National University of Singapore · 61 records

dateidtitleauthorsvenue↑
2026-11-202611.26617One Prior, Two Dynamics: Repurposing Video Diffusion Transformers for Self-Supervised Latent Action LearningYuxin Jiang et al.arXiv2
2026-11-192611.23499BiCAT: Bimanual Coordination Action TokenizationXuanwei Liu et al.arXiv3
2026-11-182611.24738Jeveration: System One Models Now GenerateYuhe Liu et al.arXiv146
2026-11-132611.18482MMFlash: Faster Speculative Decoding for Vision-Language ModelsSicheng Feng et al.arXiv39
2026-11-12W7148678813Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approachAnurag K. S. V. et al.The Journal of Chemical Physics
2026-11-112611.15777SO-Arena: Benchmarking the Capability Spectrum and Reliability of System One ModelsLichen Li et al.arXiv135
2026-11-112611.12200Learning to Act with World Models by Valuing What Comes NextChenghao Gao et al.arXiv3
2026-10-302610.42148Flide: Accelerating Decision Inference with Exact State ForkingBobo Li et al.arXiv35
2026-10-282610.36192Beyond the Estimate: Wiener-Informed Decoding for Easier Training of Pixel-Space Flow ModelsZiqiao Wang et al.arXiv6
2026-10-272610.35434Skill^{infty}: Recursive Self-Improvement for Agent Skill OptimizationWangcheng Tao et al.arXiv8
2026-10-262610.36077The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal AgentsWeidong Tang et al.arXiv44
2026-10-262610.29552Generalized CycleGRPO for Fine-Grained Multimodal Understanding and GroundingXin Zhang et al.arXiv14
2026-10-252610.34467DynaFlow: Learning Free-Form Multi-Agent OrchestrationRuhao Liu et al.arXiv47
2026-10-232610.29677Toward Recursive Self-Improvement in Physical AIMeng Luo et al.arXiv32
2026-10-212610.18604LLM2R: From Large Language Models to 3D ReassemblersZhouhan Zhong et al.arXiv25
2026-10-192610.20628PRISM: PRIvileged Sibling-guided Multi-Agent Topology DesignPhuong Tuan Dat et al.arXiv10
2026-10-192610.21467DuET: Note-Frame Dual Encoding for Piano Performance-to-Score TranscriptionWei Zeng et al.arXiv3
2026-10-162610.22311Velox: Spatial Memory Makes World Models EfficientQiuhong Shen et al.arXiv35
2026-10-162610.22282Routing Policy Optimization for Online Diffusion Reinforcement LearningQiuhong Shen et al.arXiv29
2026-10-142610.18773Jeva: System One Models from First PrinciplesQi Li et al.arXiv180
2026-10-142610.18206SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding AgentsDo Xuan Trong et al.arXiv14
2026-10-092610.08215Learn2Play Bench: How Well Do LLM Agents Learn from Experience in Unfamiliar Environments?Yibo Li et al.arXiv55
2026-10-072610.07753From Evidence to Action: How Tool-Using Agents FailHongzhan Lin et al.arXiv45
2026-10-052609.38839FrameMorrow: Future-guided Frame Selection with Prospective Tokens for Long-Horizon Video GenerationBo Yin et al.arXiv55
2026-10-052609.26425QuantWM: Temporally Consistent 2-Bit KV Cache Quantization for Video World ModelsJiaqi Zhao et al.arXiv11
2026-10-022609.35690Agent Priors-guided Policy LearningPuming Jiang et al.arXiv54
2026-09-302609.34309MaLiang-Harness: A Programmable Path to Image and Video GenerationHaoyu Zhao et al.arXiv97
2026-09-302609.31847Omni-IO Skills: Harnessing Your Agent Omni-NativeYanlin Li et al.arXiv63
2026-09-302609.33627StoryEngine: A State-Grounded Agentic Framework for Video StorytellingYingrui Wang et al.arXiv8
2026-09-292609.33203Structured Residual Connectivity Matters for Diffusion TransformersYuhe Liu et al.arXiv16
2026-09-222609.23863Grounded Action Model: 3D Grounding as a Foundation for RoboticsGehao Zhang et al.arXiv30
2026-09-222609.22220Measuring the Checker: Mutation Analysis for GPU-Kernel Benchmark OraclesMingzhe Du et al.arXiv6
2026-09-182609.20612What Does Privileged Information Add to On-Policy Self-Distillation?XiuYu Zhang et al.arXiv11
2026-09-162609.16679AI for Games in the Foundation Model EraMeng Luo et al.arXiv124
2026-09-142609.12641Breaking the Vision-Action Shortcut: Latent Interface Training for Generalizable Robotics Foundation ModelsJianman Lin et al.arXiv64
2026-09-072609.04061When Models Edit Too Much: On the Fidelity of Minimal Code EditsTongyao Zhu et al.arXiv8
2026-08-312608.25697LMSM: LLM Security Framework Inspired by Linux Security ModulesXiuYu Zhang et al.arXiv6
2026-08-282608.25518Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World ModelsPengfei Zhou et al.arXiv35
2026-08-262608.23566Best Practice Critic OptimizationPenghui Qi et al.arXiv14
2026-08-202608.13558OmniScientist: An Omni-Modal Omni-Discipline AI ScientistBobo Li et al.arXiv29
2026-08-192608.17310Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU RequirementsZhi Zheng et al.arXiv76
2026-08-192608.13556V-RAE: Rethinking Video Latent Spaces for GenerationMinghui Guo et al.arXiv29
2026-08-172608.13040Latent On-Policy Self-DistillationGuibin Zhang et al.arXiv25
2026-08-052608.03972ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct ReasoningJinhe Bi et al.arXiv6
2026-08-042608.01397SG-WAM: Self-Guided World Modeling in Geometry-Aware Policy SpaceRuiteng Zhao et al.arXiv11
2026-08-032607.26991RL^2-VLA: Adaptive RL Latent Compositional Steering with Test-Time Scaling for Vision-Language-Action ModelsDerek Ming Siang Tan et al.arXiv6
2026-07-082607.02963Parallelized Autoregressive Decoding for Omni-Modal Dense Video CaptioningWenzheng Zeng et al.arXiv26
2026-07-082607.05765Image2Sim: Scaling Embodied Navigation via Generative Neural SimulatorZihan Wang et al.arXiv3
2026-07-022607.00529NoPA: Non-Parametric Online 3D Scene Graph GenerationQi Xun Yeo et al.arXiv4
2026-06-182606.20781World Action Models: A SurveyQiuhong Shen et al.arXiv59
2026-06-172606.17628OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy DistillationGuibin Zhang et al.arXiv29
2026-06-152606.06036Memory is Reconstructed, Not Retrieved: Graph Memory for LLM AgentsShuo Ji et al.arXiv76
2026-06-152606.01462An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning ModelsMingzhong Sun et al.arXiv4
2026-06-102606.10572One Token per Multimodal Evidence: Latent Memory for Resource-Constrained QAZhi Zheng et al.arXiv18
2026-06-092606.05122Self-Evaluation Is Already There: Eliciting Latent Judge Calibration in Base LLMs with Minimal DataXiuYu Zhang et al.arXiv2
2026-06-092606.05753Cosine Misleads: Auxiliary Losses Reshape Vision Language Models, Not Their LatentsXiuYu Zhang et al.arXiv1
2026-06-042606.05121Audio Interaction ModelZhifei Xie et al.arXiv88
2026-05-222605.21072Q-ARVD: Quantizing Autoregressive Video Diffusion ModelsSiao Tang et al.arXiv18
2026-05-212605.19833Mega-ASR: Towards In-the-wild^2 Speech Recognition via Scaling up Real-world Acoustic SimulationZhifei Xie et al.arXiv132
2026-05-212605.20315Mix-Quant: Quantized Prefilling, Precise Decoding for Agentic LLMsHaiquan Lu et al.arXiv25
2026-05-082605.04045Audio-Visual Intelligence in Large Foundation ModelsYou Qin et al.arXiv31