Yu Liu
10 papers · Latest:
FAVOR: Efficient Filter-Agnostic Vector ANNS Based on Selectivity-Aware Exclusion Distances
FAVOR is a new ANNS method that efficiently integrates complex attribute filtering, achieving stable high throughput across varying selectivity levels.
Walk With Me: Long-Horizon Social Navigation for Human-Centric Outdoor Assistance
Walk with Me is a map-free framework enabling robots to perform safe, long-horizon social navigation outdoors using high-level human instructions.
Mutual Forcing: Dual-Mode Self-Evolution for Fast Autoregressive Audio-Video Character Generation
Mutual Forcing enables fast, high-quality autoregressive audio-video character generation by integrating dual-mode self-evolution for improved efficiency.
Harmonizing Generative Retrieval and Ranking in Chain-of-Recommendation
RecoChain unifies generative retrieval and ranking in a single Transformer, bridging the gap between generation and ranking for improved Top-K recommendations.
ASAP: An Azimuth-Priority Strip-Based Search Approach to Planar Microphone Array DOA Estimation in 3D
ASAP is a novel azimuth-priority strip-based search approach for fast and accurate 3D DOA estimation using planar microphone arrays.
Imaging Exploration of Molecular Subtypes in Tongue Squamous Cell Carcinoma
This study shows radiomic features can non-invasively distinguish molecular subtypes in tongue squamous cell carcinoma (TSCC), offering a new diagnostic tool.
Black-Box Skill Stealing Attack from Proprietary LLM Agents: An Empirical Study
This paper empirically studies black-box skill stealing from proprietary LLM agents, demonstrating easy extraction and highlighting overlooked copyright risks.
Seasonal Variability of Pluto's Haze Formation Revealed by Laboratory Simulations
Lab simulations show how seasonal methane variations on Pluto impact haze formation pathways and composition, enhancing nitrogen incorporation into organic solids.
Trans-RAG: Query-Centric Vector Transformation for Secure Cross-Organizational Retrieval
Trans-RAG introduces query-centric vector transformation for secure, efficient cross-organizational RAG, eliminating decryption overhead.
Adaptor: Advancing Assistive Teleoperation with Few-Shot Learning and Cross-Operator Generalization
Adaptor is a few-shot learning framework that achieves robust cross-operator intent recognition in assistive teleoperation by bridging domain gaps.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.