NVIDIA · 102 records

dateidtitleauthorsvenue↑
2026-11-202611.26629Posterior Matching DistillationHidir Yesiltepe et al.arXiv31
2026-11-202611.23835NarrowSpec: Draft Hidden Width Need Not Match the Target in Speculative DecodingAlagappan Valliappan et al.arXiv3
2026-11-132611.17331TTB: Test-time Baking for Efficient Rendering of Decoder-only View Synthesis ModelsChin-Yang Lin et al.arXiv22
2026-11-102611.14815BirdDreams: Joint Perspective and BEV Video Generation for a Plausible, Reactive Driving World ModelJunyoung Seo et al.arXiv50
2026-11-092610.35959AgentBaton: Efficient Model Collaboration in Agentic TasksJunlong Li et al.arXiv19
2026-11-092611.08290AutoCompress: Automated Hardware-Guided Model CompressionSheldon Liang et al.arXiv11
2026-11-092611.09891CPU-ARC: Agentic Reasoning Caching for on-CPU Reasoning ReusePeter Belcak et al.arXiv10
2026-11-052611.08449VFlash: Fast, Accurate Visual Planning via Flash Primitive DecodingYu-Cheng Chou et al.arXiv100
2026-11-042611.05306You don't need long contextPeter Belcak et al.arXiv14
2026-11-042611.03998Dory1.0-4B: A Looped Normalized-GPTMaxim Fishman et al.arXiv11
2026-10-272610.37666ThinkWorld: World Imagination Reinforces Chain-of-Dynamics ReasoningChaehyun Kim et al.arXiv64
2026-10-272610.36466Nemotron-Labs-Diffusion-Uni: Elastic Self-Refining Unified Diffusion LLM with Cross-Paradigm On-Policy DistillationShufan Li et al.arXiv30
2026-10-262610.31779FLORA: An Open Model Suite for Scaling Robotics Data EnginesSicong Leng et al.arXiv3
2026-10-232610.30340Foundation4D: Towards a unified perception and reconstruction model for autonomous vehiclesNVIDIA et al.arXiv14
2026-10-212610.28632Joints, Camera, Action! Co-Generating 3D Motion and Video for Arbitrary RigsNoam Rotstein et al.arXiv27
2026-10-192610.07898FC-SWE: Failure-Conditioned RL for Long-Horizon Software Engineering AgentsJia Liufu et al.arXiv8
2026-10-162610.17560EdgeWise: Bringing Long-Horizon Agents to the Edge via Selective Cloud DelegationZeru Shi et al.arXiv30
2026-10-142610.18596ACCRUE: Continual Off-Policy Reinforcement Learning for Robot Foundation ModelsChangyeon Kim et al.arXiv20
2026-10-142609.38377PhyProbe: Rethinking Physical Consistency Evaluation in Generated VideosMax Ku et al.arXiv15
2026-10-142610.17540TacitVLA: Reasoning Vision-Language-Action Models Without Reasoning LatencyJack Lu et al.arXiv4
2026-10-122610.11291When Do We Need On-Policy Distillation? Distilling on Offline Student Rollouts Is Often BetterSiyan Zhao et al.arXiv28
2026-10-122610.12582One Model to Loop Them All: Elastic Latent Recurrence for Language ModelsFaisal Ladhak et al.arXiv16
2026-10-122610.11215GATOR: Generative and Agentic 3D Object Reconstruction From Casual ImagesQirui Wu et al.arXiv2
2026-10-092606.03335A GPU-Parallel Framework for Heterogeneous Multi-Task Reinforcement LearningRui Zhang et al.arXiv22
2026-10-092610.11322USDCraft: Geometrically Grounded Programmatic Modeling of Articulated 3D Assets for SimulationChuanrui Zhang et al.arXiv17
2026-10-092610.09146Frozen Models, Evolving Expertise: Model-Agnostic Learning from Deployment Experience for Multimodal Medical AIYexiao He et al.arXiv3
2026-10-082610.10528Long-WAM: Scaling the Context of World-Action ModelsWei Huang et al.arXiv97
2026-10-072610.08430NeMo-DCR: Bit-Exact Delta-Compressed Refit for Scalable Agentic RL at Trillion-Parameter ScaleSonglin Jiang et al.arXiv28
2026-10-072610.08463UNREAL: Unifying Retrieval and Long-Context with a Single ModelEdan Kinderman et al.arXiv25
2026-10-072610.08761VeriFine: Scaling Verification for Self-Improvement in Embodied ReasoningZewei Zhou et al.arXiv9
2026-10-062610.06647LoGRA: Scaling LLM Reinforcement Learning with Low-Rank Gradient SketchesShaokun Zhang et al.arXiv45
2026-10-062609.24890OSWorld-Pro: Process-based Evaluation for Computer Use AgentsZhilin Wang et al.arXiv23
2026-10-062610.05750Beyond Semantic Similarity: Performance and Costs of Agentic Retrieval for Complex TasksReza Esfandiarpoor et al.arXiv17
2026-10-012609.39982Mid-Harness: Scaling Actions Between Model and Harness for Terminal AgentsMinki Kang et al.arXiv95
2026-10-012609.38597PixelUMM: Encoder-Free Unified Image and Video Understanding and GenerationCong Wei et al.arXiv33
2026-10-012609.40285PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn AgentsYinghui He et al.arXiv18
2026-10-012609.40358Physis-Lang: Self-Evolving Language as a Physical Representation for Video World ModelLiming Lu et al.arXiv16
2026-09-302609.38154LongLive-Plug: Once-for-All Distillation for Video GenerationShuai Yang et al.arXiv29
2026-09-302609.36864Where the Model Changes Its Mind: Hindsight-Divergence Localization for Efficient Reinforcement Learning with Verifiable RewardsFanchao Chen et al.arXiv7
2026-09-292609.35505An RL View of OPD: Least Square Policy Distillation for Sample-Efficient LLM ReasoningShangzhe Li et al.arXiv28
2026-09-292609.33419TT-VidT: Decoupling the Temporal Axis for Efficient Motion-Centric Video PretrainingShih-Ying Yeh et al.arXiv18
2026-09-182609.20519SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent HarnessHaozhe Liu et al.arXiv112
2026-09-172609.18094Agora: Git as Shared Memory for Collective AutoResearchYifan Zhang et al.arXiv35
2026-09-112609.10712An Open Recipe for IMO Gold: Training Nemotron for Olympiad MathematicsIvan Moshkov et al.arXiv21
2026-09-092609.07108Online Draft Co-Training for Speculative Decoding in Large-Scale, Long-Context RL Post-TrainingZili Wang et al.arXiv17
2026-09-092609.07941ReactVAU: A Slow-Fast Decoupled Framework for Streaming Video Anomaly UnderstandingChia-Hui Chen et al.arXiv4
2026-09-032609.02849Post-Training Language Models for Gold-Medal Performance in Coding CompetitionsAleksander Ficek et al.arXiv12
2026-08-242608.18077Hydra-0: Action Flow for Generalist World Modeling and ControlHongyu Li et al.arXiv8
2026-08-172608.10835UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal RepresentationsDvir Samuel et al.arXiv4
2026-08-142608.13391Context-Matched Distillation: Teacher Causality for Autoregressive Video DistillationHmrishav Bandyopadhyay et al.arXiv11
2026-08-102608.07408Addressable Memory for Video World ModelsXindi Wu et al.arXiv15
2026-07-302607.26410Voice Memory for Agentic Speech RecognitionChao-Han Huck Yang et al.arXiv10
2026-07-292607.26004Parallel Decoding Distillation for Fast Image and Video GenerationNeta Shaul et al.arXiv16
2026-07-282607.24027Sol-Attn: Accelerating Video Generation Inference via On-the-Fly Attention SparsificationHaopeng Li et al.arXiv38
2026-07-272607.21653Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement LearningJian Hu et al.arXiv33
2026-07-242607.21553SANA-Video 2.0: Hybrid Linear Attention with Attention Residuals for Efficient Video GenerationJunsong Chen et al.arXiv38
2026-07-242607.20709NVIDIA-labs OO Agents: Native Python Object-Oriented AgentsPaul Furgale et al.arXiv35
2026-07-202607.16107Audio-Visual Flamingo: Open Audio-Visual Intelligence for Long and Complex VideosSreyan Ghosh et al.arXiv9
2026-07-172607.15275RoboTTT: Context Scaling for Robot PoliciesYunfan Jiang et al.arXiv17
2026-07-102607.07740Jet-Long: Efficient Long-Context Extension with Dynamic Bifocal RoPEHaozhan Tang et al.arXiv24
2026-07-102607.08741ARDY: Autoregressive Diffusion with Hybrid Representation for Interactive Human Motion GenerationKaifeng Zhao et al.arXiv10
2026-07-092607.06838WildCity: A Real-World City-Scale Testbed for Rendering, Simulation, and Spatial IntelligenceXiangyu Han et al.arXiv13
2026-07-082607.03509Flex-Forcing: Towards a Unified Autoregressive and Bidirectional Video Diffusion ModelXinyin Ma et al.arXiv14
2026-07-082607.05722Nemotron-Labs-Diffusion: A Tri-Mode Language Model Unifying Autoregressive, Diffusion, and Self-Speculation DecodingYonggan Fu et al.arXiv11
2026-07-072607.05196Unified Audio Intelligence Without Regressing on Text IntelligenceZhifeng Kong et al.arXiv21
2026-07-022607.00272ASPIRE: Agentic /Skills Discovery for RoboticsRunyu Lu et al.arXiv24
2026-06-302606.29814Nemotron-Labs-Diffusion-Image: Advancing Masked Discrete Diffusion for High-Resolution Image SynthesisShufan Li et al.arXiv14
2026-06-302606.25621One Model, Many Latencies: Universal Speech Enhancement for Diverse Real-Time ApplicationsSzu-Wei Fu et al.arXiv5
2026-06-292606.28128PhysisForcing: Physics Reinforced World Simulator for Robotic ManipulationPeiwen Zhang et al.arXiv54
2026-06-292606.28276SimFoundry: Modular and Automated Scene Generation for Policy Learning and EvaluationNadun Ranawaka et al.arXiv16
2026-06-292606.20905Vesta: A Generalist Embodied Reasoning ModelJohan Bjorck et al.arXiv13
2026-06-252606.25473Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World ModelsKaiwen Zheng et al.arXiv27
2026-06-192606.19980ENPIRE: Agentic Robot Policy Self-Improvement in the Real WorldWenli Xiao et al.arXiv15
2026-06-192606.18231Adaptive Volumetric Mechanical Property Fields Invariant to ResolutionRishit Dagli et al.arXiv3
2026-06-172606.18216Zone of Proximal Policy Optimization: Teacher in Prompts, Not GradientsByung-Kwan Lee et al.arXiv64
2026-06-172606.17321ProCUA-SFT Technical ReportJaehun Jung et al.arXiv10
2026-06-162606.15007Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic ReasoningNVIDIA et al.arXiv19
2026-06-122606.13673SpatialClaw: Rethinking Action Interface for Agentic Spatial ReasoningSeokju Cho et al.arXiv115
2026-06-112606.04511SparDA: Sparse Decoupled Attention for Efficient Long-Context LLM InferenceYaosheng Fu et al.arXiv5
2026-06-092606.07723VoLo: A Physical Orchestrator for Open-Vocabulary Long-Horizon ManipulationSiyi Chen et al.arXiv6
2026-06-042606.02800Cosmos 3: Omnimodal World Models for Physical AIAditi et al.arXiv142
2026-06-042606.05160GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video PriorsTianyi Xie et al.arXiv9
2026-06-032606.03159NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle SimulationNVIDIA et al.arXiv21
2026-06-032606.03911Bootstrap Your Generator: Unpaired Visual Editing with Flow MatchingYoad Tewel et al.arXiv19
2026-06-022606.02553LongLive-RAG: A General Retrieval-Augmented Framework for Long Video GenerationQixin Hu et al.arXiv17
2026-06-022605.29318FreeForm: Reduced-Order Deformable Simulation from Particle-Based Skinning EigenmodesDonglai Xiang et al.arXiv2
2026-06-012605.30409SANA-Streaming: Real-time Streaming Video Editing with Hybrid Diffusion TransformerYuyang Zhao et al.arXiv38
2026-06-012605.31158Light Interaction: Training-Free Inference Acceleration for Interactive Video World ModelsJiacheng Lu et al.arXiv3
2026-05-292605.30161Why Far Looks Up: Probing Spatial Representation in Vision-Language ModelsCheolhong Min et al.arXiv53
2026-05-282605.28816Gamma-World: Generative Multi-Agent World Modeling Beyond Two PlayersFangfu Liu et al.arXiv95
2026-05-282605.28774Agent Explorative Policy Optimization for Multimodal Agentic ReasoningMinki Kang et al.arXiv89
2026-05-282605.23163Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous DrivingKewei Zhang et al.arXiv16
2026-05-272605.27365LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box DecodingShihao Wang et al.arXiv142
2026-05-252605.23902PiD: Fast and High-Resolution Latent Decoding with Pixel DiffusionYifan Lu et al.arXiv45
2026-05-222605.22791Gated DeltaNet-2: Decoupling Erase and Write in Linear AttentionAli Hatamizadeh et al.arXiv28
2026-05-202605.19786Fast 4D Mesh Generation by Spatio-Temporal Attention ChainsDvir Samuel et al.arXiv11
2026-05-192605.18739LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video GenerationYukang Chen et al.arXiv116
2026-05-152605.15178SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion TransformerHaoyi Zhu et al.arXiv88
2026-05-152605.14906MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language ModelsXiyu Ren et al.arXiv77
2026-05-142605.13724AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map DistillationYuchao Gu et al.arXiv101
2026-05-142605.05806Retrieval from Within: An Intrinsic Capability of Attention-Based ModelsElad Hoffer et al.arXiv10
2026-05-012604.24954Nemotron 3 Nano Omni: Efficient and Open Multimodal IntelligenceNVIDIA et al.arXiv16