Search

We consider deep deterministic policy gradient (DDPG) in the context of reinforcement learning with sparse rewards. To enhance exploration, we introduce a search procedure, \emph{${\epsilon}{t}$-greedy}, which generates exploratory options for exploring less-visited states. We prove that search using $\epsilon t$-greedy has polynomial sample complexity under mild MDP assumptions. To more efficiently use the information provided by rewarded transitions, we develop a new dual experience replay buffer framework, \emph{GDRB}, and implement \emph{longest n-step returns}. The resulting algorithm, \emph{ETGL-DDPG}, integrates all three techniques: \bm{$\epsilon t$}-greedy, \textbf{G}DRB, and \textbf{L}ongest $n$-step, into DDPG. We evaluate ETGL-DDPG on standard benchmarks and demonstrate that it outperforms DDPG, as well as other state-of-the-art methods, across all tested sparse-reward continuous environments. Ablation studies further highlight how each strategy individually enhances the performance of DDPG in this setting.

A common problem in contextual preference ranking is that a single preferred action is compared against several choices, thereby blowing up the complexity and skewing the preference distribution. In this work, we show how one can solve this problem via a suitable adaptation of the CLIP framework.This adaptation is not entirely straight-forward, because although the InfoNCE loss used by CLIP has achieved great success in computer vision and multi-modal domains, its batch-construction technique requires the ability to compare arbitrary items, and is not well-defined if one item has multiple positive associations in the same batch. We empirically demonstrate the utility of our adapted version of the InfoNCE loss in the domain of collectable card games, where we aim to learn an embedding space that captures the associations between single cards and whole card pools based on human selections. Such selection data only exists for restricted choices, thus generating concrete preferences of one item over a set of other items rather than a perfect fit between the card and the pool. Our results show that vanilla CLIP does not perform well due to the aforementioned intuitive issues. However, by adapting CLIP to the problem, we receive a model outperforming previous work trained with the triplet loss, while also alleviating problems associated with mining triplets.

A defining feature of collectable card games is the deck building process prior to actual gameplay, in which players form their decks according to some restrictions. Learning to build decks is difficult for players and models alike due to the large card variety and highly complex semantics, as well as requiring meaningful card and deck representations when aiming to utilise AI. In addition, regular releases of new card sets lead to unforeseeable fluctuations in the available card pool, thus affecting possible deck configurations and requiring continuous updates. Previous Game AI approaches to building decks have often been limited to fixed sets of possible cards, which greatly limits their utility in practice. In this work, we explore possible card representations that generalise to unseen cards, thus greatly extending the real-world utility of AI-based deck building for the game "Magic: The Gathering".We study such representations based on numerical, nominal, and text-based features of cards, card images, and meta information about card usage from third-party services. Our results show that while the particular choice of generalised input representation has little effect on learning to predict human card selections among known cards, the performance on new, unseen cards can be greatly improved. Our generalised model is able to predict 55\% of human choices on completely unseen cards, thus showing a deep understanding of card quality and strategy., Comment: Best paper award nominee at IEEE Conference on Games 2024

In imperfect information games, the evaluation of a game state not only depends on the observable world but also relies on hidden parts of the environment. As accessing the obstructed information trivialises state evaluations, one approach to tackle such problems is to estimate the value of the imperfect state as a combination of all states in the information set, i.e., all possible states that are consistent with the current imperfect information. In this work, the goal is to learn a function that maps from the imperfect game information state to its expected value. However, constructing a perfect training set, i.e. an enumeration of the whole information set for numerous imperfect states, is often infeasible. To compute the expected values for an imperfect information game like \textit{Reconnaissance Blind Chess}, one would need to evaluate thousands of chess positions just to obtain the training target for a single state. Still, the expected value of a state can already be approximated with appropriate accuracy from a much smaller set of evaluations. Thus, in this paper, we empirically investigate how a budget of perfect information game evaluations should be distributed among training samples to maximise the return. Our results show that sampling a small number of states, in our experiments roughly 3, for a larger number of separate positions is preferable over repeatedly sampling a smaller quantity of states. Thus, we find that in our case, the quantity of different samples seems to be more important than higher target quality., Comment: KI 2024 - 47th German Conference on Artificial Intelligence

In imperfect information games, the game state is generally not fully observable to players. Therefore, good gameplay requires policies that deal with the different information that is hidden from each player. To combat this, effective algorithms often reason about information sets; the sets of all possible game states that are consistent with a player's observations. While there is no way to distinguish between the states within an information set, this property does not imply that all states are equally likely to occur in play. We extend previous research on assigning weights to the states in an information set in order to facilitate better gameplay in the imperfect information game of Reconnaissance Blind Chess. For this, we train two different neural networks which estimate the likelihood of each state in an information set from historical game data. Experimentally, we find that a Siamese neural network is able to achieve higher accuracy and is more efficient than a classical convolutional neural network for the given domain. Finally, we evaluate an RBC-playing agent that is based on the generated weightings and compare different parameter settings that influence how strongly it should rely on them. The resulting best player is ranked 5th on the public leaderboard., Comment: Extended version of IEEE Conference on Games 2023 paper

We propose Expected Work Search (EWS), a new game solving algorithm. EWS combines win rate estimation, as used in Monte Carlo Tree Search, with proof size estimation, as used in Proof Number Search. The search efficiency of EWS stems from minimizing a novel notion of Expected Work, which predicts the expected computation required to solve a position. EWS outperforms traditional solving algorithms on the games of Go and Hex. For Go, we present the first solution to the empty 5x5 board with the commonly used positional superko ruleset. For Hex, our algorithm solves the empty 8x8 board in under 4 minutes. Experiments show that EWS succeeds both with and without extensive domain-specific knowledge.

Monte Carlo Tree Search (MCTS) is an immensely popular search-based framework used for decision making. It is traditionally applied to domains where a perfect simulation model of the environment is available. We study and improve MCTS in the context where the environment model is given but imperfect. We show that the discrepancy between the model and the actual environment can lead to significant performance degradation with standard MCTS. We therefore develop Uncertainty Adapted MCTS (UA-MCTS), a more robust algorithm within the MCTS framework. We estimate the transition uncertainty in the given model, and direct the search towards more certain transitions in the state space. We modify all four MCTS phases to improve the search behavior by considering these estimates. We prove, in the corrupted bandit case, that adding uncertainty information to adapt UCB leads to tighter regret bound than standard UCB. Empirically, we evaluate UA-MCTS and its individual components on the deterministic domains from the MinAtar test suite. Our results demonstrate that UA-MCTS strongly improves MCTS in the presence of model transition errors.

In this work, we adapt a training approach inspired by the original AlphaGo system to play the imperfect information game of Reconnaissance Blind Chess. Using only the observations instead of a full description of the game state, we first train a supervised agent on publicly available game records. Next, we increase the performance of the agent through self-play with the on-policy reinforcement learning algorithm Proximal Policy Optimization. We do not use any search to avoid problems caused by the partial observability of game states and only use the policy network to generate moves when playing. With this approach, we achieve an ELO of 1330 on the RBC leaderboard, which places our agent at position 27 at the time of this writing. We see that self-play significantly improves performance and that the agent plays acceptably well without search and without making assumptions about the true game state., Comment: 4 Pages, IEEE Conference on Games 2022 short paper

In this paper, we study learning in probabilistic domains where the learner may receive incorrect labels but can improve the reliability of labels by repeatedly sampling them. In such a setting, one faces the problem of whether the fixed budget for obtaining training examples should rather be used for obtaining all different examples or for improving the label quality of a smaller number of examples by re-sampling their labels. We motivate this problem in an application to compare the strength of poker hands where the training signal depends on the hidden community cards, and then study it in depth in an artificial setting where we insert controlled noise levels into the MNIST database. Our results show that with increasing levels of noise, resampling previous examples becomes increasingly more important than obtaining new examples, as classifier performance deteriorates when the number of incorrect labels is too high. In addition, we propose two different validation strategies; switching from lower to higher validations over the course of training and using chi-square statistics to approximate the confidence in obtained labels., Comment: Preliminary work under review for ICML 2022

Scaling up the size and training of autoregressive language models has enabled novel ways of solving Natural Language Processing tasks using zero-shot and few-shot learning. While extreme-scale language models such as GPT-3 offer multilingual capabilities, zero-shot learning for languages other than English remain largely unexplored. Here, we introduce Cedille, a large open source auto-regressive language model, specifically trained for the French language. Our results show that Cedille outperforms existing French language models and is competitive with GPT-3 on a range of French zero-shot benchmarks. Furthermore, we provide an in-depth comparison of the toxicity exhibited by these models, showing that Cedille marks an improvement in language model safety thanks to dataset filtering., Comment: 8 pages, 1 figure, 7 tables

Bayesian Active Learning (BAL) is an efficient framework for learning the parameters of a model, in which input stimuli are selected to maximize the mutual information between the observations and the unknown parameters. However, the applicability of BAL to experiments is limited as it requires performing high-dimensional integrations and optimizations in real time: current methods are either too time consuming, or only applicable to specific models. Here, we propose an Efficient Sampling-Based Bayesian Active Learning (ESB-BAL) framework, which is efficient enough to be used in real-time biological experiments. We apply our method to the problem of estimating the parameters of a chemical synapse from the postsynaptic responses to evoked presynaptic action potentials. Using synthetic data and synaptic whole-cell patch-clamp recordings, we show that our method can improve the precision of model-based inferences, thereby paving the way towards more systematic and efficient experimental designs in physiology., Comment: Major review after submission: - Change of title - Add biological recordings

In deep reinforcement learning (RL) systems, abnormal states pose significant risks by potentially triggering unpredictable behaviors and unsafe actions, thus impeding the deployment of RL systems in real-world scenarios. It is crucial for reliable decision-making systems to have the capability to cast an alert whenever they encounter unfamiliar observations that they are not equipped to handle. In this paper, we propose a novel Mahalanobis distance-based (MD) anomaly detection framework, called \textit{MDX}, for deep RL algorithms. MDX simultaneously addresses random, adversarial, and out-of-distribution (OOD) state outliers in both offline and online settings. It utilizes Mahalanobis distance within class-conditional distributions for each action and operates within a statistical hypothesis testing framework under the Gaussian assumption. We further extend it to robust and distribution-free versions by incorporating Robust MD and conformal inference techniques. Through extensive experiments on classical control environments, Atari games, and autonomous driving scenarios, we demonstrate the effectiveness of our MD-based detection framework. MDX offers a simple, unified, and practical anomaly detection tool for enhancing the safety and reliability of RL systems in real-world applications., Comment: 19 pages, 21 figures

Materials' microstructures are signatures of their alloying composition and processing history. Therefore, microstructures exist in a wide variety. As materials become increasingly complex to comply with engineering demands, advanced computer vision (CV) approaches such as deep learning (DL) inevitably gain relevance for quantifying microstrucutures' constituents from micrographs. While DL can outperform classical CV techniques for many tasks, shortcomings are poor data efficiency and generalizability across datasets. This is inherently in conflict with the expense associated with annotating materials data through experts and extensive materials diversity. To tackle poor domain generalizability and the lack of labeled data simultaneously, we propose to apply a sub-class of transfer learning methods called unsupervised domain adaptation (UDA). These algorithms address the task of finding domain-invariant features when supplied with annotated source data and unannotated target data, such that performance on the latter distribution is optimized despite the absence of annotations. Exemplarily, this study is conducted on a lath-shaped bainite segmentation task in complex phase steel micrographs. Here, the domains to bridge are selected to be different metallographic specimen preparations (surface etchings) and distinct imaging modalities. We show that a state-of-the-art UDA approach surpasses the na\"ive application of source domain trained models on the target domain (generalization baseline) to a large extent. This holds true independent of the domain shift, despite using little data, and even when the baseline models were pre-trained or employed data augmentation. Through UDA, mIoU was improved over generalization baselines from 82.2%, 61.0%, 49.7% to 84.7%, 67.3%, 73.3% on three target datasets, respectively. This underlines this techniques' potential to cope with materials variance., Comment: 20 pages, 7 figures

In this paper, we study the problem of evaluating the addition of elements to a set. This problem is difficult, because it can, in the general case, not be reduced to unconditional preferences between the choices. Therefore, we model preferences based on the context of the decision. We discuss and compare two different Siamese network architectures for this task: a twin network that compares the two sets resulting after the addition, and a triplet network that models the contribution of each candidate to the existing set. We evaluate the two settings on a real-world task; learning human card preferences for deck building in the collectible card game Magic: The Gathering. We show that the triplet approach achieves a better result than the twin network and that both outperform previous results on this task., Comment: arXiv admin note: substantial text overlap with arXiv:2105.11864

Drafting, i.e., the selection of a subset of items from a larger candidate set, is a key element of many games and related problems. It encompasses team formation in sports or e-sports, as well as deck selection in many modern card games. The key difficulty of drafting is that it is typically not sufficient to simply evaluate each item in a vacuum and to select the best items. The evaluation of an item depends on the context of the set of items that were already selected earlier, as the value of a set is not just the sum of the values of its members - it must include a notion of how well items go together. In this paper, we study drafting in the context of the card game Magic: The Gathering. We propose the use of a contextual preference network, which learns to compare two possible extensions of a given deck of cards. We demonstrate that the resulting network is better able to evaluate card decks in this game than previous attempts., Comment: IEEE Conference on Games 2021 version

Boolean Satisfiability (SAT) is a well-known NP-complete problem. Despite this theoretical hardness, SAT solvers based on Conflict Driven Clause Learning (CDCL) can solve large SAT instances from many important domains. CDCL learns clauses from conflicts, a technique that allows a solver to prune its search space. The selection heuristics in CDCL prioritize variables that are involved in recent conflicts. While only a fraction of decisions generate any conflicts, many generate multiple conflicts. In this paper, we study conflict-generating decisions in CDCL in detail. We investigate the impact of single conflict (sc) decisions, which generate only one conflict, and multi-conflict (mc) decisions which generate two or more. We empirically characterize these two types of decisions based on the quality of the learned clauses produced by each type of decision. We also show an important connection between consecutive clauses learned within the same mc decision, where one learned clause triggers the learning of the next one forming a chain of clauses. This leads to the consideration of similarity between conflicts, for which we formulate the notion of conflictsproximity as a similarity measure. We show that conflicts in mc decisions are more closely related than consecutive conflicts generated from sc decisions. Finally, we develop Common Reason Variable Reduction (CRVR) as a new decision strategy that reduces the selection priority of some variables from the learned clauses of mc decisions. Our empirical evaluation of CRVR implemented in three leading solvers demonstrates performance gains in benchmarks from the main track of SAT Competition-2020.

If we wish to integrate a function $h|\Omega\subset\Re^{n}\to\Re$ along a single $T$-level surface of a function $\psi |\Omega\subset\Re^{n}\to\Re$, then a number of different methods for extracting finite elements appropriate to the dimension of the level surface may be employed to obtain an explicit representation over which the integration may be performed using standard numerical quadrature techniques along each element. However, when the goal is to compute an entire continuous family $m(T)$ of integrals over all the $T$-level surfaces of $\psi$, then this method of explicit level set extraction is no longer practical. We introduce a novel method to perform this type of numerical integration efficiently by making use of the coarea formula. We present the technique for discretization of the coarea formula and present the algorithms to compute the integrals over families of T-level surfaces. While validation of our method in the special case of a single level surface demonstrates accuracies close to more explicit isosurface integration methods, we show a sizable boost in computational efficiency in the case of multiple T-level surfaces, where our coupled integration algorithms significantly outperform sequential one-at-a-time application of explicit methods.

This paper proposes a novel training model based on shape and appearance features for object segmentation in images and videos. Whereas most such models rely on two-dimensional appearance templates or a finite set of descriptors, our appearance-based feature is a one-dimensional function, which is efficiently coupled with the object's shape by integrating intensities along the object's iso-contours. Joint PCA training on these shape and appearance features further exploits shape-appearance correlations and the resulting training model is incorporated in an active-contour-type energy functional for recognition-segmentation tasks. Experiments on synthetic and infrared images demonstrate how this shape and appearance training model improves accuracy compared to methods based on the Chan-Vese energy.

Social media analysis has become a common approach to assess public opinion on various topics, including those about health, in near real-time. The growing volume of social media posts has led to an increased usage of modern machine learning methods in natural language processing. While the rapid dynamics of social media can capture underlying trends quickly, it also poses a technical problem: algorithms trained on annotated data in the past may underperform when applied to contemporary data. This phenomenon, known as concept drift, can be particularly problematic when rapid shifts occur either in the topic of interest itself, or in the way the topic is discussed. Here, we explore the effect of machine learning concept drift by focussing on vaccine sentiments expressed on Twitter, a topic of central importance especially during the COVID-19 pandemic. We show that while vaccine sentiment has declined considerably during the COVID-19 pandemic in 2020, algorithms trained on pre-pandemic data would have largely missed this decline due to concept drift. Our results suggest that social media analysis systems must address concept drift in a continuous fashion in order to avoid the risk of systematic misclassification of data, which is particularly likely during a crisis when the underlying data can change suddenly and rapidly., Comment: 9 pages, 4 figures, 3 pages of SI; Minor correction in Figure 1: Bracket was not visible

COVID-19 represents the most severe global crisis to date whose public conversation can be studied in real time. To do so, we use a data set of over 350 million tweets and retweets posted by over 26 million English speaking Twitter users from January 13 to June 7, 2020. We characterize the retweet network to identify spontaneous clustering of users and the evolution of their interaction over time in relation to the pandemic's emergence. We identify several stable clusters (super-communities), and are able to link them to international groups mainly involved in science and health topics, national elites, and political actors. The science- and health-related super-community received disproportionate attention early on during the pandemic, and was leading the discussion at the time. However, as the pandemic unfolded, the attention shifted towards both national elites and political actors, paralleled by the introduction of country-specific containment measures and the growing politicization of the debate. Scientific super-community remained present in the discussion, but experienced less reach and became more isolated within the network. Overall, the emerging network communities are characterized by an increased self-amplification and polarization. This makes it generally harder for information from international health organizations or scientific authorities to directly reach a broad audience through Twitter for prolonged time. These results may have implications for information dissemination along the unfolding of long-term events like epidemic diseases on a world-wide scale., Comment: 13 pages, LaTeX. Major changes after peer-review rebuttal

The Mu3e experiment aims to find or exclude the lepton flavour violating decay $\mu^+\to e^+e^-e^+$ with a sensitivity of one in 10$^{16}$ muon decays. The first phase of the experiment is currently under construction at the Paul Scherrer Institute (PSI, Switzerland), where beams with up to 10$^8$ muons per second are available. The detector will consist of an ultra-thin pixel tracker made from High-Voltage Monolithic Active Pixel Sensors (HV-MAPS), complemented by scintillating tiles and fibres for precise timing measurements. The experiment produces about 100 Gbit/s of zero-suppressed data which are transported to a filter farm using a network of FPGAs and fast optical links. On the filter farm, tracks and three-particle vertices are reconstructed using highly parallel algorithms running on graphics processing units, leading to a reduction of the data to 100 Mbyte/s for mass storage and offline analysis. The paper introduces the system design and hardware implementation of the Mu3e data acquisition and filter farm., Comment: 8 pages, 9 figures Submitted to IEEE TNS

Timely access to accurate information is crucial during the COVID-19 pandemic. Prompted by key stakeholders' cautioning against an "infodemic", we study information sharing on Twitter from January through May 2020. We observe an overall surge in the volume of general as well as COVID-19-related tweets around peak lockdown in March/April 2020. With respect to engagement (retweets and likes), accounts related to healthcare, science, government and politics received by far the largest boosts, whereas accounts related to religion and sports saw a relative decrease in engagement. While the threat of an "infodemic" remains, our results show that social media also provide a platform for experts and public authorities to be widely heard during a global crisis., Comment: Kristina Gligori\'c, Manoel Horta Ribeiro and Martin M\"uller contributed equally to this work

The morphology of spherically confined flexoelectric fluid membrane vesicles in an external uniform electric field is studied numerically. Due to the deformations induced by the confinement, the membrane becomes polarized resulting in an interaction with the external field. The equilibrium shapes of the vesicle without electric field can be classified in a geometrical phase diagram as a function of scaled area and reduced volume [1,2]. When the area of the membrane is only slightly larger than the area of the confining sphere, a single axisymmetric invagination appears. A non-vanishing electric field induces an additional elongation of the confined vesicle which is either perpendicular or parallel depending on the sign of the electric field parameter. Higher values of the surface area or the electric field parameter can reduce the symmetry of the system leading to more complex folding. We present the resulting shapes and show that transition lines are shifted in the presence of an electric field. The obtained folding patterns could be of interest for biophysical and technological applications alike., Comment: main text: 9 pages, 4 figures supplemental: 5 pages, 2 figures

In this work, we release COVID-Twitter-BERT (CT-BERT), a transformer-based model, pretrained on a large corpus of Twitter messages on the topic of COVID-19. Our model shows a 10-30% marginal improvement compared to its base model, BERT-Large, on five different classification datasets. The largest improvements are on the target domain. Pretrained transformer models, such as CT-BERT, are trained on a specific target domain and can be used for a wide variety of natural language processing tasks, including classification, question-answering and chatbots. CT-BERT is optimised to be used on COVID-19 content, in particular social media posts from Twitter.

Despite its potential to improve sample complexity versus model-free approaches, model-based reinforcement learning can fail catastrophically if the model is inaccurate. An algorithm should ideally be able to trust an imperfect model over a reasonably long planning horizon, and only rely on model-free updates when the model errors get infeasibly large. In this paper, we investigate techniques for choosing the planning horizon on a state-dependent basis, where a state's planning horizon is determined by the maximum cumulative model error around that state. We demonstrate that these state-dependent model errors can be learned with Temporal Difference methods, based on a novel approach of temporally decomposing the cumulative model errors. Experimental results show that the proposed method can successfully adapt the planning horizon to account for state-dependent model accuracy, significantly improving the efficiency of policy learning compared to model-based and model-free baselines.

A state-of-the-art criterion to evaluate the importance of a given learned clause is called Literal Block Distance (LBD) score. It measures the number of distinct decision levels in a given learned clause. The lower the LBD score of a learned clause, the better is its quality. The learned clauses with LBD score of 2, called glue clauses, are known to possess high pruning power which are never deleted from the clause databases of the modern CDCL SAT solvers. In this work, we relate glue clauses to decision variables. We call the variables that appeared in at least one glue clause up to the current search state Glue Variables. We first show experimentally, by running the state-of-the-art CDCL SAT solver MapleLCMDist on benchmarks from SAT Competition-2017 and 2018, that branching decisions with glue variables are categorically more inference and conflict efficient than nonglue variables. Based on this observation, we develop a structure aware CDCL variable bumping scheme, which bumps the activity score of a glue variable based on its appearance count in the glue clauses that are learned so far by the search. Empirical evaluation shows effectiveness of the new method over the main track instances from SAT Competition 2017 and 2018.

While the shape equations describing the equilibrium of an unstretchable thin sheet that is free to bend are known, the boundary conditions that supplement these equations on free edges have remained elusive. Intuitively, unstretchability is captured by a constraint on the metric within the bulk. Naively one would then guess that this constraint is enough to ensure that the deformations determining the boundary conditions on these edges respect the isometry constraint. If matters were this simple, unfortunately, it would imply unbalanced torques (as well as forces) along the edge unless manifestly unphysical constraints are met by the boundary geometry. In this paper we identify the source of the problem: not only the local arc-length but also the geodesic curvature need to be constrained explicitly on all free edges. We derive the boundary conditions which follow. Contrary to conventional wisdom, there is no need to introduce boundary layers. This framework is applied to isolated conical defects, both with deficit as well, but more briefly, as surplus angles. Using these boundary conditions, we show that the lateral tension within a circular cone of fixed radius is equal but opposite to the radial compression, and independent of the deficit angle itself. We proceed to examine the effect of an oblique outer edge on this cone perturbatively demonstrating that both the correction to the geometry as well as the stress distribution in the cone kicks in at second order in the eccentricity of the edge., Comment: 25 pages, 3 figures

We present a numerical study of the time-dependent motion of a two-dimensional vesicle in a channel under an imposed flow. In a Poiseuille flow the shape of the vesicle depends on the flow strength, the mechanical properties of the membrane, and the width of the channel as reported in the past. This study is focused on the centered snaking (CSn) shape, where the vesicle shows an oscillatory motion like a swimmer flagella even though the flow is stationary. We quantify this behavior by the amplitude and frequency of the oscillations of the vesicle's center of mass. We observe regions in parameter space, where the CSn coexists with the parachute or the unconfined slipper. The influence of an amplitude modulation of the imposed flow on the dynamics and shape of the snaking vesicle is also investigated. For large modulation amplitudes transitions to static shapes are observed. A smaller modulation amplitude induces a modulation in amplitude and frequency of the center of mass of the snaking vesicle. In a certain parameter range we find that the center of mass oscillates with a constant envelope indicating the presence of at least two stable states., Comment: 10 pages, 7 figures

In the past decade, tracking health trends using social media data has shown great promise, due to a powerful combination of massive adoption of social media around the world, and increasingly potent hardware and software that enables us to work with these new big data streams. At the same time, many challenging problems have been identified. First, there is often a mismatch between how rapidly online data can change, and how rapidly algorithms are updated, which means that there is limited reusability for algorithms trained on past data as their performance decreases over time. Second, much of the work is focusing on specific issues during a specific past period in time, even though public health institutions would need flexible tools to assess multiple evolving situations in real time. Third, most tools providing such capabilities are proprietary systems with little algorithmic or data transparency, and thus little buy-in from the global public health and research community. Here, we introduce Crowdbreaks, an open platform which allows tracking of health trends by making use of continuous crowdsourced labelling of public social media content. The system is built in a way which automatizes the typical workflow from data collection, filtering, labelling and training of machine learning classifiers and therefore can greatly accelerate the research process in the public health domain. This work introduces the technical aspects of the platform and explores its future use cases.

Intermediate filaments are the least explored among the large cytoskeletal elements. We show here that they display conformational anomalies in narrow microfluidic channels. Their unusual behavior can be understood as the consequence of a previously undetected, large scale helically curved superstructure. Confinement in a channel orders the otherwise soft, strongly fluctuating helical filaments and enhances their structural correlations, giving rise to experimentally detectable, strongly oscillating tangent correlation functions. We propose an explanation for the detected intrinsic curving phenomenon - an elastic shape instability that we call autocoiling. The mechanism involves self-induced filament buckling via a surface stress located at the outside of the cross-section. The results agree with ultrastructural findings and rationalize for the commonly observed looped intermediate filament shapes.

We study the problem of identifying the best action among a set of possible options when the value of each action is given by a mapping from a number of noisy micro-observables in the so-called fixed confidence setting. Our main motivation is the application to the minimax game search, which has been a major topic of interest in artificial intelligence. In this paper we introduce an abstract setting to clearly describe the essential properties of the problem. While previous work only considered a two-move game tree search problem, our abstract setting can be applied to the general minimax games where the depth can be non-uniform and arbitrary, and transpositions are allowed. We introduce a new algorithm (LUCB-micro) for the abstract setting, and give its lower and upper sample complexity results. Our bounds recover some previous results, which were only available in more limited settings, while they also shed further light on how the structure of minimax problems influence sample complexity.

We theoretically study the conformations of a helical semi-flexible filament confined to a flat surface. This squeezed helix exhibits a variety of unexpected shapes resembling circles, waves or spirals depending on the material parameters. We explore the conformation space in detail and show that the shapes can be understood as the mutual elastic interaction of conformational quasi-particles. Our theoretical results are potentially useful to determine the material parameters of such helical filaments in an experimental setting., Comment: 32 pages, 16 figures, 5 tables

We investigate the morphology of a toroidal fluid membrane vesicle confined inside a spherical container. The equilibrium shapes are assembled in a geometrical phase diagram as a function of scaled area and reduced volume of the membrane. For small area the vesicle can adopt its free form. When increasing the area, the membrane cannot avoid contact and touches the confining sphere along a circular contact line, which extends to a zone of contact for higher area. The elastic energies of the equilibrium shapes are compared to those of their confined counterparts of spherical topology to predict under which conditions a topology change is favored energetically., Comment: 16 pages, 7 figures

The demand for a broadband wireless connection is nowadays no longer limited to stationary situations, but also required while traveling. Therefore, there exist combined efforts to provide wireless access also on High Speed Trains (HSTs), in order to add to the attractiveness of this means for transportation. Installing an additional relay on the train, to facilitate the communication, is an approach that has already been extensively discussed in literature. The possibility of a direct communication between the base station and the passenger has been neglected until now, despite it having numerous advantages. Therefore, a comparison between these two opposing approaches is presented in this paper, accompanied by a detailed discussion of the related aspects. The focus is set on the feasibility of the direct link approach, including simulation results. Further technical issues are also presented, especially regarding the interdependencies of the different aspects and providing a view of mobile- and train-operators on the topic., Comment: 6 pages, 3 figures

We consider the axial compression of a thin sheet wrapped around a rigid cylindrical substrate. In contrast to the wrinkling-to-fold transitions exhibited in similar systems, we find that the sheet always buckles into a single symmetric fold, while periodic solutions are unstable. Upon further compression, the solution breaks symmetry and stabilizes into a recumbent fold. Using linear analysis and numerics, we theoretically predict the buckling force and energy as a function of the compressive displacement. We compare our theory to experiments employing cylindrical neoprene sheets and find remarkably good agreement., Comment: 20 pages, 5 figures

We discuss a curious example for the collective mechanical behavior of coupled non-linear monomer units entrapped in a circular filament. Within a simple model we elucidate how multistability of monomer units and exponentially large degeneracy of the filament's ground state emerge as a collective feature of the closed filament. Surprisingly, increasing the monomer frustration, i.e., the bending prestrain within the circular filament, leads to a conformational softening of the system. The phenomenon, that we term polymorphic crunching, is discussed and applied to a possible scenario for membrane tube deformation by switchable dynamin or FtsZ filaments. We find an important role of cooperative inter-unit interaction for efficient ring induced membrane fission.

Tubular lattices are ubiquitous in nature and technology. Microtubules and nanotubes of all kinds act as important pillars of biological cells and the man-made nano-world. We show that when prestress is introduced in such structures, localized conformational quasiparticles emerge and govern the collective shape dynamics of the lattice. When coupled via cooperative interactions these quasiparticles form larger-scale quasipolymer superstructures exhibiting collective dynamic modes and giving rise to a hallmark behavior radically different from semiflexible beams.

In this paper, we describe an approach that enables an autonomous system to infer the semantics of a command (i.e. a symbol sequence representing an action) in terms of the relations between changes in the observations and the action instances. We present a method of how to induce a theory (i.e. a semantic description) of the meaning of a command in terms of a minimal set of background knowledge. The only thing we have is a sequence of observations from which we extract what kinds of effects were caused by performing the command. This way, we yield a description of the semantics of the action and, hence, a definition., Comment: KRR Workshop at ICLP 2013

Steady, dihedrally symmetric patterns with sharp peaks may be observed on a spinning skirt, lagging behind the material flow of the fabric. These qualitative features are captured with a minimal model of traveling waves on an inextensible, flexible, generalized-conical sheet rotating about a fixed axis. Conservation laws are used to reduce the dynamics to a quadrature describing a particle in a three-parameter family of potentials. One parameter is associated with the stress in the sheet, aNoether is the current associated with rotational invariance, and the third is a Rossby number which indicates the relative strength of Coriolis forces. Solutions are quantized by enforcing a topology appropriate to a skirt and a particular choice of dihedral symmetry. A perturbative analysis of nearly axisymmetric cones shows that Coriolis effects are essential in establishing skirt-like solutions. Fully non-linear solutions with three-fold symmetry are presented which bear a suggestive resemblance to the observed patterns., Comment: two additional figures, changes to text throughout. journal version will have a wordier abstract

A flat elastic sheet may contain pointlike conical singularities that carry a metrical "charge" of Gaussian curvature. Adding such elementary defects to a sheet allows one to make many shapes, in a manner broadly analogous to the familiar multipole construction in electrostatics. However, here the underlying field theory is non-linear, and superposition of intrinsic defects is non-trivial as it must respect the immersion of the resulting surface in three dimensions. We consider a "charge-neutral" dipole composed of two conical singularities of opposite sign. Unlike the relatively simple electrostatic case, here there are two distinct stable minima and an infinity of unstable equilibria. We determine the shapes of the minima and evaluate their energies in the thin-sheet regime where bending dominates over stretching. Our predictions are in surprisingly good agreement with experiments on paper sheets., Comment: 20 pages, 5 figures, 2 tables

We study the morphology of a fluid membrane in spherical confinement. When the area of the membrane is slightly larger than the area of the outer container, a single axisymmetric invagination is observed. For higher area, self-contact occurs: the invagination breaks symmetry and deforms into an ellipsoid-like shape connected to its outer part via a small slit. For even higher areas, a second invagination forms inside the original invagination. The folding patterns observed could constitute basic building blocks in the morphogenesis of biological tissues and organelles., Comment: 4 pages, 3 figures, 1 table

The stability of the fundamental defects of an unstretchable flat sheet is examined. This involves expanding the bending energy to second order in deformations about the defect. The modes of deformation occur as eigenstates of a fourth-order linear differential operator. Unstretchability places a global linear constraint on these modes. Conical defects with a surplus angle exhibit an infinite number of states. If this angle is below a critical value, these states possess an n-fold symmetry labeled by an integer, n \geq 2. A nonlinear stability analysis shows that the 2-fold ground state is stable, whereas excited states possess 2(n - 2) unstable modes which come in even and odd pairs., Comment: 25 pages, 6 figures

Particles embedded in a fluctuating interface experience forces and torques mediated by the deformations and by the thermal fluctuations of the medium. Considering a system of two cylinders bound to a fluid membrane we show that the entropic contribution enhances the curvature-mediated repulsion between the two cylinders. This is contrary to the usual attractive Casimir force in the absence of curvature-mediated interactions. For a large distance between the cylinders, we retrieve the renormalization of the surface tension of a flat membrane due to thermal fluctuations., Comment: 11 pages, 5 figures; final version, as appeared in Phys. Rev. E

We investigate the morphology of thin discs and rings growing in the circumferential direction. Recent analytical results suggest that this growth produces symmetric excess cones (e-cones). We study the stability of such solutions considering self-contact and bending stress. We show that, contrary to what was assumed in previous analytical solutions, beyond a critical growth factor, no symmetric \textit{e}-cone solution is energetically minimal any more. Instead, we obtain skewed e-cone solutions having lower energy, characterized by a skewness angle and repetitive spiral winding with increasing growth. These results are generalized to discs with varying thickness and rings with holes of different radii., Comment: 4 pages, 5 figures; New version corresponds to the one published in Phys. Rev. Lett

A growing or shrinking disc will adopt a conical shape, its intrinsic geometry characterized by a surplus angle $se$ at the apex. If growth is slow, the cone will find its equilibrium. Whereas this is trivial if $se <= 0$, the disc can fold into one of a discrete infinite number of states if $se$ is positive. We construct these states in the regime where bending dominates, determine their energies and how stress is distributed in them. For each state a critical value of $se$ is identified beyond which the cone touches itself. Before this occurs, all states are stable; the ground state has two-fold symmetry., Comment: 4 pages, 4 figures, LaTeX, RevTeX style. New version corresponds to the one published in PRL

A variational framework is introduced to describe how a surface bends when it is subject to local constraints on its geometry. This framework is applied to describe the patterns of a folded sheet of paper. The unstretchability of paper implies a constraint on the surface metric; bending is penalized by an energy quadratic in mean curvature. The local Lagrange multipliers enforcing the constraint are identified with a conserved tangential stress that couples to the extrinsic curvature of the sheet. The framework is illustrated by examining the deformation of a flat sheet into a generalized cone., Comment: 20 pages, 1 figure

When a fluid surface adheres to a substrate, the location of the contact line adjusts in order to minimize the overall energy. This adhesion balance implies boundary conditions which depend on the characteristic surface deformation energies. We develop a general geometrical framework within which these conditions can be systematically derived. We treat both adhesion to a rigid substrate as well as adhesion between two fluid surfaces, and illustrate our general results for several important Hamiltonians involving both curvature and curvature gradients. Some of these have previously been studied using very different techniques, others are to our knowledge new. What becomes clear in our approach is that, except for capillary phenomena, these boundary conditions are not the manifestation of a local force balance, even if the concept of surface stress is properly generalized. Hamiltonians containing higher order surface derivatives are not just sensitive to boundary translations but also notice changes in slope or even curvature. Both the necessity and the functional form of the corresponding additional contributions follow readily from our treatment., Comment: 8 pages, 2 figures, LaTeX, RevTeX style

Torques on interfaces can be described by a divergence-free tensor which is fully encoded in the geometry. This tensor consists of two terms, one originating in the couple of the stress, the other capturing an intrinsic contribution due to curvature. In analogy to the description of forces in terms of a stress tensor, the torque on a particle can be expressed as a line integral along any contour surrounding the particle. Interactions between particles mediated by a fluid membrane are studied within this framework. In particular, torque balance places a strong constraint on the shape of the membrane. Symmetric two-particle configurations admit simple analytical expressions which are valid in the fully nonlinear regime; in particular, the problem may be solved exactly in the case of two membrane-bound parallel cylinders. This apparently simple system provides some flavor of the remarkably subtle nonlinear behavior associated with membrane-mediated interactions., Comment: 16 pages, 10 figures, REVTeX4 style. The Gaussian curvature term was included in the membrane Hamiltonian; section II.B was rephrased to smoothen the flow of presentation

Measurements with an atomic force microscope (AFM) offer a direct way to probe elastic properties of lipid bilayer membranes locally: provided the underlying stress-strain relation is known, material parameters such as surface tension or bending rigidity may be deduced. In a recent experiment a pore-spanning membrane was poked with an AFM tip, yielding a linear behavior of the force-indentation curves. A theoretical model for this case is presented here which describes these curves in the framework of Helfrich theory. The linear behavior of the measurements is reproduced if one neglects the influence of adhesion between tip and membrane. Including it via an adhesion balance changes the situation significantly: force-distance curves cease to be linear, hysteresis and nonzero detachment forces can show up. The characteristics of this rich scenario are discussed in detail in this article., Comment: 14 pages, 9 figures, REVTeX4 style. New version corresponds to the one accepted by PRE. The result section is restructured: a comparison to experimental findings is included; the discussion on the influence of adhesion between AFM tip and membrane is extended

Particles bound to an interface interact because they deform its shape. The stresses that result are fully encoded in the geometry and described by a divergence-free surface stress tensor. This stress tensor can be used to express the force on a particle as a line integral along any conveniently chosen closed contour that surrounds the particle. The resulting expression is exact (i.e., free of any "smallness" assumptions) and independent of the chosen surface parametrization. Additional surface degrees of freedom, such as vector fields describing lipid tilt, are readily included in this formalism. As an illustration, we derive the exact force for several important surface Hamiltonians in various symmetric two-particle configurations in terms of the midplane geometry; its sign is evident in certain interesting limits. Specializing to the linear regime, where the shape can be analytically determined, these general expressions yield force-distance relations, several of which have originally been derived by using an energy based approach., Comment: 18 pages, 7 figures, REVTeX4 style; final version, as appeared in Phys. Rev. E. Compared to v2 several minor mistakes, as well as one important minus sign in Eqn. (18a) have been cured. Compared to v1, this version is significantly extended: Lipid tilt degrees of freedom for membranes are included in the stress framework, more technical details are given, estimates for the magnitude of forces are made