@inproceedings{14174, abstract = {Building sample-efficient agents that generalize out-of-distribution (OOD) in real-world settings remains a fundamental unsolved problem on the path towards achieving higher-level cognition. One particularly promising approach is to begin with low-dimensional, pretrained representations of our world, which should facilitate efficient downstream learning and generalization. By training 240 representations and over 10,000 reinforcement learning (RL) policies on a simulated robotic setup, we evaluate to what extent different properties of pretrained VAE-based representations affect the OOD generalization of downstream agents. We observe that many agents are surprisingly robust to realistic distribution shifts, including the challenging sim-to-real case. In addition, we find that the generalization performance of a simple downstream proxy task reliably predicts the generalization performance of our RL agents under a wide range of OOD settings. Such proxy tasks can thus be used to select pretrained representations that will lead to agents that generalize.}, author = {Dittadi, Andrea and Träuble, Frederik and Wüthrich, Manuel and Widmaier, Felix and Gehler, Peter and Winther, Ole and Locatello, Francesco and Bachem, Olivier and Schölkopf, Bernhard and Bauer, Stefan}, booktitle = {10th International Conference on Learning Representations}, location = {Virtual}, title = {{The role of pretrained representations for the OOD generalization of reinforcement learning agents}}, year = {2022}, } @inproceedings{14175, abstract = {Predicting the future trajectory of a moving agent can be easy when the past trajectory continues smoothly but is challenging when complex interactions with other agents are involved. Recent deep learning approaches for trajectory prediction show promising performance and partially attribute this to successful reasoning about agent-agent interactions. However, it remains unclear which features such black-box models actually learn to use for making predictions. This paper proposes a procedure that quantifies the contributions of different cues to model performance based on a variant of Shapley values. Applying this procedure to state-of-the-art trajectory prediction methods on standard benchmark datasets shows that they are, in fact, unable to reason about interactions. Instead, the past trajectory of the target is the only feature used for predicting its future. For a task with richer social interaction patterns, on the other hand, the tested models do pick up such interactions to a certain extent, as quantified by our feature attribution method. We discuss the limits of the proposed method and its links to causality.}, author = {Makansi, Osama and Kügelgen, Julius von and Locatello, Francesco and Gehler, Peter and Janzing, Dominik and Brox, Thomas and Schölkopf, Bernhard}, booktitle = {10th International Conference on Learning Representations}, location = {Virtual}, title = {{You mostly walk alone: Analyzing feature attribution in trajectory prediction}}, year = {2022}, } @inproceedings{14215, abstract = {Geospatial Information Systems are used by researchers and Humanitarian Assistance and Disaster Response (HADR) practitioners to support a wide variety of important applications. However, collaboration between these actors is difficult due to the heterogeneous nature of geospatial data modalities (e.g., multi-spectral images of various resolutions, timeseries, weather data) and diversity of tasks (e.g., regression of human activity indicators or detecting forest fires). In this work, we present a roadmap towards the construction of a general-purpose neural architecture (GPNA) with a geospatial inductive bias, pre-trained on large amounts of unlabelled earth observation data in a self-supervised manner. We envision how such a model may facilitate cooperation between members of the community. We show preliminary results on the first step of the roadmap, where we instantiate an architecture that can process a wide variety of geospatial data modalities and demonstrate that it can achieve competitive performance with domain-specific architectures on tasks relating to the U.N.'s Sustainable Development Goals.}, author = {Rahaman, Nasim and Weiss, Martin and Träuble, Frederik and Locatello, Francesco and Lacoste, Alexandre and Bengio, Yoshua and Pal, Chris and Li, Li Erran and Schölkopf, Bernhard}, booktitle = {36th Conference on Neural Information Processing Systems}, location = {New Orleans, LA, United States}, title = {{A general purpose neural architecture for geospatial systems}}, year = {2022}, } @unpublished{14216, abstract = {CLIP proved that aligning visual and language spaces is key to solving many vision tasks without explicit training, but required to train image and text encoders from scratch on a huge dataset. LiT improved this by only training the text encoder and using a pre-trained vision network. In this paper, we show that a common space can be created without any training at all, using single-domain encoders (trained with or without supervision) and a much smaller amount of image-text pairs. Furthermore, our model has unique properties. Most notably, deploying a new version with updated training samples can be done in a matter of seconds. Additionally, the representations in the common space are easily interpretable as every dimension corresponds to the similarity of the input to a unique entry in the multimodal dataset. Experiments on standard zero-shot visual benchmarks demonstrate the typical transfer ability of image-text models. Overall, our method represents a simple yet surprisingly strong baseline for foundation multi-modal models, raising important questions on their data efficiency and on the role of retrieval in machine learning.}, author = {Norelli, Antonio and Fumero, Marco and Maiorca, Valentino and Moschella, Luca and Rodolà, Emanuele and Locatello, Francesco}, booktitle = {arXiv}, title = {{ASIF: Coupled data turns unimodal models to multimodal without training}}, doi = {10.48550/arXiv.2210.01738}, year = {2022}, } @unpublished{14220, abstract = {Although reinforcement learning has seen remarkable progress over the last years, solving robust dexterous object-manipulation tasks in multi-object settings remains a challenge. In this paper, we focus on models that can learn manipulation tasks in fixed multi-object settings and extrapolate this skill zero-shot without any drop in performance when the number of objects changes. We consider the generic task of bringing a specific cube out of a set to a goal position. We find that previous approaches, which primarily leverage attention and graph neural network-based architectures, do not generalize their skills when the number of input objects changes while scaling as K2. We propose an alternative plug-and-play module based on relational inductive biases to overcome these limitations. Besides exceeding performances in their training environment, we show that our approach, which scales linearly in K, allows agents to extrapolate and generalize zero-shot to any new object number.}, author = {Mambelli, Davide and Träuble, Frederik and Bauer, Stefan and Schölkopf, Bernhard and Locatello, Francesco}, booktitle = {arXiv}, title = {{Compositional multi-object reinforcement learning with linear relation networks}}, doi = {10.48550/arXiv.2201.13388}, year = {2022}, } @unpublished{14236, abstract = {We show an $(1+\epsilon)$-approximation algorithm for maintaining maximum $s$-$t$ flow under $m$ edge insertions in $m^{1/2+o(1)} \epsilon^{-1/2}$ amortized update time for directed, unweighted graphs. This constitutes the first sublinear dynamic maximum flow algorithm in general sparse graphs with arbitrarily good approximation guarantee.}, author = {Goranci, Gramoz and Henzinger, Monika H}, booktitle = {arXiv}, title = {{Incremental approximate maximum flow in m1/2+o(1) update time}}, doi = {10.48550/arXiv.2211.09606}, year = {2022}, } @article{14248, abstract = {Recent work by Forsgård indicates that not every convex lattice polygon arises as the characteristic polygon of an affine dimer or, equivalently, an admissible oriented line arrangement on the torus in general position. We begin the classication of convex lattice polygons arising as characteristic polygons of affine dimers. We present several general constructions of new affine dimers from old, and an algorithm for finding affine dimers with prescribed polygon. With these tools we prove that all lattice triangles, generalised parallelograms, and polygons of genus at most two admit an affine dimer.}, author = {Holmes, Daniel}, issn = {2576-3725}, journal = {PUMP Journal of Undergraduate Research}, keywords = {dimer model, hyperplane arrangement, torus, lattice polygon}, pages = {24--51}, publisher = {California State University}, title = {{Affine dimers from characteristic polygons}}, volume = {5}, year = {2022}, } @article{9199, abstract = {We associate a certain tensor product lattice to any primitive integer lattice and ask about its typical shape. These lattices are related to the tangent bundle of Grassmannians and their study is motivated by Peyre's programme on "freeness" for rational points of bounded height on Fano varieties.}, author = {Browning, Timothy D and Horesh, Tal and Wilsch, Florian Alexander}, issn = {1944-7833}, journal = {Algebra & Number Theory}, number = {10}, pages = {2385--2407}, publisher = {Mathematical Sciences Publishers}, title = {{Equidistribution and freeness on Grassmannians}}, doi = {10.2140/ant.2022.16.2385}, volume = {16}, year = {2022}, } @article{9311, abstract = {Partially observable Markov decision processes (POMDPs) are standard models for dynamic systems with probabilistic and nondeterministic behaviour in uncertain environments. We prove that in POMDPs with long-run average objective, the decision maker has approximately optimal strategies with finite memory. This implies notably that approximating the long-run value is recursively enumerable, as well as a weak continuity property of the value with respect to the transition function. }, author = {Chatterjee, Krishnendu and Saona Urmeneta, Raimundo J and Ziliotto, Bruno}, issn = {1526-5471}, journal = {Mathematics of Operations Research}, keywords = {Management Science and Operations Research, General Mathematics, Computer Science Applications}, number = {1}, pages = {100--119}, publisher = {Institute for Operations Research and the Management Sciences}, title = {{Finite-memory strategies in POMDPs with long-run average objectives}}, doi = {10.1287/moor.2020.1116}, volume = {47}, year = {2022}, } @article{9364, abstract = {Let t : Fp → C be a complex valued function on Fp. A classical problem in analytic number theory is bounding the maximum M(t) := max 0≤H

0 there exists a large subset of a ∈ F×p such that for kl a,1,p : x → e((ax+x) / p) we have M(kla,1,p) ≥ (1−ε/√2π + o(1)) log log p, as p→∞. Finally, we prove a result on the growth of the moments of {M (kla,1,p)}a∈F×p. 2020 Mathematics Subject Classification: 11L03, 11T23 (Primary); 14F20, 60F10 (Secondary).}, author = {Bonolis, Dante}, issn = {1469-8064}, journal = {Mathematical Proceedings of the Cambridge Philosophical Society}, number = {3}, pages = {563 -- 590}, publisher = {Cambridge University Press}, title = {{On the size of the maximum of incomplete Kloosterman sums}}, doi = {10.1017/S030500412100030X}, volume = {172}, year = {2022}, } @article{9649, abstract = {Isomanifolds are the generalization of isosurfaces to arbitrary dimension and codimension, i.e. manifolds defined as the zero set of some multivariate vector-valued smooth function f : Rd → Rd−n. A natural (and efficient) way to approximate an isomanifold is to consider its Piecewise-Linear (PL) approximation based on a triangulation T of the ambient space Rd. In this paper, we give conditions under which the PL-approximation of an isomanifold is topologically equivalent to the isomanifold. The conditions are easy to satisfy in the sense that they can always be met by taking a sufficiently fine triangulation T . This contrasts with previous results on the triangulation of manifolds where, in arbitrary dimensions, delicate perturbations are needed to guarantee topological correctness, which leads to strong limitations in practice. We further give a bound on the Fréchet distance between the original isomanifold and its PL-approximation. Finally we show analogous results for the PL-approximation of an isomanifold with boundary.}, author = {Boissonnat, Jean-Daniel and Wintraecken, Mathijs}, issn = {1615-3383}, journal = {Foundations of Computational Mathematics }, pages = {967--1012}, publisher = {Springer Nature}, title = {{The topological correctness of PL approximations of isomanifolds}}, doi = {10.1007/s10208-021-09520-0}, volume = {22}, year = {2022}, } @article{10016, abstract = {Auxin has always been at the forefront of research in plant physiology and development. Since the earliest contemplations by Julius von Sachs and Charles Darwin, more than a century-long struggle has been waged to understand its function. This largely reflects the failures, successes, and inevitable progress in the entire field of plant signaling and development. Here I present 14 stations on our long and sometimes mystical journey to understand auxin. These highlights were selected to give a flavor of the field and to show the scope and limits of our current knowledge. A special focus is put on features that make auxin unique among phytohormones, such as its dynamic, directional transport network, which integrates external and internal signals, including self-organizing feedback. Accented are persistent mysteries and controversies. The unexpected discoveries related to rapid auxin responses and growth regulation recently disturbed our contentment regarding understanding of the auxin signaling mechanism. These new revelations, along with advances in technology, usher us into a new, exciting era in auxin research. }, author = {Friml, Jiří}, issn = {1943-0264}, journal = {Cold Spring Harbor Perspectives in Biology}, number = {5}, publisher = {Cold Spring Harbor Laboratory}, title = {{Fourteen stations of auxin}}, doi = {10.1101/cshperspect.a039859 }, volume = {14}, year = {2022}, } @article{10018, abstract = {In order to study integral points of bounded log-anticanonical height on weak del Pezzo surfaces, we classify weak del Pezzo pairs. As a representative example, we consider a quartic del Pezzo surface of singularity type A1 + A3 and prove an analogue of Manin's conjecture for integral points with respect to its singularities and its lines.}, author = {Derenthal, Ulrich and Wilsch, Florian Alexander}, issn = {1475-3030 }, journal = {Journal of the Institute of Mathematics of Jussieu}, keywords = {Integral points, del Pezzo surface, universal torsor, Manin’s conjecture}, publisher = {Cambridge University Press}, title = {{Integral points on singular del Pezzo surfaces}}, doi = {10.1017/S1474748022000482}, year = {2022}, } @article{10042, abstract = {SnSe has emerged as one of the most promising materials for thermoelectric energy conversion due to its extraordinary performance in its single-crystal form and its low-cost constituent elements. However, to achieve an economic impact, the polycrystalline counterpart needs to replicate the performance of the single crystal. Herein, we optimize the thermoelectric performance of polycrystalline SnSe produced by consolidating solution-processed and surface-engineered SnSe particles. In particular, the SnSe particles are coated with CdSe molecular complexes that crystallize during the sintering process, forming CdSe nanoparticles. The presence of CdSe nanoparticles inhibits SnSe grain growth during the consolidation step due to Zener pinning, yielding a material with a high density of grain boundaries. Moreover, the resulting SnSe–CdSe nanocomposites present a large number of defects at different length scales, which significantly reduce the thermal conductivity. The produced SnSe–CdSe nanocomposites exhibit thermoelectric figures of merit up to 2.2 at 786 K, which is among the highest reported for solution-processed SnSe.}, author = {Liu, Yu and Calcabrini, Mariano and Yu, Yuan and Lee, Seungho and Chang, Cheng and David, Jérémy and Ghosh, Tanmoy and Spadaro, Maria Chiara and Xie, Chenyang and Cojocaru-Mirédin, Oana and Arbiol, Jordi and Ibáñez, Maria}, issn = {1936-086X}, journal = {ACS Nano}, keywords = {tin selenide, nanocomposite, grain growth, Zener pinning, thermoelectricity, annealing, solution processing}, number = {1}, pages = {78--88}, publisher = {American Chemical Society }, title = {{Defect engineering in solution-processed polycrystalline SnSe leads to high thermoelectric performance}}, doi = {10.1021/acsnano.1c06720}, volume = {16}, year = {2022}, } @article{10208, abstract = {It is practical to collect a huge amount of movement data and environmental context information along with the health signals of individuals because there is the emergence of new generations of positioning and tracking technologies and rapid advancements of health sensors. The study of the relations between these datasets and their sequence similarity analysis is of interest to many applications such as health monitoring and recommender systems. However, entering all movement parameters and health signals can lead to the complexity of the problem and an increase in its computational load. In this situation, dimension reduction techniques can be used to avoid consideration of simultaneous dependent parameters in the process of similarity measurement of the trajectories. The present study provides a framework, named CaDRAW, to use spatial–temporal data and movement parameters along with independent context information in the process of measuring the similarity of trajectories. In this regard, the omission of dependent movement characteristic signals is conducted by using an unsupervised feature selection dimension reduction technique. To evaluate the effectiveness of the proposed framework, it was applied to a real contextualized movement and related health signal datasets of individuals. The results indicated the capability of the proposed framework in measuring the similarity and in decreasing the characteristic signals in such a way that the similarity results -before and after reduction of dependent characteristic signals- have small differences. The mean differences between the obtained results before and after reducing the dimension were 0.029 and 0.023 for the round path, respectively.}, author = {Goudarzi, Samira and Sharif, Mohammad and Karimipour, Farid}, issn = {1868-5145}, journal = {Journal of Ambient Intelligence and Humanized Computing}, keywords = {general computer science}, pages = {2621–2635}, publisher = {Springer Nature}, title = {{A context-aware dimension reduction framework for trajectory and health signal analyses}}, doi = {10.1007/s12652-021-03569-z}, volume = {13}, year = {2022}, } @article{10284, abstract = {Infections early in life can have enduring effects on an organism's development and immunity. In this study, we show that this equally applies to developing ‘superorganisms’––incipient social insect colonies. When we exposed newly mated Lasius niger ant queens to a low pathogen dose, their colonies grew more slowly than controls before winter, but reached similar sizes afterwards. Independent of exposure, queen hibernation survival improved when the ratio of pupae to workers was small. Queens that reared fewer pupae before worker emergence exhibited lower pathogen levels, indicating that high brood rearing efforts interfere with the ability of the queen's immune system to suppress pathogen proliferation. Early-life queen pathogen exposure also improved the immunocompetence of her worker offspring, as demonstrated by challenging the workers to the same pathogen a year later. Transgenerational transfer of the queen's pathogen experience to her workforce can hence durably reduce the disease susceptibility of the whole superorganism.}, author = {Casillas Perez, Barbara E and Pull, Christopher and Naiser, Filip and Naderlinger, Elisabeth and Matas, Jiri and Cremer, Sylvia}, issn = {1461-0248}, journal = {Ecology Letters}, number = {1}, pages = {89--100}, publisher = {Wiley}, title = {{Early queen infection shapes developmental dynamics and induces long-term disease protection in incipient ant colonies}}, doi = {10.1111/ele.13907}, volume = {25}, year = {2022}, } @article{10335, abstract = {Van der Holst and Pendavingh introduced a graph parameter σ, which coincides with the more famous Colin de Verdière graph parameter μ for small values. However, the definition of a is much more geometric/topological directly reflecting embeddability properties of the graph. They proved μ(G) ≤ σ(G) + 2 and conjectured σ(G) ≤ σ(G) for any graph G. We confirm this conjecture. As far as we know, this is the first topological upper bound on σ(G) which is, in general, tight. Equality between μ and σ does not hold in general as van der Holst and Pendavingh showed that there is a graph G with μ(G) ≤ 18 and σ(G) ≥ 20. We show that the gap appears at much smaller values, namely, we exhibit a graph H for which μ(H) ≥ 7 and σ(H) ≥ 8. We also prove that, in general, the gap can be large: The incidence graphs Hq of finite projective planes of order q satisfy μ(Hq) ∈ O(q3/2) and σ(Hq) ≥ q2.}, author = {Kaluza, Vojtech and Tancer, Martin}, issn = {0209-9683}, journal = {Combinatorica}, pages = {1317--1345}, publisher = {Springer Nature}, title = {{Even maps, the Colin de Verdière number and representations of graphs}}, doi = {10.1007/s00493-021-4443-7}, volume = {42}, year = {2022}, } @article{10364, abstract = {This paper characterizes the latency of the simplified successive-cancellation (SSC) decoding scheme for polar codes under hardware resource constraints. In particular, when the number of processing elements P that can perform SSC decoding operations in parallel is limited, as is the case in practice, the latency of SSC decoding is O(N1-1/μ + N/P log2 log2 N/P), where N is the block length of the code and μ is the scaling exponent of the channel. Three direct consequences of this bound are presented. First, in a fully-parallel implementation where P = N/2, the latency of SSC decoding is O(N1-1/μ), which is sublinear in the block length. This recovers a result from our earlier work. Second, in a fully-serial implementation where P = 1, the latency of SSC decoding scales as O(N log2 log2 N). The multiplicative constant is also calculated: we show that the latency of SSC decoding when P = 1 is given by (2 + o(1))N log2 log2 N. Third, in a semi-parallel implementation, the smallest P that gives the same latency as that of the fully-parallel implementation is P = N1/μ. The tightness of our bound on SSC decoding latency and the applicability of the foregoing results is validated through extensive simulations.}, author = {Hashemi, Seyyed Ali and Mondelli, Marco and Fazeli, Arman and Vardy, Alexander and Cioffi, John and Goldsmith, Andrea}, issn = {1558-2248}, journal = {IEEE Transactions on Wireless Communications}, number = {6}, pages = {3909--3920}, publisher = {Institute of Electrical and Electronics Engineers}, title = {{Parallelism versus latency in simplified successive-cancellation decoding of polar codes}}, doi = {10.1109/TWC.2021.3125626}, volume = {21}, year = {2022}, } @article{10411, abstract = {The phytohormone auxin is the major growth regulator governing tropic responses including gravitropism. Auxin build-up at the lower side of stimulated shoots promotes cell expansion, whereas in roots it inhibits growth, leading to upward shoot bending and downward root bending, respectively. Yet it remains an enigma how the same signal can trigger such opposite cellular responses. In this review, we discuss several recent unexpected insights into the mechanisms underlying auxin regulation of growth, challenging several existing models. We focus on the divergent mechanisms of apoplastic pH regulation in shoots and roots revisiting the classical Acid Growth Theory and discuss coordinated involvement of multiple auxin signaling pathways. From this emerges a more comprehensive, updated picture how auxin regulates growth.}, author = {Li, Lanxin and Gallei, Michelle C and Friml, Jiří}, issn = {1360-1385}, journal = {Trends in Plant Science}, number = {5}, pages = {440--449}, publisher = {Cell Press}, title = {{Bending to auxin: Fast acid growth for tropisms}}, doi = {10.1016/j.tplants.2021.11.006}, volume = {27}, year = {2022}, } @article{10530, abstract = {Cell dispersion from a confined area is fundamental in a number of biological processes, including cancer metastasis. To date, a quantitative understanding of the interplay of single cell motility, cell proliferation, and intercellular contacts remains elusive. In particular, the role of E- and N-Cadherin junctions, central components of intercellular contacts, is still controversial. Combining theoretical modeling with in vitro observations, we investigate the collective spreading behavior of colonies of human cancer cells (T24). The spreading of these colonies is driven by stochastic single-cell migration with frequent transient cell-cell contacts. We find that inhibition of E- and N-Cadherin junctions decreases colony spreading and average spreading velocities, without affecting the strength of correlations in spreading velocities of neighboring cells. Based on a biophysical simulation model for cell migration, we show that the behavioral changes upon disruption of these junctions can be explained by reduced repulsive excluded volume interactions between cells. This suggests that in cancer cell migration, cadherin-based intercellular contacts sharpen cell boundaries leading to repulsive rather than cohesive interactions between cells, thereby promoting efficient cell spreading during collective migration. }, author = {Zisis, Themistoklis and Brückner, David and Brandstätter, Tom and Siow, Wei Xiong and d’Alessandro, Joseph and Vollmar, Angelika M. and Broedersz, Chase P. and Zahler, Stefan}, issn = {0006-3495}, journal = {Biophysical Journal}, keywords = {Biophysics}, number = {1}, pages = {P44--60}, publisher = {Elsevier}, title = {{Disentangling cadherin-mediated cell-cell interactions in collective cancer cell migration}}, doi = {10.1016/j.bpj.2021.12.006}, volume = {121}, year = {2022}, }