United States

Game theory has been increasingly applied in settings where the game is not known outright, but has to be estimated by sampling.
For example, meta-games that arise in multi-agent evaluation can only be accessed by running a succession of expensive experiments that may involve simultaneous deployment of several agents.
In this paper, we focus on Î±-rank, a popular game-theoretic solution concept designed to perform well in such scenarios.
We aim to estimate the Î±-rank of the game using as few samples as possible.
Our algorithm maximizes information gain between an epistemic belief over the Î±-ranks and the observed payoff. This approach has two main benefits. 
First, it allows us to focus our sampling on the entries that matter the most for identifying the Î±-rank. Second, the Bayesian formulation provides a facility to build in modeling assumptions by using a prior over game payoffs.
We show the benefits of using information gain as compared to the confidence interval criterion of ResponseGraphUCB, 
and provide theoretical results justifying our method.

AAAI 2021

Estimating alpha-Rank by Maximizing Information Gain

The purpose of the AAAI conference is to promote research in artificial intelligence (AI) and scientific exchange among AI researchers, practitioners, scientists, and engineers in affiliated disciplines. 

technical paper

This paper deals with distributed reinforcement learning problems with safety constraints. In particular, we consider that a team of agents cooperate in a shared environment, where each agent has its individual reward function and safety constraints that involve all agents' joint actions. As such, the agents aim to maximize the team-average long-term return, subject to all the safety constraints. More intriguingly, no central controller is assumed to coordinate the agents, and both the rewards and constraints are only known to each agent locally/privately. Instead, the agents are connected by a peer-to-peer communication network to share information with their neighbors. In this work, we first formulate this problem as a distributed constrained Markov decision process (D-CMDP) with networked agents. Then, we propose a decentralized policy gradient (PG) method, Safe Dec-PG, to perform policy optimization based on this D-CMDP model over a network. Convergence guarantees, together with numerical results, showcase the superiority of the proposed algorithm. To the best of our knowledge, this is the first decentralized PG algorithm that accounts for the coupled safety constraints with a quantifiable convergence rate in multi-agent reinforcement learning. Finally, we emphasize that our algorithm is also novel in solving a class of decentralized stochastic nonconvex-concave minimax optimization problems, where both the algorithm design and corresponding theoretical analysis are of independent interest.

Decentralized Policy Gradient Descent Ascent for Safe Multi-Agent Reinforcement Learning

We study the problem of differentially private constrained maximization of decomposable submodular functions. A submodular function is decomposable if it takes the form of a sum of submodular functions. The special case of maximizing a monotone, decomposable submodular function under cardinality constraints is known as the Combinatorial Public Projects (CPP) problem (Papadimitriou, Schapira, and Singer 2008). Previous work by Gupta et al. (2010) gave a differentially private algorithm for the CPP problem. 
We extend this work by designing differentially private algorithms for both monotone and non-monotone decomposable submodular maximization under general matroid constraints, with competitive utility guarantees. We complement our theoretical bounds with experiments demonstrating improved empirical performance.

Differentially Private Decomposable Submodular Maximization

We address the problem of crowd localization, i.e., the prediction of dots corresponding to people in a crowded scene. Due to various challenges, a localization method is prone to spatial semantic errors, i.e., predicting multiple dots within a same person or collapsing multiple dots in a cluttered region. We propose a topological approach targeting these semantic errors. We introduce a topological constraint that teaches the model to reason about the spatial arrangement of dots. To enforce this constraint, we define a persistence loss based on the theory of persistent homology. The loss compares the topographic landscape of the likelihood map and the topology of the ground truth. Topological reasoning improves the quality of the localization algorithm especially near cluttered regions. On multiple public benchmarks, our method outperforms previous localization methods. Additionally, we demonstrate the potential of our method in improving the performance in the crowd counting task.

Localization in the Crowd with Topological Constraints

Analysing and computing with Gaussian processes arising from infinitely wide neural networks has recently seen a resurgence in popularity. Despite this, many explicit covariance functions of networks with activation functions used in modern networks remain unknown. Furthermore, while the kernels of deep networks can be computed iteratively, theoretical understanding of deep kernels is lacking, particularly with respect to fixed-point dynamics. Firstly, we derive the covariance functions of multi-layer perceptrons (MLPs) with exponential linear units (ELU) and Gaussian error linear units (GELU) and evaluate the performance of the limiting Gaussian processes on some benchmarks. Secondly, and more generally, we analyse the fixed-point dynamics of iterated kernels corresponding to a broad range of activation functions. We find that unlike some previously studied neural network kernels, these new kernels exhibit non-trivial fixed-point dynamics which are mirrored in finite-width neural networks. The fixed point behaviour present in some networks explains a mechanism for implicit regularisation in overparameterised deep models. Our results relate to both the static iid parameter conjugate kernel and the dynamic neural tangent kernel constructions

Avoiding Kernel Fixed Points: Computing with ELU and GELU Infinite Networks

Propositional model counting, or #SAT, is the problem of computing the number of satisfying assignments of a Boolean formula. Many problems from different application areas, including many discrete probabilistic inference problems, can be translated into model counting problems to be solved by #SAT solvers. Exact #SAT solvers, however, are often not scalable to industrial size instances. In this paper, we present Neuro#, an approach for learning branching heuristics to improve the performance of exact #SAT solvers on instances from a given family of problems. We experimentally show that our method reduces the step count
on similarly distributed held-out instances and generalizes to much larger instances from the same problem family. It is able to achieve these results on a number of different problem families having very different structures.
In addition to step count improvements, Neuro# can also achieve orders of magnitude wall-clock speedups over the vanilla solver on larger instances in some problem families, despite the runtime overhead of querying the model.

Learning Branching Heuristics for Propositional Model Counting

Slides are commonly used to present information and tell stories. In academic and research communities, slides are typically used to summarize findings in accepted papers for presentation in meetings and conferences.
These slides for academic papers usually contain common and essential topics such as major contributions, model design, experiment details and future work. In this paper, we aim to automatically generate slides for academic papers. We first conducted an in-depth analysis of how humans create slides. We then mined frequently used slide topics. Given a topic, our approach extracts relevant sentences in the paper to provide the draft slides. Due to the lack of labeling data, we integrate prior knowledge of ground truth sentences into a log-linear model to create an initial pseudo-target distribution. Two sentence extractors are learned collaboratively and bootstrap the performance of each other. Evaluation results on a labeled test set show that our model can extract more relevant sentences than baseline methods. Human evaluation also shows slides generated by our model can serve as a good basis for preparing the final presentations.

Towards Topic-Aware Slide Generation for Academic Papers with Unsupervised Mutual Learning

Retrosynthesis is the task of predicting reactant molecules from a given product molecule and is, important in organic chemistry because the identification of a synthetic path is as demanding as the discovery of new chemical compounds. Recently, the retrosynthesis task has been solved automatically without human expertise using powerful deep learning models. Recent deep models are primarily based on seq2seq or graph neural networks depending on the function of molecular representation, sequence, or graph. Current state-of-the-art models represent a molecule as a graph, but they require joint training with auxiliary prediction tasks, such as the most probable reaction template or reaction center prediction. Furthermore, they require additional labels by experienced chemists, thereby incurring additional cost. Herein, we propose a novel template-free model, i.e., Graph Truncated Attention (GTA), which leverages both sequence and graph representations by inserting graphical information into a seq2seq model. The proposed GTA model masks the self-attention layer using the adjacency matrix of product molecule in the encoder and applies a new loss using atom mapping acquired from an automated algorithm to the cross-attention layer in the decoder. Our model achieves new state-of-the-art records, i.e., exact match top-1 and top-10 accuracies of 51.1% and 81.6% on the USPTO-50k benchmark dataset, respectively, and 46.0% and 70.0% on the USPTO-full dataset, respectively, both without any reaction class information. The GTA model surpasses prior graph-based template-free models by 2% and 7% in terms of the top-1 and top-10 accuracies on the USPTO-50k dataset, respectively, and by over 6% for both the top-1 and top-10 accuracies on the USPTO-full dataset.

GTA: Graph Truncated Attention for Retrosynthesis

We study the sample complexity of teaching, termed as ``teaching dimension" (TDim) in the literature, for the teaching-by-reinforcement paradigm, where the teacher guides the student through rewards. This is distinct from the teaching-by-demonstration paradigm motivated by robotics applications, where the teacher teaches by providing demonstrations of state/action trajectories. The teaching-by-reinforcement paradigm applies to a wider range of real-world settings where a demonstration is inconvenient, but has not been studied systematically.  In this paper, we focus on a specific family of reinforcement learning algorithms, Q-learning, and characterize the TDim under different teachers with varying control power over the environment, and present matching optimal teaching algorithms. Our TDim results provide the minimum number of samples needed for reinforcement learning, and we discuss their connections to standard PAC-style RL sample complexity and teaching-by-demonstration sample complexity results. Our teaching algorithms have the potential to speed up RL agent learning in applications where a helpful teacher is available.

The Sample Complexity of Teaching by Reinforcement on Q-Learning

Single-view 3D object reconstruction is a challenging fundamental problem in machine perception, largely due to the morphological diversity of objects in the natural world. In particular, high curvature regions are not always represented accurately by methods trained with common set-based loss functions such as Chamfer Distance, resulting in reconstructions short-circuiting the surface or "cutting corners." To address this issue, we propose an approach to 3D reconstruction that embeds points on the surface of an object into a higher-dimensional space that captures both the original 3D surface as well as geodesic distances between points on the surface of the object. The precise specification of these additional "lifted" coordinates ultimately yields useful surface information without requiring excessive additional computation during either training or testing, in comparison with existing approaches. Our experiments show that taking advantage of these learned lifted coordinates yields better performance for estimating surface normals and generating surfaces than using point cloud reconstructions alone. Further, we find that this learned geodesic embedding space provides useful information for applications such as unsupervised object decomposition.

Geodesic-HOF: 3D Reconstruction without Cutting Corners

We introduce Wasserstein Adversarial Proximal Policy Optimization (WAPPO), a novel algorithm for visual transfer in Reinforcement Learning that explicitly learns to align the distributions of extracted features between a source and target task. WAPPO approximates and minimizes the Wasserstein-1 distance between the distributions of features from source and target domains via a novel Wasserstein Confusion objective. WAPPO outperforms the prior state-of-the-art in visual transfer and successfully transfers policies across Visual Cartpole and both the easy and hard settings of of 16 OpenAI Procgen environments.

Downloads

Next from AAAI 2021

Decentralized Policy Gradient Descent Ascent for Safe Multi-Agent Reinforcement Learning

Similar lecture

On Convergence of Gradient Expected Sarsa (lambda)

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAAI 2021

Decentralized Policy Gradient Descent Ascent for Safe Multi-Agent Reinforcement Learning

Similar lecture

On Convergence of Gradient Expected Sarsa (lambda)

Downloads