ずんだもんのHugging Faceニュース

3:16

Daily AI Papers Briefing (2026-10-03)

ずんだもんのHugging Faceニュース ·

2026/10/03

【本日の論文】1. OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction https://huggingface.co/papers/2610.017622. Adaptive Reward Routing: Dynamic Multi-Reward Optimization for Joint Audio-Video Diffusion via Forward-Process RL https://huggingface.co/papers/2609.372003. On-Policy or Off-Policy Learning? A Systematic Study of Distillation Dynamics https://huggingface.co/papers/2609.352594. Beyond Memory: Harnessing Long-Horizon Agents with Explicit Belief

ja

3:24

Daily AI Papers Briefing (2026-10-02)

ずんだもんのHugging Faceニュース ·

2026/10/02

【本日の論文】1. The Teacher Is a Direction, Not a Destination: Extrapolating RL-Induced Representation Residuals in On-Policy Distillation https://huggingface.co/papers/2609.364842. UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement https://huggingface.co/papers/2609.387213. Rethinking Latent Visual Reasoning: Grounding Latent Reasoning in Visual Evidence https://huggingface.co/papers/2609.345634. False Frontiers: Diagnosing and Mitigating Co-Cheating in

ja

3:15

Daily AI Papers Briefing (2026-10-01)

ずんだもんのHugging Faceニュース ·

2026/10/01

【本日の論文】1. Raven: The Harness of Harnesses for Composable Agentic Intelligence https://huggingface.co/papers/2609.334392. In-Context Learning for Robots: Methods and Applications https://huggingface.co/papers/2609.360123. MaLiang-Harness: A Programmable Path to Image and Video Generation https://huggingface.co/papers/2609.343094. Scaling Properties of Same-Family On-Policy Distillation https://huggingface.co/papers/2609.327225. Omni-IO Skills: Harnessing Your Agent Omni-Native

ja

4:10

Daily AI Papers Briefing (2026-09-30)

ずんだもんのHugging Faceニュース ·

2026/09/30

【本日の論文】1. Post-Training Leaves Behavioral Shadows on Unrelated Decisions https://huggingface.co/papers/2609.292332. YuE2: Unifying Symbolic and Audio Music Generation at Frontier Quality https://huggingface.co/papers/2609.337573. Beyond Teacher Assignment: Domain-Normalized Multi-Teacher On-Policy Distillation https://huggingface.co/papers/2609.353474. Groupwise Agentic Grading and Advantage Redistribution for Code Agent RL https://huggingface.co/papers/2609.325775. Self-Evolving Coding

ja

2:50

Daily AI Papers Briefing (2026-09-29)

ずんだもんのHugging Faceニュース ·

2026/09/29

【本日の論文】1. FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders https://huggingface.co/papers/2609.316202. Disaggregated Quantization: Specializing LLM Prefill and Decode https://huggingface.co/papers/2609.263333. RayOrch: Programming and Executing Lineage-Controlled Multi-Grain Dataflows for Foundation-Model Data Preparation https://huggingface.co/papers/2609.187034. Block Sparse Attention with Log-Linear Complexity

ja

2:49

Daily AI Papers Briefing (2026-09-28)

ずんだもんのHugging Faceニュース ·

2026/09/28

【本日の論文】1. Training Object Permanence in World Models https://huggingface.co/papers/2609.286542. Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs https://huggingface.co/papers/2609.298453. WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation https://huggingface.co/papers/2609.302214. OmniEcho: Spatial Audio Understanding for Embodied Agents https://huggingface.co/papers/2609.234075. Agent-Editing World Model: Rethinking World

ja

3:43

Daily AI Papers Briefing (2026-09-27)

ずんだもんのHugging Faceニュース ·

2026/09/27

【本日の論文】1. Training Object Permanence in World Models https://huggingface.co/papers/2609.286542. Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs https://huggingface.co/papers/2609.298453. WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation https://huggingface.co/papers/2609.302214. OmniEcho: Spatial Audio Understanding for Embodied Agents https://huggingface.co/papers/2609.234075. Agent-Editing World Model: Rethinking World

ja

3:53

Daily AI Papers Briefing (2026-09-26)

ずんだもんのHugging Faceニュース ·

2026/09/26

【本日の論文】1. Training Object Permanence in World Models https://huggingface.co/papers/2609.286542. Your Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs https://huggingface.co/papers/2609.298453. WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation https://huggingface.co/papers/2609.302214. OmniEcho: Spatial Audio Understanding for Embodied Agents https://huggingface.co/papers/2609.234075. Agent-Editing World Model: Rethinking World

ja

4:08

Daily AI Papers Briefing (2026-09-25)

ずんだもんのHugging Faceニュース ·

2026/09/25

【本日の論文】1. SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue https://huggingface.co/papers/2609.267802. Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World https://huggingface.co/papers/2609.230383. HappyWorld-Bench https://huggingface.co/papers/2609.243084. The Past Frames the Future: Memory for Autoregressive Video Generation https://huggingface.co/papers/2609.284665. Just-in-Time Memory: Learning to Curate

ja

3:40

Daily AI Papers Briefing (2026-09-24)

ずんだもんのHugging Faceニュース ·

2026/09/24

【本日の論文】1. The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks https://huggingface.co/papers/2609.258042. RULER: Instance-aware Rubric Rewards for SVG Generation https://huggingface.co/papers/2609.252703. GAE: Learning a Geometry-Native Latent Space for 3D-Consistent World Generation https://huggingface.co/papers/2609.249814. All-in-One Multilingual Scene Text Recognition with Script-aware Mixture-of-Experts https://huggingface.co/papers/2609.240585. Circuit Hypernetworks

ja

2:57

Daily AI Papers Briefing (2026-09-23)

ずんだもんのHugging Faceニュース ·

2026/09/23

【本日の論文】1. RRSI: Regularized Recursive Self-Improvement of Agent Harnesses https://huggingface.co/papers/2609.249722. WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory https://huggingface.co/papers/2609.249843. GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplay https://huggingface.co/papers/2609.250014. Transferring the Intelligence of VLMs to Robotic Control https://huggingface.co/papers/2609.229665. OmniEdu: Open Foundation Models for Learning and

ja

3:56

Daily AI Papers Briefing (2026-09-22)

ずんだもんのHugging Faceニュース ·

2026/09/22

【本日の論文】1. CodeMidas: Scaling Agentic Coding RL Environments from Code Itself https://huggingface.co/papers/2609.220682. Grounded Skill Synthesis from Code at Scale for Agentic Intelligence https://huggingface.co/papers/2609.055713. EvoOntology: A Self-Evolving Ontology Layer for Data Agents https://huggingface.co/papers/2609.157794. RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents https://huggingface.co/papers/2609.220005. IntBMoE: Integrating Block-Level

ja

3:34

Daily AI Papers Briefing (2026-09-21)

ずんだもんのHugging Faceニュース ·

2026/09/20

【本日の論文】1. DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression https://huggingface.co/papers/2609.199692. Can MiniMax-H3 Reason About the Physical World? An Evaluation of Omni-Modal Generative Model https://huggingface.co/papers/2609.183233. When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation https://huggingface.co/papers/2609.205114. SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness

ja

3:13

Daily AI Papers Briefing (2026-09-20)

ずんだもんのHugging Faceニュース ·

2026/09/19

【本日の論文】1. DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression https://huggingface.co/papers/2609.199692. Can MiniMax-H3 Reason About the Physical World? An Evaluation of Omni-Modal Generative Model https://huggingface.co/papers/2609.183233. When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation https://huggingface.co/papers/2609.205114. SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness

ja

2:36

Daily AI Papers Briefing (2026-09-19)

ずんだもんのHugging Faceニュース ·

2026/09/18

【本日の論文】1. DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression https://huggingface.co/papers/2609.199692. SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness https://huggingface.co/papers/2609.205193. An Empirical Study of Harness Design for Coding Agents https://huggingface.co/papers/2609.208044. When EOS Tokens Disagree: Understanding Length Inflation in On-Policy Distillation https://huggingface.co/papers/2609.205115. JEPA-Anything: Learning Predictive

ja

3:52

Daily AI Papers Briefing (2026-09-18)

ずんだもんのHugging Faceニュース ·

2026/09/18

【本日の論文】1. LimiX-2: A Contextual Mechanism Network Towards General Structured-Data Intelligence https://huggingface.co/papers/2609.174882. ScienceIDE: Turning World's Scientific Codebase into Agent Learnable Environments https://huggingface.co/papers/2609.191343. Rethinking Critic Learning in PPO: Understanding and Mitigating Value Flattening https://huggingface.co/papers/2609.187084. Confidence Comes from Experience: Experiential Confidence Estimation from Reasoning to Agents

ja

4:22

Daily AI Papers Briefing (2026-09-17)

ずんだもんのHugging Faceニュース ·

2026/09/17

【本日の論文】1. Continual Learning Mechanisms Compose for Long-Horizon Memorization https://huggingface.co/papers/2609.069862. AI for Games in the Foundation Model Era https://huggingface.co/papers/2609.166793. StepAudio 3 Realtime Technical Report https://huggingface.co/papers/2609.140054. StepAudio 3 Music Technical Report https://huggingface.co/papers/2609.160345. ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents

ja

3:20

Daily AI Papers Briefing (2026-09-16)

ずんだもんのHugging Faceニュース ·

2026/09/16

【本日の論文】1. Vidu S2: Real-Time Interactive, Editable, and Spatial Video Generation https://huggingface.co/papers/2609.116382. Atria Dawn: The Dawn of Agentic Superintelligence https://huggingface.co/papers/2609.158183. ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search https://huggingface.co/papers/2609.133564. Dream-RSI: Recursive Self-Improvement through Evolving Worlds https://huggingface.co/papers/2609.148585. PhysBrain 1.5: From Vision-Language

ja

4:52

Daily AI Papers Briefing (2026-09-15)

ずんだもんのHugging Faceニュース ·

2026/09/15

【本日の論文】1. DataFlex-RL: An Evaluation Platform for RLVR Data Policies https://huggingface.co/papers/2609.061072. Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation https://huggingface.co/papers/2609.111153. Feyospace-v1: How the Cyber Mercury Seven Trained Frontier Cyber Models https://huggingface.co/papers/2609.084184. Breaking the Vision-Action Shortcut: Latent Interface Training for Generalizable Robotics Foundation Models

ja

5:18

Daily AI Papers Briefing (2026-09-14)

ずんだもんのHugging Faceニュース ·

2026/09/14

【本日の論文】1. NCP-ArchPreview Technical Report: Moving towards Latent Space Language Models through Next Concept Prediction https://huggingface.co/papers/2609.107152. SenseNova-U1.5: Towards Native Unified Visual Intelligence https://huggingface.co/papers/2609.119293. SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem https://huggingface.co/papers/2609.070644. EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agents

ja

All Episodes

Similar Podcasts

Lead With AI

Tamara Nall

The Paul Barron Crypto Show

Paul Barron Network

Double Tap

Steven Scott, Shaun Preece, Steven...

The Interface

BBC

The Lunduke Journal of Technology

Bryan Lunduke