Michael K. Ng
Publications
RewardFlow: Topology-Aware Reward Propagation on State Graphs for Agentic RL with Large Language Models
Reinforcement learning (RL) holds significant promise for enhancing the agentic reasoning capabilities of large language models (LLMs) with external environments. However, the inherent sparsity of terminal rewards hinders fine-grained, state-level optimization. Although process reward modeling offers a promising alternative, training dedicated reward models often entails substantial computational costs and scaling difficulties. To address these challenges, we introduce RewardFlow, a lightweight method for estimating state-level rewards tailored to agentic reasoning tasks. RewardFlow leverages the intrinsic topological structure of states within reasoning trajectories by constructing state graphs. This enables an analysis of state-wise contributions to success, followed by topology-aware graph propagation to quantify contributions and yield objective, state-level rewards. When integrated as dense rewards for RL optimization, RewardFlow substantially outperforms prior RL baselines across four agentic reasoning benchmarks, demonstrating superior performance, robustness, and training efficiency. The implementation of RewardFlow is publicly available at https://github.com/tmlr-group/RewardFlow.
TVWorld: Foundations for Remote-Control TV Agents
Recent large vision-language models (LVLMs) have demonstrated strong potential for device control. However, existing research has primarily focused on point-and-click (PnC) interaction, while remote-control (RC) interaction commonly encountered in everyday TV usage remains largely underexplored. To fill this gap, we introduce \textbf{TVWorld}, an offline graph-based abstraction of real-world TV navigation that enables reproducible and deployment-free evaluation. On this basis, we derive two complementary benchmarks that comprehensively assess TV-use capabilities: \textbf{TVWorld-N} for topology-aware navigation and \textbf{TVWorld-G} for focus-aware grounding. These benchmarks expose a key limitation of existing agents: insufficient topology awareness for focus-based, long-horizon TV navigation. Motivated by this finding, we propose a \emph{Topology-Aware Training} framework that injects topology awareness into LVLMs. Using this framework, we develop \textbf{TVTheseus}, a foundation model specialized for TV navigation. TVTheseus achieves a success rate of $68.3\%$ on TVWorld-N, surpassing strong closed-source baselines such as Gemini 3 Flash and establishing state-of-the-art (SOTA) performance. Additional analyses further provide valuable insights into the development of effective TV-use agents.