Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
ReToken: Improving Long-Context VLMs with Visual Retrieval TokenGoogle DeepMind, Microsoft · arXiv (Cornell University)Jul 30Beyond Frame Selection: Generative Latent Evidence Aggregation for Long-Video UnderstandingAlibaba · arXiv (Cornell University)Jul 30Can LVLMs Uncover the Truth Behind Visual Illusions? An Analysis of Perceptual and Reasoning CapabilitiesAlibaba · arXiv (Cornell University)Jul 30Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at ScaleMicrosoft · arXiv (Cornell University)Jul 30ReDiPPO: Reference-Guided Value Calibration and Discrepancy-Aware Token Reweighting for Mathematical ReasoningiFlytek · arXiv (Cornell University)Jul 30SKILL-KD: Contrastive Skill Distillation for LLM AgentsAlibaba · arXiv (Cornell University)Jul 30LoopMemGR: From Behavior Logs to Evolving Memory for Generative RecommendationAlibaba · arXiv (Cornell University)Jul 30Robotic ultrasound scanning platform with autonomous control, multimodal human–machine interface and real-time image analysisNVIDIA · International Journal of Computer Assisted Radiology and SurgeryJul 30BitNet v2: Low-Bit and Sparsified Activations for 1-Bit Large Language ModelsMicrosoft · IEEE Transactions on Pattern Analysis and Machine Intelligence1 citationJul 29MCP vs CLI for AI Agents: A Comparative Analysis of Tool Integration Strategies for Autonomous SystemsMicrosoft · arXiv (Cornell University)Jul 29BridgeAlign: Bridging Preference Alignment for Humanities and Social SciencesAlibaba · arXiv (Cornell University)Jul 29Beyond the Bidirectional Promise: Re-evaluating the Robustness of Diffusion Language ModelsMicrosoft · arXiv (Cornell University)Jul 29Rethinking Self-Evolution: A Constrained Exploration-Exploitation Process for Mitigating Skill OverfittingAlibaba · arXiv (Cornell University)Jul 29SecRespond: Benchmarking AI Agents for Real-World Post-Compromise Incident ResponseAlibaba · arXiv (Cornell University)Jul 29BayesAME: Bayesian Active Model EvaluationGoogle DeepMind · arXiv (Cornell University)Jul 29Learning from the Future: Privileged Self-Distillation for Sequential RecommendationAlibaba · arXiv (Cornell University)Jul 29SERPO: Self-Evolving Rubric Policy Optimization for Open-Ended Test-Time Reinforcement LearningAlibaba · arXiv (Cornell University)Jul 29Living-Harness Is an Interactive-Agent EvolverAlibaba · arXiv (Cornell University)Jul 29Contrastive ESA: Human Evaluation of Multiple Translations at OnceAlphabet · arXiv (Cornell University)Jul 29TinyMem: Condensing Multimodal Memory for Long-Form Video Action DetectionMicrosoft · IEEE Transactions on Pattern Analysis and Machine IntelligenceJul 29Quantum meets capitalism: how to pair long-term bets on technology with commercial speedMicrosoft · NatureJul 29Weak-to-Strong On-Policy DistillationMicrosoft · arXiv (Cornell University)Jul 28DVPSFormer: Efficient Online Depth-aware Video Panoptic Segmentation for Autonomous DrivingMicrosoft · arXiv (Cornell University)Jul 28Specula: Scaling formal specifications for autonomous model checking of system codeMicrosoft · arXiv (Cornell University)Jul 28TopoGR: Revealing and Preserving Latent Structure of Semantic ID in Generative RecommendationAlibaba · arXiv (Cornell University)Jul 28