Haoyu Wang
7 papers · Latest:
Formalize, Don't Optimize: The Heuristic Trap in LLM-Generated Combinatorial Solvers
LLMs should formalize, not optimize, combinatorial solvers, as attempts at search optimization lead to a "heuristic trap" and reduced correctness.
Unsafe by Flow: Uncovering Bidirectional Data-Flow Risks in MCP Ecosystem
MCP-BiFlow is a static analysis framework that uncovers bidirectional data-flow risks in Model Context Protocol (MCP) ecosystems.
CommitSuite: A Comprehensive Benchmark for Commit Classification and Message Generation
CommitSuite is a new benchmark for commit classification and message generation, featuring 63k CCS-compliant commits and a reference-free evaluation framework.
Prop-Chromeleon: Adaptive Haptic Props in Mixed Reality through Generative Artificial Intelligence
Prop-Chromeleon uses generative AI to transform everyday objects into adaptive haptic props in Mixed Reality, enhancing realism and immersion.
Listen to the Voices of Everyday Users: Democratizing Privacy Ratings for Sensitive Data Access in Mobile Apps
This paper introduces DePRa, a system that democratizes mobile app privacy ratings by involving everyday users to assess sensitive data access.
Fisher Decorator: Refining Flow Policy via A Local Transport Map
Fisher Decorator refines flow policies in offline RL by using a local transport map and anisotropic optimization, outperforming prior isotropic methods.
MATRIX: Multi-Layer Code Watermarking via Dual-Channel Constrained Parity-Check Encoding
MATRIX is a novel multi-layer, dual-channel code watermarking framework that uses parity-check encoding for robust provenance tracking in LLM-generated code.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.