Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Formalizing Feature Flag Synchronization in Multi-Region ArchitecturesMicrosoft · International Journal of Computer ApplicationsApr 18Understanding dropout risks in online courses: SHAP-based analysis of lecture interval patterns in LMSLG AI Research · Education and Information TechnologiesApr 18A Generative Victim Model for SegmentationAlibaba · International Journal of Computer VisionApr 18Sample Complexity Bounds for Stochastic Shortest Path with a Generative ModelGoogle DeepMind, Meta · arXiv (Cornell University)7 citationsApr 17LLMs Corrupt Your Documents When You DelegateMicrosoft · arXiv (Cornell University)1 citationApr 17CoEvolve: Training LLM Agents via Agent-Data Mutual EvolutionAlibaba · arXiv (Cornell University)1 citationApr 17Enhancing Classroom Monitoring in Science Experiments Using Synthetic VideosNVIDIA · International Conference on Information and Education Technology (ICIET)Apr 17MemExplorer: Navigating the Heterogeneous Memory Design Space for Agentic Inference NPUsMicrosoft · arXiv (Cornell University)Apr 17Mind's Eye: A Benchmark of Visual Abstraction, Transformation and Composition for Multimodal LLMsMicrosoft · arXiv (Cornell University)Apr 17CPU Optimization of a Monocular 3D Biomechanics Pipeline for Low-Resource DeploymentAlphabet · arXiv (Cornell University)Apr 17Do Vision-Language Models Truly Perform Vision Reasoning? A Rigorous Study of the Modality GapAlibaba · arXiv (Cornell University)Apr 17Chain-of-Thought Degrades Visual Spatial Reasoning Capabilities of Multimodal LLMsMicrosoft · arXiv (Cornell University)Apr 17KnitLoRA: bridging low-rank adaptation as interwoven layers for deeper semantic reasoningMeta · Scientific Reports1 citationApr 16Scaling Test-Time Compute for Agentic CodingGoogle DeepMind · arXiv (Cornell University)Apr 16Modeling LLM Unlearning as an Asymmetric Two-Task Learning ProblemAlibaba · arXiv (Cornell University)Apr 16Generalization in LLM Problem Solving: The Case of the Shortest PathAlphabet · arXiv (Cornell University)Apr 16UniPASE: A Generative Model for Universal Speech Enhancement with High Fidelity and Low HallucinationsHorizon Robotics · arXiv (Cornell University)Apr 16OpenMobile: Building Open Mobile Agents with Task and Trajectory SynthesisSenseTime · arXiv (Cornell University)Apr 16Scalable Model-Based Clustering with Sequential Monte CarloMicrosoft · arXiv (Cornell University)Apr 16Text2Arch: A Dataset for Generating Scientific Architecture Diagrams from Natural Language DescriptionsAlphabet · arXiv (Cornell University)Apr 16WavAlign: Enhancing Intelligence and Expressiveness in Spoken Dialogue Models via Adaptive Hybrid Post-TrainingAlibaba · arXiv (Cornell University)Apr 16RAD-2: Scaling Reinforcement Learning in a Generator-Discriminator FrameworkHorizon Robotics · arXiv (Cornell University)Apr 16Dual-Axis Generative Reward Model Toward Semantic and Turn-taking Robustness in Interactive Spoken Dialogue ModelsAlibaba · arXiv (Cornell University)Apr 16MM-WebAgent: A Hierarchical Multimodal Web Agent for Webpage GenerationMicrosoft · arXiv (Cornell University)Apr 16User choice, hidden gems, and the quadratic horizonAlphabet · PNAS NexusApr 16