Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Ruby: Unmasking Unsafe Rust in Stripped Binaries via Machine LearningMicrosoft · International Conference on Dependable Systems & Networks/ProceedingsJun 22You Don't Need to Run Every EvalMicrosoft · arXiv (Cornell University)Jun 22Polynomial-time exact diagonalization via sparse guided eigenwalksNVIDIA · arXiv (Cornell University)Jun 22A System-Level Framework for Evaluating Privacy in Hybrid LLM DeploymentsMicrosoft · AlgorithmsJun 22FlowTTS-GRPO: Online Reinforcement Learning with Multi-Objective Reward Optimization for Flow-Matching Based Text-to-SpeechAlibaba · arXiv (Cornell University)Jun 22AudioCALM: Continuous Autoregressive Language Modeling for Universal Audio GenerationAlibaba · arXiv (Cornell University)Jun 22Multi-Vector Embeddings are Provably More Expressive than Single Vector EmbeddingsAlphabet · arXiv (Cornell University)Jun 22STAR-VAE: Structured Topology-Aware Regularization for Audio Reconstruction and GenerationAlibaba · arXiv (Cornell University)Jun 22StatABench: Dataset and Framework for Evaluating Statistical Analysis Capabilities of LLMsAlibaba · arXiv (Cornell University)Jun 22Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement LearningMicrosoft · arXiv (Cornell University)Jun 22ScalingAttention: Discovering Intrinsic Sparse Attention Topology for Video Diffusion TransformersNVIDIA · arXiv (Cornell University)Jun 22Evaluation Awareness Is Not One Capability: Evidence from Open Language ModelsMicrosoft · arXiv (Cornell University)Jun 22Bagpiper-TTS: Natural Language Guided Universal Speech SynthesisNVIDIA · arXiv (Cornell University)Jun 22Agent-as-a-Router: Agentic Model Routing for Coding TasksAlibaba · arXiv (Cornell University)Jun 22Rising From the Ashes: How Agentic AI is Unblocking Challenges in CybersecurityMicrosoft · arXiv (Cornell University)Jun 22Tmax: A simple recipe for terminal agentsAllen Institute for AI · arXiv (Cornell University)Jun 22FORGE: Fused On-Register Gradient Elimination for Memory-Efficient LLM TrainingNVIDIA · arXiv (Cornell University)Jun 22Modeling polarization in public opinion through LLM-synthesized arguments and stance treesGoogle DeepMind · PeerJ Computer ScienceJun 22Adaptive Recurrent Message Passing for Test Time Computing on GraphsAlibaba · arXiv (Cornell University)Jun 21OmniSpace: Efficient Geometry Awareness for Autonomous Vehicles MLLMsAlphabet · arXiv (Cornell University)Jun 21What are Key Factors for Updates in RL for LLM Reasoning?Microsoft · arXiv (Cornell University)Jun 21PRESTO: Prefix-Aligned Tree Drafting for Diffusion Speculative DecodingNVIDIA · arXiv (Cornell University)Jun 20Streaming T5-based Text-to-Speech Synthesis with Limited LookaheadNVIDIA · arXiv (Cornell University)Jun 20Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer DecodingAlibaba · arXiv (Cornell University)Jun 20OpenBioRQ: Unsolved Biomedical Research Questions for AgentsUpstage · arXiv (Cornell University)Jun 20