Chen Wang

The paper introduces PACT, a provenance-aware runtime monitor that enhances agent security by tracking the origin and trust of individual tool arguments, solving the granularity mismatch in LLM agent defenses.

The paper introduces AsyncTool, a new benchmark designed to evaluate LLM agents' ability to handle multiple, concurrent tasks with delayed tool feedback, demonstrating that asynchronous coordination is a significant challenge for current models.

The paper proposes GASP, a framework that injects fundamental geometric priors directly into Vision-Language Models (VLMs) using ground-truth video geometry, significantly enhancing 3D spatial reasoning without requiring 3D VQA data.

The paper introduces EASE, a method that enhances multimodal Reinforcement Learning with Verifiable Rewards (RLVR) by providing spatial attention supervision anchored to visual evidence, significantly improving visual grounding and reasoning capabilities in VLMs.

This paper systematically evaluates the consistency of popular causal discovery benchmarks against real-world scientific literature, revealing significant variability in their accuracy.

This paper proposes a method to improve few-step generation in diffusion models by introducing a marginal-alignment regularizer.

The paper introduces TradeLens, a toolkit for evaluating the agentic viability of large language model agents in trading systems by reconstructing trading trajectories and diagnosing intelligence-to-profit conversion.

This paper proposes a neural audio watermarking method that embeds a message into the continuous latent representation of a codec-like speech autoencoder for improved codec robustness.