arXiv subjects

Yutong Ren

Publications and source records attributed to Yutong Ren.

3 recordsLinked to original sources

arXiv2026

PhiBE-Q-Learning: Bridging Off-Policy Reinforcement Learning and Continuous-Time Control

In this paper, we develop an off-policy method for continuous-time reinforcement learning (CTRL), where the system dynamics are governed by an unknown stochastic differential equation (SDE) and only discrete-time trajectory data are available. A central challenge is that the classical state-action value function $Q(s,a)$, which enables off-policy learning in discrete-time RL, does not exist in CTRL (Baird, 1994; Jia and Zhou, 2023; Tallec et al., 2019). On the other hand, continuous-time control provides local notions such as the instantaneous advantage function $q(s,a)$, but these typically rely on state value function $V(s)$. To address this, we introduce a new definition of the state-action value function in CTRL and derive its governing equation. Building on the PhiBE approximation (Zhu, 2024; Zhu et al., 2025), we propose iterative algorithms to approximate the optimal $Q$-function in both model-based and model-free settings using only discrete-time off-policy data. Under linear function approximation, we establish convergence guarantees and derive explicit convergence rates for the proposed method.

math.OC↗

arXiv2026

PeriphAR: Fast and Accurate Real-World Object Selection with Peripheral Augmented Reality Displays

Gaze-based selection in XR requires visual confirmation due to eye-tracking limitations and target ambiguity in 3D contexts. Current designs for wide-FOV displays use world-locked, central overlays, which are not conducive to always-on AR glasses. This paper introduces PeriphAR (per-ree-far), a visualization technique that leverages peripheral vision for feedback during gaze-based selection on a monocular AR display. In a first user study, we isolated text, color, and shape properties of target objects to compare peripheral selection cues. Peripheral vision was more sensitive to color than shape, but this sensitivity rapidly declined at lower contrast. To preserve preattentive processing of color, we developed two strategies to enhance color in users' peripheral vision. In a second user study, our strategy that maximized contrast of the target to the neighboring object with the most similar color was subjectively preferred. As proof of concept, we implemented PeriphAR in an end-to-end system to test performance with real-world object detection.

cs.HC↗

arXiv2025

PHYBench: Holistic Evaluation of Physical Perception and Reasoning in Large Language Models

Current benchmarks for evaluating the reasoning capabilities of Large Language Models (LLMs) face significant limitations: task oversimplification, data contamination, and flawed evaluation items. These deficiencies necessitate more rigorous assessment methods. To address these limitations, we introduce PHYBench, a benchmark of 500 original physics problems ranging from high school to Physics Olympiad difficulty. PHYBench addresses data contamination through original content and employs a systematic curation pipeline to eliminate flawed items. Evaluations show that PHYBench activates more tokens and provides stronger differentiation between reasoning models compared to other baselines like AIME 2024, OlympiadBench and GPQA. Even the best-performing model, Gemini 2.5 Pro, achieves only 36.9% accuracy compared to human experts' 61.9%. To further enhance evaluation precision, we introduce the Expression Edit Distance (EED) Score for mathematical expression assessment, which improves sample efficiency by 204% over binary scoring. Moreover, PHYBench effectively elicits multi-step and multi-condition reasoning, providing a platform for examining models' reasoning robustness, preferences, and deficiencies. The benchmark results and dataset are publicly available at https://www.phybench.cn/.

cs.CL↗