17 accepted papers
Reinforcement learning (RL) faces significant challenges in real-world deployments due to the sim-to-real gap, where policies trained in simulators often underperform in practice due to mismatches between training and deployment conditions. Distributionally robust RL addresses this issue by optimizi
Well-trained multi-agent systems can fail when deployed in real-world environments due to model mismatches between the training and deployment environments, caused by environment uncertainties including noise or adversarial attacks. Distributionally Robust Markov Games (DRMGs) enhance system resilie…
Robust reinforcement learning (RL) under the average reward criterion, which seeks to optimize long-term system performance in uncertain environments, remains a largely unexplored area. To address this challenge, we propose a reduction-based framework that transforms robust average reward optimizati…
The alignment of large language models (LLMs) with human values is critical as these models become increasingly integrated into various societal and decision-making processes. Traditional methods, such as reinforcement learning from human feedback (RLHF), achieve alignment by fine-tuning model param…
Traditional adversarial attacks typically aim to alter the predicted labels of input images by generating perturbations that are imperceptible to the human eye. However, these approaches often lack explainability. Moreover, most existing work on adversarial attacks focuses on single-stage classifier…
Planning in non-Markovian environments often requires inferring task structures, such as reward machines, through interactions with the environment. Traditional active grammatical inference methods, like Angluin’s L<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/19…
Offline reinforcement learning (RL) has gained considerable attention for its ability to learn policies from pre-collected data without real-time interaction, which makes it particularly useful for high-risk applications. However, due to its reliance on offline datasets, existing works inevitably in…
Transfer reinforcement learning aims to derive a near-optimal policy for a target environment with limited data by leveraging abundant data from related source domains. However, it faces two key challenges: the lack of performance guarantees for the transferred policy, which can lead to undesired ac…
Domain shift occurs when the data distribution used for testing diverges from the one employed during training, resulting in reduced performance of machine learning models. Domain adaptation (DA) offers techniques to mitigate disparities between two domains and enhance a model’s effectiveness in the…
Robust Markov decision processes (MDPs) address the challenge of model uncertainty by optimizing the worst-case performance over an uncertainty set of MDPs. In this paper, we focus on the robust average-reward MDPs under the model-free setting. We first theoretically characterize the structure of so…
Robust tensor completion aims to recover a tensor from partially observed noisy entries that may be contaminated with large outliers by exploiting its low-rank property. While there exist several robust tensor completion algorithms, their reliance on singular value decomposition (SVD) limits their s…
Tensor ring (TR) decomposition has recently received increased attention due to its superior expressive performance for high-order tensors. However, the applicability of traditional TR decomposition algorithms to real-world applications is hindered by prevalent large data sizes, missing entries, and…
Traditional subspace clustering aims to cluster data lying in a union of linear subspaces. The vectorization of high-dimensional data to 1-D vectors to perform clustering ignores much of the structure intrinsic to such data. To preserve said structure, in this work we exploit clustering in a high-or…
Adversarial attacks can drastically reduce the accuracy and confidence level of classifiers while being imperceptible. Existing studies on the topic have largely focused on one-stage classifiers. In this paper, we study the robustness of two Two-Stage Hierarchical Classifier models, the flat and top…
Adversarial attacks have exposed the vulnerability of one-stage classifiers to carefully crafted perturbations which were shown to drastically alter their predictions while remaining imperceptible. In this paper, we examine the susceptibility of coarse-to-fine hierarchical classifiers to such types…
Repetitive visual stimuli induce periodic Visual Evoked Potentials (VEPs) in the brain that can be potentially identified in an EEG trace. The ability to distinguish frequencies and patterns due to different stimuli is the basis for brain computer interfaces (BCIs) used for communication and control…
This paper focuses on the low rank plus sparse matrix decomposition problem in big data settings. Conventional algorithms solve high-dimensional optimization problems that scale with the data dimension, which limits their scalability. In addition, existing randomized approaches mostly rely on blind…