Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Visual Late Chunking: An Empirical Study of Contextual Chunking for Efficient Visual Document RetrievalAlibaba · arXiv (Cornell University)Apr 11WaveTune: Wave-aware Bilinear Modeling for Efficient GPU Kernel Auto-tuningAlibaba · arXiv (Cornell University)Apr 11Learning from Emptiness: De-biasing Listwise Rerankers with Content-Agnostic Probability CalibrationAlibaba · arXiv (Cornell University)Apr 11Scylla: Translating an Applicative Subset of C to Safe RustAlphabet · Proceedings of the ACM on Programming Languages7 citationsApr 10InspectCoder: Dynamic Analysis-Driven Self Repair through Interactive LLM-Debugger CollaborationAlibaba · Proceedings of the ACM on Programming Languages1 citationApr 10Hunting CUDA Bugs at Scale with cuFuzzNVIDIA · Proceedings of the ACM on Programming Languages1 citationApr 10GenesisFunc: Multi-Agent Data Generation for Accurate and Generalizable Function-CallingAlibaba · arXiv (Cornell University)Apr 10EdgeRazor: A Lightweight Framework for Large Language Models via Mixed-Precision Quantization-Aware DistillationMicrosoft · arXiv (Cornell University)Apr 10Generative Circuit Design for Quantum-Selected Configuration InteractionNVIDIA · arXiv (Cornell University)Apr 10MEMENTO: Teaching LLMs to Manage Their Own ContextMicrosoft · arXiv (Cornell University)Apr 10SLM Finetuning for Natural Language to Domain Specific Code Generation in ProductionMicrosoft · arXiv (Cornell University)Apr 10Discrete Token Modeling for Multi-Stem Music Source Separation with Language ModelsAlibaba · arXiv (Cornell University)Apr 10Tantalum-Encapsulated Niobium Superconducting Resonators: High Internal Quality Factor and Improved Temporal Stability via Surface PassivationTechnology Innovation Institute · arXiv (Cornell University)Apr 10Breaking Block Boundaries: Anchor-based History-stable Decoding for Diffusion Large Language ModelsAlibaba · arXiv (Cornell University)Apr 10Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm TypesCohere · arXiv (Cornell University)Apr 10SPPO: Sequence-Level PPO for Long-Horizon Reasoning TasksMicrosoft · arXiv (Cornell University)Apr 10Online3R: Online Learning for Consistent Sequential Reconstruction Based on Geometry Foundation ModelNVIDIA · arXiv (Cornell University)Apr 10ScheMatiQ: From Research Question to Structured Data through Interactive Schema DiscoveryAllen Institute for AI · arXiv (Cornell University)Apr 10Tora3: Trajectory-Guided Audio-Video Generation with Physical CoherenceAlibaba · arXiv (Cornell University)Apr 10Interactive ASR: Towards Human-Like Interaction and Semantic Coherence Evaluation for Agentic Speech RecognitionAlibaba · arXiv (Cornell University)Apr 10BERT-as-a-Judge: A Robust Alternative to Lexical Methods for Efficient Reference-Based LLM EvaluationCohere · arXiv (Cornell University)Apr 10Automatic Propagation of Profile Information through the Optimization PipelineMicrosoft · Proceedings of the ACM on Programming LanguagesApr 10Gradient‐Domain ReSTIR Path TracingNVIDIA · Computer Graphics ForumApr 10Project Everest: Perspectives from Developing Industrial-Grade High-Assurance SoftwareMicrosoft · ACM Transactions on Programming Languages and SystemsApr 10PLUM: Adapting Pre-trained Language Models for Industrial-scale Generative RecommendationsAlphabet · International World Wide Web Conference (WWW)5 citationsApr 9