Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
OctWorld: Long-Range World-Consistent Video Generation with Octree-Based 3D MappingMicrosoft · Lecture notes in computer scienceJan 1From Visual Cues to Spoken Narration: Rethinking Audio DescriptionAlphabet · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1Efficient Diffusion with Two-Stage Dynamics Compression for High-Quality Talking Head GenerationiFlytek · IEEE Transactions on Circuits and Systems for Video TechnologyJan 1Skim and Skip: Hierarchical Adaptive Inference for Efficient Multimodal RetrievalMicrosoft · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1GUI-CC: Benchmarking Contextual Consistency of GUI World Models as Agent EnvironmentsAlibaba · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)Jan 1PCoMoE: Shifting MoE Inference from Monolithic Expert Selection to Fine-Grained Path CompositionAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1On the Design Fundamentals of Pixel Text Representation LearningAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1Visual Attention Faithfulness in Vision-Language Models is HeterogeneousMicrosoft · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1HomoDiff: Noise-Robust Multi-modal Homography Estimation via Diffusion ModelAlibaba · IEEE Transactions on Pattern Analysis and Machine IntelligenceJan 1Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use AgentsAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1Answer Probing-Guided Search for Diverse Solution Exploration of LLMsAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1PAC: Progress-Augmented Advantage Curriculum for Multi-Task Reinforcement Learning of LLMsAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1You Know What I Mean: A Benchmark for Agentic Conversational Reference GroundingAllen Institute for AI · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1Think, Look, and Revise: Inconsistency-Aware Visual Self-Correction in MLLMsNVIDIA · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1SPHERE: Automatic Music Upmixing via Audio Language Model Post-Training with Spatial Heuristics RewardsMeta · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1DICS: Exploring Data Intrinsic Consistency for Visual Instruction SelectionAlibaba · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)Jan 1The Illusion of What If: Evaluating the Breakdown of Counterfactual Reasoning in LLMsAlibaba · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)Jan 1Fully Unleashing the Multimodal Attacker: Meta-Adaptive Jailbreaking of Vision-Language ModelsAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1CultureConverse: A Multilingual Multi-turn Simulation Harness for Culturally Grounded Assistance in East and Southeast AsiaMicrosoft, Upstage · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1EvoHarmBench: Breaking Content Moderation with Iterative Human-Like EvasionAlibaba · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)Jan 1STAR : Sentence Translation Alignment Rate for Document-to-Document Machine TranslationAlibaba · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)Jan 1RubricRM: Generative Reward Modeling via Dynamic Rubrics for Image Generation and EditingAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1SeVeR: Selective Visual Exposure and Retrieval for 3D Medical Image Question AnsweringAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1AgentSpec: Speculative Decoding for Batch Inference of LLM AgentsMicrosoft · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1Context-Aware Cluster Decoding: Semantic Anchor-Driven Coherence in dMLLMsAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)Jan 1