arXiv subjects

Mumuksh Tayal

Publications and source records attributed to Mumuksh Tayal.

6 recordsLinked to original sources

arXiv2026

Safe Flow Q-Learning: Offline Safe Reinforcement Learning with Reachability-Based Flow Policies

Offline safe reinforcement learning (RL) seeks reward-maximizing policies from static datasets under strict safety constraints. Existing methods often rely on soft expected-cost objectives or iterative generative inference, which can be insufficient for safety-critical real-time control. We propose Safe Flow Q-Learning (SafeFQL), which extends FQL to safe offline RL by combining a Hamilton--Jacobi reachability-inspired safety value function with an efficient one-step flow policy. SafeFQL learns the safety value via a self-consistency Bellman recursion, trains a flow policy by behavioral cloning, and distills it into a one-step actor for reward-maximizing safe action selection without rejection sampling at deployment. Empirically, SafeFQL trades modestly higher offline training cost for substantially lower inference latency than diffusion-style safe generative baselines, which is advantageous for real-time safety-critical deployment. Across boat navigation, and Safety Gymnasium MuJoCo tasks, SafeFQL matches or exceeds prior offline safe RL performance while substantially reducing constraint violations.

cs.LG↗

arXiv2026

Epigraph-Guided Flow Matching for Safe and Performant Offline Reinforcement Learning

Offline reinforcement learning (RL) provides a compelling paradigm for training autonomous systems without the risks of online exploration, particularly in safety-critical domains. However, jointly achieving strong safety and performance from fixed datasets remains challenging. Existing safe offline RL methods often rely on soft constraints that allow violations, introduce excessive conservatism, or struggle to balance safety, reward optimization, and adherence to the data distribution. To address this, we propose Epigraph-Guided Flow Matching (EpiFlow), a framework that formulates safe offline RL as a state-constrained optimal control problem to co-optimize safety and performance. We learn a feasibility value function derived from an epigraph reformulation of the optimal control problem, thereby avoiding the decoupled objectives or post-hoc filtering common in prior work. Policies are synthesized by reweighting the behavior distribution based on this epigraph value function and fitting a generative policy via flow matching, enabling efficient, distribution-consistent sampling. Across various safety-critical tasks, including Safety-Gymnasium benchmarks, EpiFlow achieves competitive returns with near-zero empirical safety violations, demonstrating the effectiveness of epigraph-guided policy synthesis.

cs.LG↗

arXiv2025

V-OCBF: Learning Safety Filters from Offline Data via Value-Guided Offline Control Barrier Functions

Ensuring safety in autonomous systems requires controllers that aim to satisfy state-wise constraints without relying on online interaction.While existing Safe Offline RL methods typically enforce soft expected-cost constraints, they struggle to ensure strict state-wise safety. Conversely, Control Barrier Functions (CBFs) offer a principled mechanism to enforce forward invariance, but often rely on expert-designed barrier functions or knowledge of the system dynamics. We introduce Value-Guided Offline Control Barrier Functions (V-OCBF), a framework that learns a neural CBF entirely from offline demonstrations. Unlike prior approaches, V-OCBF does not assume access to the dynamics model; instead, it derives a recursive finite-difference barrier update, enabling model-free learning of a barrier that propagates safety information over time. Moreover, V-OCBF incorporates an expectile-based objective that avoids querying the barrier on out-of-distribution actions and restricts updates to the dataset-supported action set. The learned barrier is then used with a Quadratic Program (QP) formulation to synthesize real-time safe control. Across multiple case studies, V-OCBF yields substantially fewer safety violations than baseline methods while maintaining strong task performance, highlighting its scalability for offline synthesis of safety-critical controllers without online interaction or hand-engineered barriers.

cs.AI↗

arXiv2025

RISE: Robust Imitation through Stochastic Encoding

Ensuring safety in robotic systems remains a fundamental challenge, especially when deploying offline policy-learning methods such as imitation learning in dynamic environments. Traditional behavior cloning (BC) often fails to generalize when deployed without fine-tuning as it does not account for disturbances in observations that arises in real-world, changing environments. To address this limitation, we propose RISE (Robust Imitation through Stochastic Encodings), a novel imitation-learning framework that explicitly addresses erroneous measurements of environment parameters into policy learning via a variational latent representation. Our framework encodes parameters such as obstacle state, orientation, and velocity into a smooth variational latent space to improve test time generalization. This enables an offline-trained policy to produce actions that are more robust to perceptual noise and environment uncertainty. We validate our approach on two robotic platforms, an autonomous ground vehicle and a Franka Emika Panda manipulator and demonstrate improved safety robustness while maintaining goal-reaching performance compared to baseline methods.

cs.RO↗

arXiv2025

A Fixed Parameter Tractable Approach for Solving the Vertex Cover Problem in Polynomial Time Complexity

The Minimum Vertex Cover problem, a classical NP-complete problem, presents significant challenges for exact solution on large graphs. Fixed-Parameter Tractability (FPT) offers a powerful paradigm to address such problems by exploiting a parameter of the input, typically related to the size of the desired solution. This paper presents an implementation and empirical evaluation of an FPT algorithm for the Minimum Vertex Cover problem parameterized by the size of the vertex cover, $k$. The algorithm utilizes a branching strategy based on selecting adjacent vertices and recursively solving subproblems on a reduced graph. We describe the algorithmic approach, implementation details in Python, and present experimental results comparing its performance against the SageMath computational system. The results demonstrate that the FPT implementation achieves significant performance improvements for instances with large numbers of vertices ($n$) but relatively small values of the parameter ($k$), aligning with theoretical FPT complexity guarantees. We also discuss potential optimizations that could further improve the algorithm's performance, particularly concerning the branching factor.

cs.DS↗

arXiv2025

Evaluating Multi-Instance DNN Inferencing on Multiple Accelerators of an Edge Device

Edge devices like Nvidia Jetson platforms now offer several on-board accelerators -- including GPU CUDA cores, Tensor Cores, and Deep Learning Accelerators (DLA) -- which can be concurrently exploited to boost deep neural network (DNN) inferencing. In this paper, we extend previous work by evaluating the performance impacts of running multiple instances of the ResNet50 model concurrently across these heterogeneous components. We detail the effects of varying batch sizes and hardware combinations on throughput and latency. Our expanded analysis highlights not only the benefits of combining CUDA and Tensor Cores, but also the performance degradation from resource contention when integrating DLAs. These findings, together with insights on precision constraints and workload allocation challenges, motivate further exploration of intelligent scheduling mechanisms to optimize resource utilization on edge platforms.

cs.DC↗