TY - THES AB - Decades of studies have revealed the mechanisms of gene regulation in molecular detail. We make use of such well-described regulatory systems to explore how the molecular mechanisms of protein-protein and protein-DNA interactions shape the dynamics and evolution of gene regulation. i) We uncover how the biophysics of protein-DNA binding determines the potential of regulatory networks to evolve and adapt, which can be captured using a simple mathematical model. ii) The evolution of regulatory connections can lead to a significant amount of crosstalk between binding proteins. We explore the effect of crosstalk on gene expression from a target promoter, which seems to be modulated through binding competition at non-specific DNA sites. iii) We investigate how the very same biophysical characteristics as in i) can generate significant fitness costs for cells through global crosstalk, meaning non-specific DNA binding across the genomic background. iv) Binding competition between proteins at a target promoter is a prevailing regulatory feature due to the prevalence of co-regulation at bacterial promoters. However, the dynamics of these systems are not always straightforward to determine even if the molecular mechanisms of regulation are known. A detailed model of the biophysical interactions reveals that interference between the regulatory proteins can constitute a new, generic form of system memory that records the history of the input signals at the promoter. We demonstrate how the biophysics of protein-DNA binding can be harnessed to investigate the principles that shape and ultimately limit cellular gene regulation. These results provide a basis for studies of higher-level functionality, which arises from the underlying regulation. AU - Igler, Claudia ID - 6371 KW - gene regulation KW - biophysics KW - transcription factor binding KW - bacteria SN - 2663-337X TI - On the nature of gene regulatory design - The biophysics of transcription factor binding shapes gene regulation ER - TY - JOUR AB - In this paper, we evaluate clock signals generated in ring oscillators and self-timed rings and the way their jitter can be transformed into random numbers. We show that counting the periods of the jittery clock signal produces random numbers of significantly better quality than the methods in which the jittery signal is simply sampled (the case in almost all current methods). Moreover, we use the counter values to characterize and continuously monitor the source of randomness. However, instead of using the widely used statistical variance, we propose to use Allan variance to do so. There are two main advantages: Allan variance is insensitive to low frequency noises such as flicker noise that are known to be autocorrelated and significantly less circuitry is required for its computation than that used to compute commonly used variance. We also show that it is essential to use a differential principle of randomness extraction from the jitter based on the use of two identical oscillators to avoid autocorrelations originating from external and internal global jitter sources and that this fact is valid for both kinds of rings. Last but not least, we propose a method of statistical testing based on high order Markov model to show the reduced dependencies when the proposed randomness extraction is applied. AU - Allini, Elie Noumon AU - Skórski, Maciej AU - Petura, Oto AU - Bernard, Florent AU - Laban, Marek AU - Fischer, Viktor ID - 10286 IS - 3 JF - IACR Transactions on Cryptographic Hardware and Embedded Systems TI - Evaluation and monitoring of free running oscillators serving as source of randomness VL - 2018 ER - TY - CONF AB - Solving parity games, which are equivalent to modal μ-calculus model checking, is a central algorithmic problem in formal methods, with applications in reactive synthesis, program repair, verification of branching-time properties, etc. Besides the standard compu- tation model with the explicit representation of games, another important theoretical model of computation is that of set-based symbolic algorithms. Set-based symbolic algorithms use basic set operations and one-step predecessor operations on the implicit description of games, rather than the explicit representation. The significance of symbolic algorithms is that they provide scalable algorithms for large finite-state systems, as well as for infinite-state systems with finite quotient. Consider parity games on graphs with n vertices and parity conditions with d priorities. While there is a rich literature of explicit algorithms for parity games, the main results for set-based symbolic algorithms are as follows: (a) the basic algorithm that requires O(nd) symbolic operations and O(d) symbolic space; and (b) an improved algorithm that requires O(nd/3+1) symbolic operations and O(n) symbolic space. In this work, our contributions are as follows: (1) We present a black-box set-based symbolic algorithm based on the explicit progress measure algorithm. Two important consequences of our algorithm are as follows: (a) a set-based symbolic algorithm for parity games that requires quasi-polynomially many symbolic operations and O(n) symbolic space; and (b) any future improvement in progress measure based explicit algorithms immediately imply an efficiency improvement in our set-based symbolic algorithm for parity games. (2) We present a set-based symbolic algorithm that requires quasi-polynomially many symbolic operations and O(d · log n) symbolic space. Moreover, for the important special case of d ≤ log n, our algorithm requires only polynomially many symbolic operations and poly-logarithmic symbolic space. AU - Chatterjee, Krishnendu AU - Dvořák, Wolfgang AU - Henzinger, Monika H AU - Svozil, Alexander ID - 10883 SN - 2398-7340 T2 - 22nd International Conference on Logic for Programming, Artificial Intelligence and Reasoning TI - Quasipolynomial set-based symbolic algorithms for parity games VL - 57 ER - TY - CONF AB - We report on a novel strategy to derive mean-field limits of quantum mechanical systems in which a large number of particles weakly couple to a second-quantized radiation field. The technique combines the method of counting and the coherent state approach to study the growth of the correlations among the particles and in the radiation field. As an instructional example, we derive the Schrödinger–Klein–Gordon system of equations from the Nelson model with ultraviolet cutoff and possibly massless scalar field. In particular, we prove the convergence of the reduced density matrices (of the nonrelativistic particles and the field bosons) associated with the exact time evolution to the projectors onto the solutions of the Schrödinger–Klein–Gordon equations in trace norm. Furthermore, we derive explicit bounds on the rate of convergence of the one-particle reduced density matrix of the nonrelativistic particles in Sobolev norm. AU - Leopold, Nikolai K AU - Pickl, Peter ID - 11 TI - Mean-field limits of particles in interaction with quantised radiation fields VL - 270 ER - TY - JOUR AB - Two generalizations of Itô formula to infinite-dimensional spaces are given. The first one, in Hilbert spaces, extends the classical one by taking advantage of cancellations when they occur in examples and it is applied to the case of a group generator. The second one, based on the previous one and a limit procedure, is an Itô formula in a special class of Banach spaces having a product structure with the noise in a Hilbert component; again the key point is the extension due to a cancellation. This extension to Banach spaces and in particular the specific cancellation are motivated by path-dependent Itô calculus. AU - Flandoli, Franco AU - Russo, Francesco AU - Zanco, Giovanni A ID - 1215 IS - 2 JF - Journal of Theoretical Probability TI - Infinite-dimensional calculus under weak spatial regularity of the processes VL - 31 ER - TY - CONF AB - We resolve in the affirmative conjectures of A. Skopenkov and Repovš (1998), and M. Skopenkov (2003) generalizing the classical Hanani-Tutte theorem to the setting of approximating maps of graphs on 2-dimensional surfaces by embeddings. Our proof of this result is constructive and almost immediately implies an efficient algorithm for testing whether a given piecewise linear map of a graph in a surface is approximable by an embedding. More precisely, an instance of this problem consists of (i) a graph G whose vertices are partitioned into clusters and whose inter-cluster edges are partitioned into bundles, and (ii) a region R of a 2-dimensional compact surface M given as the union of a set of pairwise disjoint discs corresponding to the clusters and a set of pairwise disjoint "pipes" corresponding to the bundles, connecting certain pairs of these discs. We are to decide whether G can be embedded inside M so that the vertices in every cluster are drawn in the corresponding disc, the edges in every bundle pass only through its corresponding pipe, and every edge crosses the boundary of each disc at most once. AU - Fulek, Radoslav AU - Kynčl, Jan ID - 185 SN - 978-3-95977-066-8 TI - Hanani-Tutte for approximating maps of graphs VL - 99 ER - TY - CONF AB - Smallest enclosing spheres of finite point sets are central to methods in topological data analysis. Focusing on Bregman divergences to measure dissimilarity, we prove bounds on the location of the center of a smallest enclosing sphere. These bounds depend on the range of radii for which Bregman balls are convex. AU - Edelsbrunner, Herbert AU - Virk, Ziga AU - Wagner, Hubert ID - 188 TI - Smallest enclosing spheres and Chernoff points in Bregman geometry VL - 99 ER - TY - JOUR AB - A cornerstone of statistical inference, the maximum entropy framework is being increasingly applied to construct descriptive and predictive models of biological systems, especially complex biological networks, from large experimental data sets. Both its broad applicability and the success it obtained in different contexts hinge upon its conceptual simplicity and mathematical soundness. Here we try to concisely review the basic elements of the maximum entropy principle, starting from the notion of ‘entropy’, and describe its usefulness for the analysis of biological systems. As examples, we focus specifically on the problem of reconstructing gene interaction networks from expression data and on recent work attempting to expand our system-level understanding of bacterial metabolism. Finally, we highlight some extensions and potential limitations of the maximum entropy approach, and point to more recent developments that are likely to play a key role in the upcoming challenges of extracting structures and information from increasingly rich, high-throughput biological data. AU - De Martino, Andrea AU - De Martino, Daniele ID - 306 IS - 4 JF - Heliyon TI - An introduction to the maximum entropy approach and its application to inference problems in biology VL - 4 ER - TY - BOOK AB - This book first explores the origins of this idea, grounded in theoretical work on temporal logic and automata. The editors and authors are among the world's leading researchers in this domain, and they contributed 32 chapters representing a thorough view of the development and application of the technique. Topics covered include binary decision diagrams, symbolic model checking, satisfiability modulo theories, partial-order reduction, abstraction, interpolation, concurrency, security protocols, games, probabilistic model checking, and process algebra, and chapters on the transfer of theory to industrial practice, property specification languages for hardware, and verification of real-time systems and hybrid systems. The book will be valuable for researchers and graduate students engaged with the development of formal methods and verification tools. AU - Clarke, Edmund M. AU - Henzinger, Thomas A AU - Veith, Helmut AU - Bloem, Roderick ID - 3300 SN - 978-3-319-10574-1 TI - Handbook of Model Checking ER - TY - CHAP AB - Developmental processes are inherently dynamic and understanding them requires quantitative measurements of gene and protein expression levels in space and time. While live imaging is a powerful approach for obtaining such data, it is still a challenge to apply it over long periods of time to large tissues, such as the embryonic spinal cord in mouse and chick. Nevertheless, dynamics of gene expression and signaling activity patterns in this organ can be studied by collecting tissue sections at different developmental stages. In combination with immunohistochemistry, this allows for measuring the levels of multiple developmental regulators in a quantitative manner with high spatiotemporal resolution. The mean protein expression levels over time, as well as embryo-to-embryo variability can be analyzed. A key aspect of the approach is the ability to compare protein levels across different samples. This requires a number of considerations in sample preparation, imaging and data analysis. Here we present a protocol for obtaining time course data of dorsoventral expression patterns from mouse and chick neural tube in the first 3 days of neural tube development. The described workflow starts from embryo dissection and ends with a processed dataset. Software scripts for data analysis are included. The protocol is adaptable and instructions that allow the user to modify different steps are provided. Thus, the procedure can be altered for analysis of time-lapse images and applied to systems other than the neural tube. AU - Zagórski, Marcin P AU - Kicheva, Anna ID - 37 SN - 1064-3745 T2 - Morphogen Gradients TI - Measuring dorsoventral pattern and morphogen signaling profiles in the growing neural tube VL - 1863 ER - TY - JOUR AB - The hanging-drop network (HDN) is a technology platform based on a completely open microfluidic network at the bottom of an inverted, surface-patterned substrate. The platform is predominantly used for the formation, culturing, and interaction of self-assembled spherical microtissues (spheroids) under precisely controlled flow conditions. Here, we describe design, fabrication, and operation of microfluidic hanging-drop networks. AU - Misun, Patrick AU - Birchler, Axel AU - Lang, Moritz AU - Hierlemann, Andreas AU - Frey, Olivier ID - 305 JF - Methods in Molecular Biology TI - Fabrication and operation of microfluidic hanging drop networks VL - 1771 ER - TY - CONF AB - Probabilistic programs extend classical imperative programs with real-valued random variables and random branching. The most basic liveness property for such programs is the termination property. The qualitative (aka almost-sure) termination problem asks whether a given program program terminates with probability 1. While ranking functions provide a sound and complete method for non-probabilistic programs, the extension of them to probabilistic programs is achieved via ranking supermartingales (RSMs). Although deep theoretical results have been established about RSMs, their application to probabilistic programs with nondeterminism has been limited only to programs of restricted control-flow structure. For non-probabilistic programs, lexicographic ranking functions provide a compositional and practical approach for termination analysis of real-world programs. In this work we introduce lexicographic RSMs and show that they present a sound method for almost-sure termination of probabilistic programs with nondeterminism. We show that lexicographic RSMs provide a tool for compositional reasoning about almost-sure termination, and for probabilistic programs with linear arithmetic they can be synthesized efficiently (in polynomial time). We also show that with additional restrictions even asymptotic bounds on expected termination time can be obtained through lexicographic RSMs. Finally, we present experimental results on benchmarks adapted from previous work to demonstrate the effectiveness of our approach. AU - Agrawal, Sheshansh AU - Chatterjee, Krishnendu AU - Novotny, Petr ID - 325 IS - POPL TI - Lexicographic ranking supermartingales: an efficient approach to termination of probabilistic programs VL - 2 ER - TY - CHAP AB - Adventitious roots (AR) are de novo formed roots that emerge from any part of the plant or from callus in tissue culture, except root tissue. The plant tissue origin and the method by which they are induced determine the physiological properties of emerged ARs. Hence, a standard method encompassing all types of AR does not exist. Here we describe a method for the induction and analysis of AR that emerge from the etiolated hypocotyl of dicot plants. The hypocotyl is formed during embryogenesis and shows a determined developmental pattern which usually does not involve AR formation. However, the hypocotyl shows propensity to form de novo roots under specific circumstances such as removal of the root system, high humidity or flooding, or during de-etiolation. The hypocotyl AR emerge from a pericycle-like cell layer surrounding the vascular tissue of the central cylinder, which is reminiscent to the developmental program of lateral roots. Here we propose an easy protocol for in vitro hypocotyl AR induction from etiolated Arabidopsis seedlings. AU - Trinh, Hoang AU - Verstraeten, Inge AU - Geelen, Danny ID - 408 SN - 1064-3745 T2 - Root Development TI - In vitro assay for induction of adventitious rooting on intact arabidopsis hypocotyls VL - 1761 ER - TY - CHAP AB - Immunolocalization is a valuable tool for cell biology research that allows to rapidly determine the localization and expression levels of endogenous proteins. In plants, whole-mount in situ immunolocalization remains a challenging method, especially in tissues protected by waxy layers and complex cell wall carbohydrates. Here, we present a robust method for whole-mount in situ immunolocalization in primary root meristems and lateral root primordia in Arabidopsis thaliana. For good epitope preservation, fixation is done in an alkaline paraformaldehyde/glutaraldehyde mixture. This fixative is suitable for detecting a wide range of proteins, including integral transmembrane proteins and proteins peripherally attached to the plasma membrane. From initiation until emergence from the primary root, lateral root primordia are surrounded by several layers of differentiated tissues with a complex cell wall composition that interferes with the efficient penetration of all buffers. Therefore, immunolocalization in early lateral root primordia requires a modified method, including a strong solvent treatment for removal of hydrophobic barriers and a specific cocktail of cell wall-degrading enzymes. The presented method allows for easy, reliable, and high-quality in situ detection of the subcellular localization of endogenous proteins in primary and lateral root meristems without the need of time-consuming crosses or making translational fusions to fluorescent proteins. AU - Karampelias, Michael AU - Tejos, Ricardo AU - Friml, Jirí AU - Vanneste, Steffen ED - Ristova, Daniela ED - Barbez, Elke ID - 411 T2 - Root Development. Methods and Protocols TI - Optimized whole mount in situ immunolocalization for Arabidopsis thaliana root meristems and lateral root primordia VL - 1761 ER - TY - JOUR AB - Inhibition of the endoplasmic reticulum stress pathway may hold the key to Zika virus-associated microcephaly treatment. AU - Novarino, Gaia ID - 456 IS - 423 JF - Science Translational Medicine TI - Zika-associated microcephaly: Reduce the stress and race for the treatment VL - 10 ER - TY - JOUR AB - In 2013, a publication repository was implemented at IST Austria and 2015 after a thorough preparation phase a data repository was implemented - both based on the Open Source Software EPrints. In this text, designed as field report, we will reflect on our experiences with Open Source Software in general and specifically with EPrints regarding technical aspects but also regarding their characteristics of the user community. The second part is a pleading for including the end users in the process of implementation, adaption and evaluation. AU - Petritsch, Barbara AU - Porsche, Jana ID - 53 IS - 1 JF - VÖB Mitteilungen TI - IST PubRep and IST DataRep: the institutional repositories at IST Austria VL - 71 ER - TY - JOUR AB - We consider the problem of consensus in the challenging classic model. In this model, the adversary is adaptive; it can choose which processors crash at any point during the course of the algorithm. Further, communication is via asynchronous message passing: there is no known upper bound on the time to send a message from one processor to another, and all messages and coin flips are seen by the adversary. We describe a new randomized consensus protocol with expected message complexity O(n2log2n) when fewer than n / 2 processes may fail by crashing. This is an almost-linear improvement over the best previously known protocol, and within logarithmic factors of a known Ω(n2) message lower bound. The protocol further ensures that no process sends more than O(nlog3n) messages in expectation, which is again within logarithmic factors of optimal. We also present a generalization of the algorithm to an arbitrary number of failures t, which uses expected O(nt+t2log2t) total messages. Our approach is to build a message-efficient, resilient mechanism for aggregating individual processor votes, implementing the message-passing equivalent of a weak shared coin. Roughly, in our protocol, a processor first announces its votes to small groups, then propagates them to increasingly larger groups as it generates more and more votes. To bound the number of messages that an individual process might have to send or receive, the protocol progressively increases the weight of generated votes. The main technical challenge is bounding the impact of votes that are still “in flight” (generated, but not fully propagated) on the final outcome of the shared coin, especially since such votes might have different weights. We achieve this by leveraging the structure of the algorithm, and a technical argument based on martingale concentration bounds. Overall, we show that it is possible to build an efficient message-passing implementation of a shared coin, and in the process (almost-optimally) solve the classic consensus problem in the asynchronous message-passing model. AU - Alistarh, Dan-Adrian AU - Aspnes, James AU - King, Valerie AU - Saia, Jared ID - 536 IS - 6 JF - Distributed Computing SN - 01782770 TI - Communication-efficient randomized consensus VL - 31 ER - TY - JOUR AB - We analyse the canonical Bogoliubov free energy functional in three dimensions at low temperatures in the dilute limit. We prove existence of a first-order phase transition and, in the limit (Formula presented.), we determine the critical temperature to be (Formula presented.) to leading order. Here, (Formula presented.) is the critical temperature of the free Bose gas, ρ is the density of the gas and a is the scattering length of the pair-interaction potential V. We also prove asymptotic expansions for the free energy. In particular, we recover the Lee–Huang–Yang formula in the limit (Formula presented.). AU - Napiórkowski, Marcin M AU - Reuvers, Robin AU - Solovej, Jan ID - 554 IS - 1 JF - Communications in Mathematical Physics SN - 00103616 TI - The Bogoliubov free energy functional II: The dilute Limit VL - 360 ER - TY - CHAP AB - Primary neuronal cell culture preparations are widely used to investigate synaptic functions. This chapter describes a detailed protocol for the preparation of a neuronal cell culture in which giant calyx-type synaptic terminals are formed. This chapter also presents detailed protocols for utilizing the main technical advantages provided by such a preparation, namely, labeling and imaging of synaptic organelles and electrophysiological recordings directly from presynaptic terminals. AU - Dimitrov, Dimitar AU - Guillaud, Laurent AU - Eguchi, Kohgaku AU - Takahashi, Tomoyuki ED - Skaper, Stephen D. ID - 562 T2 - Neurotrophic Factors TI - Culture of mouse giant central nervous system synapses and application for imaging and electrophysiological analyses VL - 1727 ER - TY - CHAP AB - Graph-based games are an important tool in computer science. They have applications in synthesis, verification, refinement, and far beyond. We review graphbased games with objectives on infinite plays. We give definitions and algorithms to solve the games and to give a winning strategy. The objectives we consider are mostly Boolean, but we also look at quantitative graph-based games and their objectives. Synthesis aims to turn temporal logic specifications into correct reactive systems. We explain the reduction of synthesis to graph-based games (or equivalently tree automata) using synthesis of LTL specifications as an example. We treat the classical approach that uses determinization of parity automata and more modern approaches. AU - Bloem, Roderick AU - Chatterjee, Krishnendu AU - Jobstmann, Barbara ED - Henzinger, Thomas A ED - Clarke, Edmund M. ED - Veith, Helmut ED - Bloem, Roderick ID - 59 SN - 978-3-319-10574-1 T2 - Handbook of Model Checking TI - Graph games and reactive synthesis ER - TY - CHAP AB - Model checking is a computer-assisted method for the analysis of dynamical systems that can be modeled by state-transition systems. Drawing from research traditions in mathematical logic, programming languages, hardware design, and theoretical computer science, model checking is now widely used for the verification of hardware and software in industry. This chapter is an introduction and short survey of model checking. The chapter aims to motivate and link the individual chapters of the handbook, and to provide context for readers who are not familiar with model checking. AU - Clarke, Edmund AU - Henzinger, Thomas A AU - Veith, Helmut ED - Henzinger, Thomas A ID - 60 T2 - Handbook of Model Checking TI - Introduction to model checking ER - TY - JOUR AB - Blood platelets are critical for hemostasis and thrombosis, but also play diverse roles during immune responses. We have recently reported that platelets migrate at sites of infection in vitro and in vivo. Importantly, platelets use their ability to migrate to collect and bundle fibrin (ogen)-bound bacteria accomplishing efficient intravascular bacterial trapping. Here, we describe a method that allows analyzing platelet migration in vitro, focusing on their ability to collect bacteria and trap bacteria under flow. AU - Fan, Shuxia AU - Lorenz, Michael AU - Massberg, Steffen AU - Gärtner, Florian R ID - 6354 IS - 18 JF - Bio-Protocol KW - Platelets KW - Cell migration KW - Bacteria KW - Shear flow KW - Fibrinogen KW - E. coli SN - 2331-8325 TI - Platelet migration and bacterial trapping assay under flow VL - 8 ER - TY - GEN AU - Petritsch, Barbara ID - 6459 KW - Open Access KW - Publication Analysis TI - Open Access at IST Austria 2009-2017 ER - TY - CHAP AB - This chapter finds an agreement of equivariant indices of semi-classical homomorphisms between pairwise mirror branes in the GL2 Higgs moduli space on a Riemann surface. On one side of the agreement, components of the Lagrangian brane of U(1,1) Higgs bundles, whose mirror was proposed by Hitchin to be certain even exterior powers of the hyperholomorphic Dirac bundle on the SL2 Higgs moduli space, are present. The agreement arises from a mysterious functional equation. This gives strong computational evidence for Hitchin’s proposal. AU - Hausel, Tamás AU - Mellit, Anton AU - Pei, Du ID - 6525 SN - 9780198802013 T2 - Geometry and Physics: Volume I TI - Mirror symmetry with branes by equivariant verlinde formulas ER - TY - JOUR AB - We consider spectral properties and the edge universality of sparse random matrices, the class of random matrices that includes the adjacency matrices of the Erdős–Rényi graph model G(N, p). We prove a local law for the eigenvalue density up to the spectral edges. Under a suitable condition on the sparsity, we also prove that the rescaled extremal eigenvalues exhibit GOE Tracy–Widom fluctuations if a deterministic shift of the spectral edge due to the sparsity is included. For the adjacency matrix of the Erdős–Rényi graph this establishes the Tracy–Widom fluctuations of the second largest eigenvalue when p is much larger than N−2/3 with a deterministic shift of order (Np)−1. AU - Lee, Jii AU - Schnelli, Kevin ID - 690 IS - 1-2 JF - Probability Theory and Related Fields TI - Local law and Tracy–Widom limit for sparse random matrices VL - 171 ER - TY - JOUR AB - We consider the NP-hard problem of MAP-inference for undirected discrete graphical models. We propose a polynomial time and practically efficient algorithm for finding a part of its optimal solution. Specifically, our algorithm marks some labels of the considered graphical model either as (i) optimal, meaning that they belong to all optimal solutions of the inference problem; (ii) non-optimal if they provably do not belong to any solution. With access to an exact solver of a linear programming relaxation to the MAP-inference problem, our algorithm marks the maximal possible (in a specified sense) number of labels. We also present a version of the algorithm, which has access to a suboptimal dual solver only and still can ensure the (non-)optimality for the marked labels, although the overall number of the marked labels may decrease. We propose an efficient implementation, which runs in time comparable to a single run of a suboptimal dual solver. Our method is well-scalable and shows state-of-the-art results on computational benchmarks from machine learning and computer vision. AU - Shekhovtsov, Alexander AU - Swoboda, Paul AU - Savchynskyy, Bogdan ID - 703 IS - 7 JF - IEEE Transactions on Pattern Analysis and Machine Intelligence SN - 01628828 TI - Maximum persistency via iterative relaxed inference with graphical models VL - 40 ER - TY - CONF AB - Training deep learning models has received tremendous research interest recently. In particular, there has been intensive research on reducing the communication cost of training when using multiple computational devices, through reducing the precision of the underlying data representation. Naturally, such methods induce system trade-offs—lowering communication precision could de-crease communication overheads and improve scalability; but, on the other hand, it can also reduce the accuracy of training. In this paper, we study this trade-off space, and ask:Can low-precision communication consistently improve the end-to-end performance of training modern neural networks, with no accuracy loss?From the performance point of view, the answer to this question may appear deceptively easy: compressing communication through low precision should help when the ratio between communication and computation is high. However, this answer is less straightforward when we try to generalize this principle across various neural network architectures (e.g., AlexNet vs. ResNet),number of GPUs (e.g., 2 vs. 8 GPUs), machine configurations(e.g., EC2 instances vs. NVIDIA DGX-1), communication primitives (e.g., MPI vs. NCCL), and even different GPU architectures(e.g., Kepler vs. Pascal). Currently, it is not clear how a realistic realization of all these factors maps to the speed up provided by low-precision communication. In this paper, we conduct an empirical study to answer this question and report the insights. AU - Grubic, Demjan AU - Tam, Leo AU - Alistarh, Dan-Adrian AU - Zhang, Ce ID - 7116 SN - 2367-2005 T2 - Proceedings of the 21st International Conference on Extending Database Technology TI - Synchronous multi-GPU training for deep learning with low-precision communications: An empirical study ER - TY - CONF AB - Proofs of space (PoS) [Dziembowski et al., CRYPTO'15] are proof systems where a prover can convince a verifier that he "wastes" disk space. PoS were introduced as a more ecological and economical replacement for proofs of work which are currently used to secure blockchains like Bitcoin. In this work we investigate extensions of PoS which allow the prover to embed useful data into the dedicated space, which later can be recovered. Our first contribution is a security proof for the original PoS from CRYPTO'15 in the random oracle model (the original proof only applied to a restricted class of adversaries which can store a subset of the data an honest prover would store). When this PoS is instantiated with recent constructions of maximally depth robust graphs, our proof implies basically optimal security. As a second contribution we show three different extensions of this PoS where useful data can be embedded into the space required by the prover. Our security proof for the PoS extends (non-trivially) to these constructions. We discuss how some of these variants can be used as proofs of catalytic space (PoCS), a notion we put forward in this work, and which basically is a PoS where most of the space required by the prover can be used to backup useful data. Finally we discuss how one of the extensions is a candidate construction for a proof of replication (PoR), a proof system recently suggested in the Filecoin whitepaper. AU - Pietrzak, Krzysztof Z ID - 7407 SN - 1868-8969 T2 - 10th Innovations in Theoretical Computer Science Conference (ITCS 2019) TI - Proofs of catalytic space VL - 124 ER - TY - JOUR AB - The concurrent memory reclamation problem is that of devising a way for a deallocating thread to verify that no other concurrent threads hold references to a memory block being deallocated. To date, in the absence of automatic garbage collection, there is no satisfactory solution to this problem; existing tracking methods like hazard pointers, reference counters, or epoch-based techniques like RCU are either prohibitively expensive or require significant programming expertise to the extent that implementing them efficiently can be worthy of a publication. None of the existing techniques are automatic or even semi-automated. In this article, we take a new approach to concurrent memory reclamation. Instead of manually tracking access to memory locations as done in techniques like hazard pointers, or restricting shared accesses to specific epoch boundaries as in RCU, our algorithm, called ThreadScan, leverages operating system signaling to automatically detect which memory locations are being accessed by concurrent threads. Initial empirical evidence shows that ThreadScan scales surprisingly well and requires negligible programming effort beyond the standard use of Malloc and Free. AU - Alistarh, Dan-Adrian AU - Leiserson, William AU - Matveev, Alexander AU - Shavit, Nir ID - 6001 IS - 4 JF - ACM Transactions on Parallel Computing SN - 2329-4949 TI - ThreadScan: Automatic and scalable memory reclamation VL - 4 ER - TY - CONF AB - Deep neural networks (DNNs) continue to make significant advances, solving tasks from image classification to translation or reinforcement learning. One aspect of the field receiving considerable attention is efficiently executing deep models in resource-constrained environments, such as mobile or embedded devices. This paper focuses on this problem, and proposes two new compression methods, which jointly leverage weight quantization and distillation of larger teacher networks into smaller student networks. The first method we propose is called quantized distillation and leverages distillation during the training process, by incorporating distillation loss, expressed with respect to the teacher, into the training of a student network whose weights are quantized to a limited set of levels. The second method, differentiable quantization, optimizes the location of quantization points through stochastic gradient descent, to better fit the behavior of the teacher model. We validate both methods through experiments on convolutional and recurrent architectures. We show that quantized shallow students can reach similar accuracy levels to full-precision teacher models, while providing order of magnitude compression, and inference speedup that is linear in the depth reduction. In sum, our results enable DNNs for resource-constrained environments to leverage architecture and accuracy advances developed on more powerful devices. AU - Polino, Antonio AU - Pascanu, Razvan AU - Alistarh, Dan-Adrian ID - 7812 T2 - 6th International Conference on Learning Representations TI - Model compression via distillation and quantization ER - TY - GEN AB - The cerebral cortex contains multiple hierarchically organized areas with distinctive cytoarchitectonical patterns, but the cellular mechanisms underlying the emergence of this diversity remain unclear. Here, we have quantitatively investigated the neuronal output of individual progenitor cells in the ventricular zone of the developing mouse neocortex using a combination of methods that together circumvent the biases and limitations of individual approaches. We found that individual cortical progenitor cells show a high degree of stochasticity and generate pyramidal cell lineages that adopt a wide range of laminar configurations. Mathematical modelling these lineage data suggests that a small number of progenitor cell populations, each generating pyramidal cells following different stochastic developmental programs, suffice to generate the heterogenous complement of pyramidal cell lineages that collectively build the complex cytoarchitecture of the neocortex. AU - Llorca, Alfredo AU - Ciceri, Gabriele AU - Beattie, Robert J AU - Wong, Fong K. AU - Diana, Giovanni AU - Serafeimidou, Eleni AU - Fernández-Otero, Marian AU - Streicher, Carmen AU - Arnold, Sebastian J. AU - Meyer, Martin AU - Hippenmeyer, Simon AU - Maravall, Miguel AU - Marín, Oscar ID - 8547 T2 - bioRxiv TI - Heterogeneous progenitor cell behaviors underlie the assembly of neocortical cytoarchitecture ER - TY - CHAP AB - Responsiveness—the requirement that every request to a system be eventually handled—is one of the fundamental liveness properties of a reactive system. Average response time is a quantitative measure for the responsiveness requirement used commonly in performance evaluation. We show how average response time can be computed on state-transition graphs, on Markov chains, and on game graphs. In all three cases, we give polynomial-time algorithms. AU - Chatterjee, Krishnendu AU - Henzinger, Thomas A AU - Otop, Jan ED - Lohstroh, Marten ED - Derler, Patricia ED - Sirjani, Marjan ID - 86 T2 - Principles of Modeling TI - Computing average response time VL - 10760 ER - TY - JOUR AU - Danzl, Johann G ID - 9229 IS - S1 JF - Opera Medica et Physiologica SN - 2500-2287 TI - Diffraction-unlimited optical imaging for synaptic physiology VL - 4 ER - TY - CONF AB - Network games are widely used as a model for selfish resource-allocation problems. In the classicalmodel, each player selects a path connecting her source and target vertices. The cost of traversingan edge depends on theload; namely, number of players that traverse it. Thus, it abstracts the factthat different users may use a resource at different times and for different durations, which playsan important role in determining the costs of the users in reality. For example, when transmittingpackets in a communication network, routing traffic in a road network, or processing a task in aproduction system, actual sharing and congestion of resources crucially depends on time.In [13], we introducedtimed network games, which add a time component to network games.Each vertexvin the network is associated with a cost function, mapping the load onvto theprice that a player pays for staying invfor one time unit with this load. Each edge in thenetwork is guarded by the time intervals in which it can be traversed, which forces the players tospend time in the vertices. In this work we significantly extend the way time can be referred toin timed network games. In the model we study, the network is equipped withclocks, and, as intimed automata, edges are guarded by constraints on the values of the clocks, and their traversalmay involve a reset of some clocks. We argue that the stronger model captures many realisticnetworks. The addition of clocks breaks the techniques we developed in [13] and we developnew techniques in order to show that positive results on classic network games carry over to thestronger timed setting. AU - Avni, Guy AU - Guha, Shibashis AU - Kupferman, Orna ID - 6005 SN - 1868-8969 TI - Timed network games with clocks VL - 117 ER - TY - JOUR AB - More than 100 years after Grigg’s influential analysis of species’ borders, the causes of limits to species’ ranges still represent a puzzle that has never been understood with clarity. The topic has become especially important recently as many scientists have become interested in the potential for species’ ranges to shift in response to climate change—and yet nearly all of those studies fail to recognise or incorporate evolutionary genetics in a way that relates to theoretical developments. I show that range margins can be understood based on just two measurable parameters: (i) the fitness cost of dispersal—a measure of environmental heterogeneity—and (ii) the strength of genetic drift, which reduces genetic diversity. Together, these two parameters define an ‘expansion threshold’: adaptation fails when genetic drift reduces genetic diversity below that required for adaptation to a heterogeneous environment. When the key parameters drop below this expansion threshold locally, a sharp range margin forms. When they drop below this threshold throughout the species’ range, adaptation collapses everywhere, resulting in either extinction or formation of a fragmented metapopulation. Because the effects of dispersal differ fundamentally with dimension, the second parameter—the strength of genetic drift—is qualitatively different compared to a linear habitat. In two-dimensional habitats, genetic drift becomes effectively independent of selection. It decreases with ‘neighbourhood size’—the number of individuals accessible by dispersal within one generation. Moreover, in contrast to earlier predictions, which neglected evolution of genetic variance and/or stochasticity in two dimensions, dispersal into small marginal populations aids adaptation. This is because the reduction of both genetic and demographic stochasticity has a stronger effect than the cost of dispersal through increased maladaptation. The expansion threshold thus provides a novel, theoretically justified, and testable prediction for formation of the range margin and collapse of the species’ range. AU - Polechova, Jitka ID - 315 IS - 6 JF - PLoS Biology SN - 15449173 TI - Is the sky the limit? On the expansion threshold of a species’ range VL - 16 ER - TY - CONF AB - A drawing of a graph on a surface is independently even if every pair of nonadjacent edges in the drawing crosses an even number of times. The ℤ2-genus of a graph G is the minimum g such that G has an independently even drawing on the orientable surface of genus g. An unpublished result by Robertson and Seymour implies that for every t, every graph of sufficiently large genus contains as a minor a projective t × t grid or one of the following so-called t-Kuratowski graphs: K3, t, or t copies of K5 or K3,3 sharing at most 2 common vertices. We show that the ℤ2-genus of graphs in these families is unbounded in t; in fact, equal to their genus. Together, this implies that the genus of a graph is bounded from above by a function of its ℤ2-genus, solving a problem posed by Schaefer and Štefankovič, and giving an approximate version of the Hanani-Tutte theorem on orientable surfaces. AU - Fulek, Radoslav AU - Kynčl, Jan ID - 186 TI - The ℤ2-Genus of Kuratowski minors VL - 99 ER - TY - CONF AB - A thrackle is a graph drawn in the plane so that every pair of its edges meet exactly once: either at a common end vertex or in a proper crossing. We prove that any thrackle of n vertices has at most 1.3984n edges. Quasi-thrackles are defined similarly, except that every pair of edges that do not share a vertex are allowed to cross an odd number of times. It is also shown that the maximum number of edges of a quasi-thrackle on n vertices is 3/2(n-1), and that this bound is best possible for infinitely many values of n. AU - Fulek, Radoslav AU - Pach, János ID - 433 TI - Thrackles: An improved upper bound VL - 10692 ER - TY - GEN AB - Both classical and recent studies suggest that chromosomal inversion polymorphisms are important in adaptation and speciation. However, biases in discovery and reporting of inversions make it difficult to assess their prevalence and biological importance. Here, we use an approach based on linkage disequilibrium among markers genotyped for samples collected across a transect between contrasting habitats to detect chromosomal rearrangements de novo. We report 17 polymorphic rearrangements in a single locality for the coastal marine snail, Littorina saxatilis. Patterns of diversity in the field and of recombination in controlled crosses provide strong evidence that at least the majority of these rearrangements are inversions. Most show clinal changes in frequency between habitats, suggestive of divergent selection, but only one appears to be fixed for different arrangements in the two habitats. Consistent with widespread evidence for balancing selection on inversion polymorphisms, we argue that a combination of heterosis and divergent selection can explain the observed patterns and should be considered in other systems spanning environmental gradients. AU - Faria, Rui AU - Chaube, Pragya AU - Morales, Hernán E. AU - Larsson, Tomas AU - Lemmon, Alan R. AU - Lemmon, Emily M. AU - Rafajlović, Marina AU - Panova, Marina AU - Ravinet, Mark AU - Johannesson, Kerstin AU - Westram, Anja M AU - Butlin, Roger K. ID - 9837 TI - Data from: Multiple chromosomal rearrangements in a hybrid zone between Littorina saxatilis ecotypes ER - TY - GEN AB - We consider the problem of expected cost analysis over nondeterministic probabilistic programs, which aims at automated methods for analyzing the resource-usage of such programs. Previous approaches for this problem could only handle nonnegative bounded costs. However, in many scenarios, such as queuing networks or analysis of cryptocurrency protocols, both positive and negative costs are necessary and the costs are unbounded as well. In this work, we present a sound and efficient approach to obtain polynomial bounds on the expected accumulated cost of nondeterministic probabilistic programs. Our approach can handle (a) general positive and negative costs with bounded updates in variables; and (b) nonnegative costs with general updates to variables. We show that several natural examples which could not be handled by previous approaches are captured in our framework. Moreover, our approach leads to an efficient polynomial-time algorithm, while no previous approach for cost analysis of probabilistic programs could guarantee polynomial runtime. Finally, we show the effectiveness of our approach by presenting experimental results on a variety of programs, motivated by real-world applications, for which we efficiently synthesize tight resource-usage bounds. AU - Anonymous, 1 AU - Anonymous, 2 AU - Anonymous, 3 AU - Anonymous, 4 AU - Anonymous, 5 AU - Anonymous, 6 ID - 5457 SN - 2664-1690 TI - Cost analysis of nondeterministic probabilistic programs ER - TY - CHAP AB - We prove that every congruence distributive variety has directed Jónsson terms, and every congruence modular variety has directed Gumm terms. The directed terms we construct witness every case of absorption witnessed by the original Jónsson or Gumm terms. This result is equivalent to a pair of claims about absorption for admissible preorders in congruence distributive and congruence modular varieties, respectively. For finite algebras, these absorption theorems have already seen significant applications, but until now, it was not clear if the theorems hold for general algebras as well. Our method also yields a novel proof of a result by P. Lipparini about the existence of a chain of terms (which we call Pixley terms) in varieties that are at the same time congruence distributive and k-permutable for some k. AU - Kazda, Alexandr AU - Kozik, Marcin AU - McKenzie, Ralph AU - Moore, Matthew ED - Czelakowski, J ID - 10864 SN - 2211-2758 T2 - Don Pigozzi on Abstract Algebraic Logic, Universal Algebra, and Computer Science TI - Absorption and directed Jónsson terms VL - 16 ER - TY - CONF AB - We prove that for every d ≥ 2, deciding if a pure, d-dimensional, simplicial complex is shellable is NP-hard, hence NP-complete. This resolves a question raised, e.g., by Danaraj and Klee in 1978. Our reduction also yields that for every d ≥ 2 and k ≥ 0, deciding if a pure, d-dimensional, simplicial complex is k-decomposable is NP-hard. For d ≥ 3, both problems remain NP-hard when restricted to contractible pure d-dimensional complexes. AU - Goaoc, Xavier AU - Paták, Pavel AU - Patakova, Zuzana AU - Tancer, Martin AU - Wagner, Uli ID - 184 TI - Shellability is NP-complete VL - 99 ER - TY - CONF AB - In graph theory, as well as in 3-manifold topology, there exist several width-type parameters to describe how "simple" or "thin" a given graph or 3-manifold is. These parameters, such as pathwidth or treewidth for graphs, or the concept of thin position for 3-manifolds, play an important role when studying algorithmic problems; in particular, there is a variety of problems in computational 3-manifold topology - some of them known to be computationally hard in general - that become solvable in polynomial time as soon as the dual graph of the input triangulation has bounded treewidth. In view of these algorithmic results, it is natural to ask whether every 3-manifold admits a triangulation of bounded treewidth. We show that this is not the case, i.e., that there exists an infinite family of closed 3-manifolds not admitting triangulations of bounded pathwidth or treewidth (the latter implies the former, but we present two separate proofs). We derive these results from work of Agol and of Scharlemann and Thompson, by exhibiting explicit connections between the topology of a 3-manifold M on the one hand and width-type parameters of the dual graphs of triangulations of M on the other hand, answering a question that had been raised repeatedly by researchers in computational 3-manifold topology. In particular, we show that if a closed, orientable, irreducible, non-Haken 3-manifold M has a triangulation of treewidth (resp. pathwidth) k then the Heegaard genus of M is at most 48(k+1) (resp. 4(3k+1)). AU - Huszár, Kristóf AU - Spreer, Jonathan AU - Wagner, Uli ID - 285 SN - 18688969 TI - On the treewidth of triangulated 3-manifolds VL - 99 ER - TY - GEN AB - This dataset contains a GitHub repository containing all the data, analysis, Nextflow workflows and Jupyter notebooks to replicate the manuscript titled "Fast and accurate large multiple sequence alignments with a root-to-leaf regressive method". It also contains the Multiple Sequence Alignments (MSAs) generated and well as the main figures and tables from the manuscript. The repository is also available at GitHub (https://github.com/cbcrg/dpa-analysis) release `v1.2`. For details on how to use the regressive alignment algorithm, see the T-Coffee software suite (https://github.com/cbcrg/tcoffee). AU - Garriga, Edgar AU - di Tommaso, Paolo AU - Magis, Cedrik AU - Erb, Ionas AU - Mansouri, Leila AU - Baltzis, Athanasios AU - Laayouni, Hafid AU - Kondrashov, Fyodor AU - Floden, Evan AU - Notredame, Cedric ID - 13059 TI - Fast and accurate large multiple sequence alignments with a root-to-leaf regressive method ER - TY - THES AB - Nowadays, quantum computation is receiving more and more attention as an alternative to the classical way of computing. For realizing a quantum computer, different devices are investigated as potential quantum bits. In this thesis, the focus is on Ge hut wires, which turned out to be promising candidates for implementing hole spin quantum bits. The advantages of Ge as a material system are the low hyperfine interaction for holes and the strong spin orbit coupling, as well as the compatibility with the highly developed CMOS processes in industry. In addition, Ge can also be isotopically purified which is expected to boost the spin coherence times. The strong spin orbit interaction for holes in Ge on the one hand enables the full electrical control of the quantum bit and on the other hand should allow short spin manipulation times. Starting with a bare Si wafer, this work covers the entire process reaching from growth over the fabrication and characterization of hut wire devices up to the demonstration of hole spin resonance. From experiments with single quantum dots, a large g-factor anisotropy between the in-plane and the out-of-plane direction was found. A comparison to a theoretical model unveiled the heavy-hole character of the lowest energy states. The second part of the thesis addresses double quantum dot devices, which were realized by adding two gate electrodes to a hut wire. In such devices, Pauli spin blockade was observed, which can serve as a read-out mechanism for spin quantum bits. Applying oscillating electric fields in spin blockade allowed the demonstration of continuous spin rotations and the extraction of a lower bound for the spin dephasing time. Despite the strong spin orbit coupling in Ge, the obtained value for the dephasing time is comparable to what has been recently reported for holes in Si. All in all, the presented results point out the high potential of Ge hut wires as a platform for long-lived, fast and fully electrically tunable hole spin quantum bits. AU - Watzinger, Hannes ID - 49 SN - 2663-337X TI - Ge hut wires - from growth to hole spin resonance ER - TY - THES AB - We describe arrangements of three-dimensional spheres from a geometrical and topological point of view. Real data (fitting this setup) often consist of soft spheres which show certain degree of deformation while strongly packing against each other. In this context, we answer the following questions: If we model a soft packing of spheres by hard spheres that are allowed to overlap, can we measure the volume in the overlapped areas? Can we be more specific about the overlap volume, i.e. quantify how much volume is there covered exactly twice, three times, or k times? What would be a good optimization criteria that rule the arrangement of soft spheres while making a good use of the available space? Fixing a particular criterion, what would be the optimal sphere configuration? The first result of this thesis are short formulas for the computation of volumes covered by at least k of the balls. The formulas exploit information contained in the order-k Voronoi diagrams and its closely related Level-k complex. The used complexes lead to a natural generalization into poset diagrams, a theoretical formalism that contains the order-k and degree-k diagrams as special cases. In parallel, we define different criteria to determine what could be considered an optimal arrangement from a geometrical point of view. Fixing a criterion, we find optimal soft packing configurations in 2D and 3D where the ball centers lie on a lattice. As a last step, we use tools from computational topology on real physical data, to show the potentials of higher-order diagrams in the description of melting crystals. The results of the experiments leaves us with an open window to apply the theories developed in this thesis in real applications. AU - Iglesias Ham, Mabel ID - 201 SN - 2663-337X TI - Multiple covers with balls ER - TY - THES AB - The most common assumption made in statistical learning theory is the assumption of the independent and identically distributed (i.i.d.) data. While being very convenient mathematically, it is often very clearly violated in practice. This disparity between the machine learning theory and applications underlies a growing demand in the development of algorithms that learn from dependent data and theory that can provide generalization guarantees similar to the independent situations. This thesis is dedicated to two variants of dependencies that can arise in practice. One is a dependence on the level of samples in a single learning task. Another dependency type arises in the multi-task setting when the tasks are dependent on each other even though the data for them can be i.i.d. In both cases we model the data (samples or tasks) as stochastic processes and introduce new algorithms for both settings that take into account and exploit the resulting dependencies. We prove the theoretical guarantees on the performance of the introduced algorithms under different evaluation criteria and, in addition, we compliment the theoretical study by the empirical one, where we evaluate some of the algorithms on two real world datasets to highlight their practical applicability. AU - Zimin, Alexander ID - 68 SN - 2663-337X TI - Learning from dependent data ER - TY - THES AB - A proof system is a protocol between a prover and a verifier over a common input in which an honest prover convinces the verifier of the validity of true statements. Motivated by the success of decentralized cryptocurrencies, exemplified by Bitcoin, the focus of this thesis will be on proof systems which found applications in some sustainable alternatives to Bitcoin, such as the Spacemint and Chia cryptocurrencies. In particular, we focus on proofs of space and proofs of sequential work. Proofs of space (PoSpace) were suggested as more ecological, economical, and egalitarian alternative to the energy-wasteful proof-of-work mining of Bitcoin. However, the state-of-the-art constructions of PoSpace are based on sophisticated graph pebbling lower bounds, and are therefore complex. Moreover, when these PoSpace are used in cryptocurrencies like Spacemint, miners can only start mining after ensuring that a commitment to their space is already added in a special transaction to the blockchain. Proofs of sequential work (PoSW) are proof systems in which a prover, upon receiving a statement x and a time parameter T, computes a proof which convinces the verifier that T time units had passed since x was received. Whereas Spacemint assumes synchrony to retain some interesting Bitcoin dynamics, Chia requires PoSW with unique proofs, i.e., PoSW in which it is hard to come up with more than one accepting proof for any true statement. In this thesis we construct simple and practically-efficient PoSpace and PoSW. When using our PoSpace in cryptocurrencies, miners can start mining on the fly, like in Bitcoin, and unlike current constructions of PoSW, which either achieve efficient verification of sequential work, or faster-than-recomputing verification of correctness of proofs, but not both at the same time, ours achieve the best of these two worlds. AU - Abusalah, Hamza M ID - 83 SN - 2663-337X TI - Proof systems for sustainable decentralized cryptocurrencies ER - TY - THES AB - Modern computer vision systems heavily rely on statistical machine learning models, which typically require large amounts of labeled data to be learned reliably. Moreover, very recently computer vision research widely adopted techniques for representation learning, which further increase the demand for labeled data. However, for many important practical problems there is relatively small amount of labeled data available, so it is problematic to leverage full potential of the representation learning methods. One way to overcome this obstacle is to invest substantial resources into producing large labelled datasets. Unfortunately, this can be prohibitively expensive in practice. In this thesis we focus on the alternative way of tackling the aforementioned issue. We concentrate on methods, which make use of weakly-labeled or even unlabeled data. Specifically, the first half of the thesis is dedicated to the semantic image segmentation task. We develop a technique, which achieves competitive segmentation performance and only requires annotations in a form of global image-level labels instead of dense segmentation masks. Subsequently, we present a new methodology, which further improves segmentation performance by leveraging tiny additional feedback from a human annotator. By using our methods practitioners can greatly reduce the amount of data annotation effort, which is required to learn modern image segmentation models. In the second half of the thesis we focus on methods for learning from unlabeled visual data. We study a family of autoregressive models for modeling structure of natural images and discuss potential applications of these models. Moreover, we conduct in-depth study of one of these applications, where we develop the state-of-the-art model for the probabilistic image colorization task. AU - Kolesnikov, Alexander ID - 197 SN - 2663-337X TI - Weakly-Supervised Segmentation and Unsupervised Modeling of Natural Images ER - TY - JOUR AB - A central problem of algebraic topology is to understand the homotopy groups 𝜋𝑑(𝑋) of a topological space X. For the computational version of the problem, it is well known that there is no algorithm to decide whether the fundamental group 𝜋1(𝑋) of a given finite simplicial complex X is trivial. On the other hand, there are several algorithms that, given a finite simplicial complex X that is simply connected (i.e., with 𝜋1(𝑋) trivial), compute the higher homotopy group 𝜋𝑑(𝑋) for any given 𝑑≥2 . However, these algorithms come with a caveat: They compute the isomorphism type of 𝜋𝑑(𝑋) , 𝑑≥2 as an abstract finitely generated abelian group given by generators and relations, but they work with very implicit representations of the elements of 𝜋𝑑(𝑋) . Converting elements of this abstract group into explicit geometric maps from the d-dimensional sphere 𝑆𝑑 to X has been one of the main unsolved problems in the emerging field of computational homotopy theory. Here we present an algorithm that, given a simply connected space X, computes 𝜋𝑑(𝑋) and represents its elements as simplicial maps from a suitable triangulation of the d-sphere 𝑆𝑑 to X. For fixed d, the algorithm runs in time exponential in size(𝑋) , the number of simplices of X. Moreover, we prove that this is optimal: For every fixed 𝑑≥2 , we construct a family of simply connected spaces X such that for any simplicial map representing a generator of 𝜋𝑑(𝑋) , the size of the triangulation of 𝑆𝑑 on which the map is defined, is exponential in size(𝑋) . AU - Filakovský, Marek AU - Franek, Peter AU - Wagner, Uli AU - Zhechev, Stephan Y ID - 6774 IS - 3-4 JF - Journal of Applied and Computational Topology SN - 2367-1726 TI - Computing simplicial representatives of homotopy group elements VL - 2 ER - TY - CONF AB - Synchronous programs are easy to specify because the side effects of an operation are finished by the time the invocation of the operation returns to the caller. Asynchronous programs, on the other hand, are difficult to specify because there are side effects due to pending computation scheduled as a result of the invocation of an operation. They are also difficult to verify because of the large number of possible interleavings of concurrent computation threads. We present synchronization, a new proof rule that simplifies the verification of asynchronous programs by introducing the fiction, for proof purposes, that asynchronous operations complete synchronously. Synchronization summarizes an asynchronous computation as immediate atomic effect. Modular verification is enabled via pending asynchronous calls in atomic summaries, and a complementary proof rule that eliminates pending asynchronous calls when components and their specifications are composed. We evaluate synchronization in the context of a multi-layer refinement verification methodology on a collection of benchmark programs. AU - Kragl, Bernhard AU - Qadeer, Shaz AU - Henzinger, Thomas A ID - 133 SN - 18688969 TI - Synchronizing the asynchronous VL - 118 ER -