Thomas and Tkačik, Gašper}, booktitle = {arXiv:1912.08579}, pages = {5}, publisher = {ArXiv}, title = {{Action at a distance in transcriptional regulation}}, year = {2019}, } @inproceedings{7576, abstract = {We present the results of a friendly competition for formal verification of continuous and hybrid systems with nonlinear continuous dynamics. The friendly competition took place as part of the workshop Applied Verification for Continuous and Hybrid Systems (ARCH) in 2019. In this year, 6 tools Ariadne, CORA, DynIbex, Flow*, Isabelle/HOL, and JuliaReach (in alphabetic order) participated. They are applied to solve reachability analysis problems on four benchmark problems, one of them with hybrid dynamics. We do not rank the tools based on the results, but show the current status and discover the potential advantages of different tools.}, author = {Immler, Fabian and Althoff, Matthias and Benet, Luis and Chapoutot, Alexandre and Chen, Xin and Forets, Marcelo and Geretti, Luca and Kochdumper, Niklas and Sanders, David P. and Schilling, Christian}, booktitle = {EPiC Series in Computing}, issn = {23987340}, location = {Montreal, Canada}, pages = {41--61}, publisher = {EasyChair Publications}, title = {{ARCH-COMP19 Category Report: Continuous and hybrid systems with nonlinear dynamics}}, doi = {10.29007/m75b}, volume = {61}, year = {2019}, } @inproceedings{7606, abstract = {We derive a tight lower bound on equivocation (conditional entropy), or equivalently a tight upper bound on mutual information between a signal variable and channel outputs. The bound is in terms of the joint distribution of the signals and maximum a posteriori decodes (most probable signals given channel output). As part of our derivation, we describe the key properties of the distribution of signals, channel outputs and decodes, that minimizes equivocation and maximizes mutual information. This work addresses a problem in data analysis, where mutual information between signals and decodes is sometimes used to lower bound the mutual information between signals and channel outputs. Our result provides a corresponding upper bound.}, author = {Hledik, Michal and Sokolowski, Thomas R and Tkačik, Gašper}, booktitle = {IEEE Information Theory Workshop, ITW 2019}, isbn = {9781538669006}, location = {Visby, Sweden}, publisher = {IEEE}, title = {{A tight upper bound on mutual information}}, doi = {10.1109/ITW44776.2019.8989292}, year = {2019}, } @inproceedings{7639, abstract = {Deep neural networks (DNNs) have become increasingly important due to their excellent empirical performance on a wide range of problems. However, regularization is generally achieved by indirect means, largely due to the complex set of functions defined by a network and the difficulty in measuring function complexity. There exists no method in the literature for additive regularization based on a norm of the function, as is classically considered in statistical learning theory. In this work, we study the tractability of function norms for deep neural networks with ReLU activations. We provide, to the best of our knowledge, the first proof in the literature of the NP-hardness of computing function norms of DNNs of 3 or more layers. We also highlight a fundamental difference between shallow and deep networks. In the light on these results, we propose a new regularization strategy based on approximate function norms, and show its efficiency on a segmentation task with a DNN.}, author = {Rannen-Triki, Amal and Berman, Maxim and Kolmogorov, Vladimir and Blaschko, Matthew B.}, booktitle = {Proceedings of the 2019 International Conference on Computer Vision Workshop}, isbn = {9781728150239}, location = {Seoul, South Korea}, publisher = {IEEE}, title = {{Function norms for neural networks}}, doi = {10.1109/ICCVW.2019.00097}, year = {2019}, } @inproceedings{7640, abstract = {We propose a new model for detecting visual relationships, such as "person riding motorcycle" or "bottle on table". This task is an important step towards comprehensive structured mage understanding, going beyond detecting individual objects. Our main novelty is a Box Attention mechanism that allows to model pairwise interactions between objects using standard object detection pipelines. The resulting model is conceptually clean, expressive and relies on well-justified training and prediction procedures. Moreover, unlike previously proposed approaches, our model does not introduce any additional complex components or hyperparameters on top of those already required by the underlying detection model. We conduct an experimental evaluation on two datasets, V-COCO and Open Images, demonstrating strong quantitative and qualitative results.}, author = {Kolesnikov, Alexander and Kuznetsova, Alina and Lampert, Christoph and Ferrari, Vittorio}, booktitle = {Proceedings of the 2019 International Conference on Computer Vision Workshop}, isbn = {9781728150239}, location = {Seoul, South Korea}, publisher = {IEEE}, title = {{Detecting visual relationships using box attention}}, doi = {10.1109/ICCVW.2019.00217}, year = {2019}, } @article{5, abstract = {In this paper, we introduce a quantum version of the wonderful compactification of a group as a certain noncommutative projective scheme. Our approach stems from the fact that the wonderful compactification encodes the asymptotics of matrix coefficients, and from its realization as a GIT quotient of the Vinberg semigroup. In order to define the wonderful compactification for a quantum group, we adopt a generalized formalism of Proj categories in the spirit of Artin and Zhang. Key to our construction is a quantum version of the Vinberg semigroup, which we define as a q-deformation of a certain Rees algebra, compatible with a standard Poisson structure. Furthermore, we discuss quantum analogues of the stratification of the wonderful compactification by orbits for a certain group action, and provide explicit computations in the case of SL2.}, author = {Ganev, Iordan V}, journal = {Journal of the London Mathematical Society}, number = {3}, pages = {778--806}, publisher = {Wiley}, title = {{The wonderful compactification for quantum groups}}, doi = {10.1112/jlms.12193}, volume = {99}, year = {2019}, } @article{5678, abstract = {The order-k Voronoi tessellation of a locally finite set 𝑋⊆ℝ𝑛 decomposes ℝ𝑛 into convex domains whose points have the same k nearest neighbors in X. Assuming X is a stationary Poisson point process, we give explicit formulas for the expected number and total area of faces of a given dimension per unit volume of space. We also develop a relaxed version of discrete Morse theory and generalize by counting only faces, for which the k nearest points in X are within a given distance threshold.}, author = {Edelsbrunner, Herbert and Nikitenko, Anton}, issn = {14320444}, journal = {Discrete and Computational Geometry}, number = {4}, pages = {865–878}, publisher = {Springer}, title = {{Poisson–Delaunay Mosaics of Order k}}, doi = {10.1007/s00454-018-0049-2}, volume = {62}, year = {2019}, }