Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
OGLS-SD: On-Policy Self-Distillation with Outcome-Guided Logit Steering for LLM ReasoningNVIDIA · arXiv (Cornell University)May 12Sobolev Regularized MMD Gradient FlowAlphabet, Google DeepMind · arXiv (Cornell University)May 12Controllable User SimulationAlphabet · arXiv (Cornell University)May 12SkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill GraphsAlibaba · arXiv (Cornell University)May 12Formalize, Don't Optimize: The Heuristic Trap in LLM-Generated Combinatorial SolversGoogle DeepMind · arXiv (Cornell University)May 12GEAR: Granularity-Adaptive Advantage Reweighting for LLM Agents via Self-DistillationMicrosoft · arXiv (Cornell University)May 12Robust LLM Unlearning Against Relearning Attacks: The Minor Components in Representations MatterAlibaba · arXiv (Cornell University)May 12Robust Promptable Video Object SegmentationAlphabet · arXiv (Cornell University)May 12From Noise to Diversity: Random Embedding Injection in LLM ReasoningMicrosoft · arXiv (Cornell University)May 12Reinforcing VLAs in Task-Agnostic World ModelsMicrosoft · arXiv (Cornell University)May 12Avoiding Cross-Datacenter Collective Congestion via Disaggregated BufferingNVIDIA · arXiv (Cornell University)May 12Learning Agentic Policy from Action GuidanceAlibaba · arXiv (Cornell University)May 12Agent-BRACE: Decoupling Beliefs from Actions in Long-Horizon Tasks via Verbalized State UncertaintyMicrosoft · arXiv (Cornell University)May 12Large Language Models (LLMs) for fisheries management research: understanding potential and navigating risksAlphabet · Reviews in Fish Biology and FisheriesMay 12A highly sensitive mixed-signal auto-range adaptable interface for capacitive sensorsAlphabet · MeasurementMay 12Analytical Modeling of Set-Associative Caches for Optimizing Tensor OperationsAlphabet · ACM Transactions on Architecture and Code OptimizationMay 12ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?Alphabet, Anthropic, OpenAI · arXiv (Cornell University)3 citationsMay 11AI in the Enterprise: How People Use M365 Copilot ChatMicrosoft · arXiv (Cornell University)May 11LoKA: Low-precision Kernel Applications for Recommendation Models At ScaleMeta · arXiv (Cornell University)May 11ShardTensor: Domain Parallelism for Scientific Machine LearningNVIDIA · arXiv (Cornell University)May 11Language Modeling with Hyperspherical FlowsMicrosoft · arXiv (Cornell University)May 11Sieve: Dynamic Expert-Aware PIM Acceleration for Evolving Mixture-of-Experts ModelsNVIDIA · arXiv (Cornell University)May 11Curriculum Learning-Guided Progressive Distillation in Large Language ModelsAlphabet · arXiv (Cornell University)May 11Interpretability Can Be ActionableGoogle DeepMind · arXiv (Cornell University)May 11Multi-domain Multi-modal Document Classification Benchmark with a Multi-level TaxonomyAlibaba · arXiv (Cornell University)May 11