Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
StreamErase: Towards Ultra-Fast Streaming Video ErasingAlibaba · arXiv (Cornell University)Oct 14Lluna: Collaborative Unimodal Models for Missing-Modality InferenceMicrosoft · arXiv (Cornell University)Oct 14What Makes a Social Agent's Interactions Effective? Learning from Verifiable Outcomes in Paired WorldsAlibaba · arXiv (Cornell University)Oct 14Learning Visual Speech Representations via Cross-Modal Distillation and Joint Face-Lip ModelingiFlytek · arXiv (Cornell University)Oct 14Progress-Aware Distillation for Mitigating Stagnation in Small Language Model AgentsMicrosoft · arXiv (Cornell University)Oct 14Beyond Expert Rankings: Contrast Calibration for Mixed-Precision MoE QuantizationMicrosoft · arXiv (Cornell University)Oct 14What Matters for World Modeling in Vision-Language-Action Models?Alibaba · arXiv (Cornell University)Oct 14The COUNT: Counting N-grams Unlocks Scaling on Host Memory and Training-Free AdaptationAllen Institute for AI · arXiv (Cornell University)Oct 14Memory-Efficient Subspace Optimization with Residual Correction: From Momentum SGD to MuonAlibaba · arXiv (Cornell University)Oct 14Asymptotic Behavior of TransformersAlphabet · arXiv (Cornell University)Oct 14VASPO: Self-Distillation with Privileged Value EstimationiFlytek · arXiv (Cornell University)Oct 14Nonlinear fixed-distance codes via real lifts of moment configurationsMicrosoft · arXiv (Cornell University)Oct 14One Block Ahead: Promoting Sampler State in Diffusion Language ModelsNVIDIA · arXiv (Cornell University)Oct 14Not All Criteria Matter: Importance-Aware Criteria Weighting for Rubric-Based LLM EvaluationAlibaba · arXiv (Cornell University)Oct 14OS-Bench: Evaluating SWE Agents on Cross-OS Code MigrationMicrosoft · arXiv (Cornell University)Oct 14DeLoop: Repeated Actions as Self-Supervision for Long-Horizon AgentsGoogle DeepMind · arXiv (Cornell University)Oct 14EvalBuilder: Can LLMs Build Efficient Benchmarks?Allen Institute for AI · arXiv (Cornell University)Oct 14Co-Evolving Acting and Forecasting for Language AgentsMeta · arXiv (Cornell University)Oct 14Component-Level Inverse Design of Transmon Qubits Using Neural NetworksNVIDIA · Research SquareOct 14AgentCIScope: Grounded Contextual Integrity Evaluation for LLM AgentsMicrosoft · arXiv (Cornell University)Oct 13Logical gate set tomographyNVIDIA · arXiv (Cornell University)Oct 13HERO: Harnessing Evidence for Test-time Refinement and Prompt Optimization in Program GenerationMicrosoft · arXiv (Cornell University)Oct 13CacheSmith: Accurate RAG Serving under Latency SLOs with Adaptive KV ReuseMicrosoft · arXiv (Cornell University)Oct 13Imagining Goals with Geometric Horizon Models for Policy GeneralizationGoogle DeepMind · arXiv (Cornell University)Oct 13Contextual Credit Policy OptimizationMicrosoft · arXiv (Cornell University)Oct 13