This paper addresses exploration in reinforcement learning (RL) in unknown environments with sparse rewards, focusing on maximum entropy exploration. It introduces a game-theoretic algorithm for visitation entropy maximization with improved sample complexity of O(H^3S^2A/ε^2). For trajectory entropy, the paper presents an algorithm with O(poly(S, A, H)/ε) complexity, showing the statistical advantage of regularized MDPs for exploration. Why it matters: The research offers new techniques to reduce the sample complexity of RL, potentially enhancing the efficiency of AI agents in complex environments.
This paper introduces Diffusion-BBO, a new online black-box optimization (BBO) framework that uses a conditional diffusion model as an inverse surrogate model. The framework employs an Uncertainty-aware Exploration (UaE) acquisition function to propose scores in the objective space for conditional sampling. The approach is shown theoretically to achieve a near-optimal solution and empirically outperforms existing online BBO baselines across 6 scientific discovery tasks.
A research paper by Dr. Reda Alami of the AI and Digital Science Research Center (AIDRC) at TII has been accepted for publication at the 14th Asian Conference on Machine Learning (ACML 2022). The paper addresses sequential decision-making under uncertainty in non-stationary environments, proposing a Bayesian Change-Point Detection with Thompson Sampling (Bayesian-CPD-TS) algorithm. The algorithm combines decision-making under uncertainty and sequential detection of abrupt changes. Why it matters: This recognition highlights the growing AI research capabilities within the UAE and its contribution to the global machine learning community.
Vaneet Aggarwal from Purdue University presented new research on discrete and continuous submodular bandits with full bandit feedback. The research introduces a framework transforming discrete offline approximation algorithms into sublinear α-regret methods using bandit feedback. Additionally, it introduces a unified approach for maximizing continuous DR-submodular functions, accommodating various settings and oracle access types. Why it matters: This research provides new methods for optimization under uncertainty, which is crucial for real-world AI applications in the region, such as resource allocation and automated decision-making.
The article discusses the importance of sample correlations in computer graphics, vision, and machine learning, highlighting how tailored randomness can improve the efficiency of existing models. It covers various correlations studied in computer graphics and tools to characterize them, including the use of neural networks for developing different correlations. Gurprit Singh from the Max Planck Institute for Informatics will be presenting on the topic. Why it matters: Optimizing sampling techniques via understanding and applying correlations can lead to significant advancements and efficiency gains across multiple AI fields.
A presentation will demonstrate the construction of well-calibrated, distribution-free neural Temporal Point Process (TPP) models from multiple event sequences using conformal prediction. The method builds a distribution-free joint prediction region for event arrival time and type with a finite-sample coverage guarantee. The refined method is based on the highest density regions, derived from the joint predictive density of event arrival time and type to address the challenge of creating a joint prediction region for a bivariate response that includes both continuous and discrete data types. Why it matters: This research from a KAUST postdoc improves uncertainty quantification in neural TPPs, which are crucial for modeling continuous-time event sequences, with applications in various fields, by providing more reliable prediction regions.
This article discusses approximating a high-dimensional distribution using Gaussian variational inference by minimizing Kullback-Leibler divergence. It builds upon previous research and approximates the minimizer using a Gaussian distribution with specific mean and variance. The study details approximation accuracy and applicability using efficient dimension, relevant for analyzing sampling schemes in optimization. Why it matters: This theoretical research can inform the development of more efficient and accurate AI algorithms, particularly in areas dealing with high-dimensional data such as machine learning and data analysis.
The paper introduces a novel actor-critic framework called Distillation Policy Optimization that combines on-policy and off-policy data for reinforcement learning. It incorporates variance reduction mechanisms like a unified advantage estimator (UAE) and a residual baseline. The empirical results demonstrate improved sample efficiency for on-policy algorithms, bridging the gap with off-policy methods.