Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Socratic-Geo: Synthetic Data Generation and Geometric Reasoning via Multi-Agent InteractionAlibaba · arXiv (Cornell University)Feb 3The Trigger in the Haystack: Extracting and Reconstructing LLM Backdoor TriggersMicrosoft · arXiv (Cornell University)Feb 3Beyond Quantity: Trajectory Diversity Scaling for Code AgentsAlibaba · arXiv (Cornell University)Feb 3A dimensional structure based knowledge distillation method for cross-modal learningiFlytek · Neural NetworksFeb 3Taming Multidimensional Analytics Costs: An Empirical Study of HLL and KLL Sketches in Cloud Data WarehousesAnthropic · IEEE SoftwareFeb 3Prompt grouping for rehearsal-free continual learningiFlytek · Multimedia SystemsFeb 3DiffusionLight-Turbo: Accelerated Light Probes for Free via Single-Pass Chrome Ball InpaintingGoogle DeepMind · IEEE Transactions on Pattern Analysis and Machine Intelligence3 citationsFeb 2Learned Image and Video Compression: Foundations, Algorithms, and Open ChallengesNVIDIA · Mile-High Video Conference (MHV)1 citationFeb 2Bespoke Ad Targeting via LLM-Generated User Profiles and RAG-Based RecommendationsMicrosoft · International Conference on AI x Data and Knowledge Engineering (AIxDKE)Feb 2Focused Frames: Enhancing Video Quality Assessment Through Perceptual Pre-CroppingAlphabet · Mile-High Video Conference (MHV)Feb 2MARS: Modular Agent with Reflective Search for Automated AI ResearchAlphabet · arXiv (Cornell University)Feb 2WAXAL: A Large-Scale Multilingual African Language Speech CorpusAlphabet · arXiv (Cornell University)Feb 2Outrunning LLM Cutoffs: A Live Kernel Crash Resolution Benchmark for AllAlphabet · arXiv (Cornell University)Feb 2Why Steering Works: Toward a Unified View of Language Model Parameter DynamicsAlibaba · arXiv (Cornell University)Feb 2Contribution-aware Token Compression for Efficient Video Understanding via Reinforcement LearningAlibaba · arXiv (Cornell University)Feb 2Read As Human: Compressing Context via Parallelizable Close Reading and SkimmingAlibaba · arXiv (Cornell University)Feb 2CoMeT: Collaborative Memory Transformer for Efficient Long Context ModelingAlibaba · arXiv (Cornell University)Feb 2D-CORE: Incentivizing Task Decomposition in Large Reasoning Models for Complex Tool UseAlibaba · arXiv (Cornell University)Feb 2Co-RedTeam: Orchestrated Security Discovery and Exploitation with LLM AgentsAlphabet · arXiv (Cornell University)Feb 2RE-TRAC: REcursive TRAjectory Compression for Deep Search AgentsMicrosoft · arXiv (Cornell University)Feb 2Think Dense, Not Long: Dynamic Decoupled Conditional Advantage for Efficient ReasoningAlibaba · arXiv (Cornell University)Feb 2Focus-dLLM: Accelerating Long-Context Diffusion LLM Inference via Confidence-Guided Context FocusingSenseTime · arXiv (Cornell University)Feb 2Entropy-Guided Data-Efficient Training for Multimodal Reasoning Reward ModelsAlibaba · arXiv (Cornell University)Feb 2COMI: Coarse-to-fine Context Compression via Marginal Information GainAlibaba · arXiv (Cornell University)Feb 2SWE-Universe: Scale Real-World Verifiable Environments to MillionsAlibaba · arXiv (Cornell University)Feb 2