Daily Papers of 2026-11-09

  1. Confucius4-R2T2: A Stable-Prefix Streaming ASR Model for Real Real-Time Transcription 92 upvotes, #1 of 2026-11-09
  2. SplatAnything: Gaussian Spacing Normalization for Feed-forward 3DGS across Arbitrary Scenes 69 upvotes, #2 of 2026-11-09
  3. P^2-MOPD: Accelerating Multi-Teacher On-Policy Distillation via Parameter Priors 68 upvotes, #3 of 2026-11-09
  4. Toward the Era of Agentic Embodied Intelligence 53 upvotes, #4 of 2026-11-09
  5. PCBRouteAgent: How Close Are LLM Agents to Becoming PCB Routing Engineers? 42 upvotes, #5 of 2026-11-09
  6. OneTrans-GSR: A Generative Slate Ranking Paradigm Unifying Reranking and Mix-Ranking with One Autoregressive Transformer 41 upvotes, #6 of 2026-11-09
  7. LLMs Don't Know What They Know: Benchmarking Temporal Self-Knowledge for Financial Backtesting 34 upvotes, #7 of 2026-11-09
  8. LongBanana: An Expert-Verified Benchmark for Long-Context Multi-Reference Image Synthesis 31 upvotes, #8 of 2026-11-09
  9. Frontier3DBench: Benchmarking Coding Agents for 3D Creation across Objects, Scenes, and Interactive Open Worlds 28 upvotes, #9 of 2026-11-09
  10. EvoWorld: Towards Out-of-Sight World Evolution in Generative Video Models 25 upvotes, #10 of 2026-11-09
  11. Beneath the Surface: What We Overlook When We Talk About Agent Evolution 25 upvotes, #10 of 2026-11-09
  12. Fast and Exact Relative-Position Attention via Structure-Aware Kernels 23 upvotes, #12 of 2026-11-09
  13. AgentBaton: Efficient Model Collaboration in Agentic Tasks 22 upvotes, #13 of 2026-11-09
  14. Video-Coder: Reinforcing Agentic Coding for Video Generation 22 upvotes, #13 of 2026-11-09
  15. Exploring Weight-Space Programming: Capabilities, Generalization, and Limitations 19 upvotes, #15 of 2026-11-09
  16. LowRAM: Low-Rank Keys and Amortised Values for Composable KV Cache Compression at Test-Time 16 upvotes, #16 of 2026-11-09
  17. AutoCompress: Automated Hardware-Guided Model Compression 12 upvotes, #17 of 2026-11-09
  18. Steerable Multi-Objective Group-Relative Preference Optimization with Jointly Trained Low-Rank Adapters 11 upvotes, #18 of 2026-11-09
  19. CPU-ARC: Agentic Reasoning Caching for on-CPU Reasoning Reuse 11 upvotes, #18 of 2026-11-09
  20. Hierarchical Residuals: A Minimalist Decoupling of Residual Width from Compute Width 9 upvotes, #20 of 2026-11-09
  21. Authorization Drift in LLM Agents: Evaluating Boundary Compliance in Multi-Step Tool Use 7 upvotes, #21 of 2026-11-09
  22. "PRISM" Is Not All You Need: Cross-Topic Name Collisions and Scientific Paper Retrieval 6 upvotes, #22 of 2026-11-09
  23. Human-Centred, Risk-Constrained Evaluation of Financial Document Recognition: Evidence, Critical Errors, and Selective Automation 6 upvotes, #22 of 2026-11-09
  24. The Last Mile to the Real World: Agent-Native Primitives as Reusable Building Blocks for Solving Real-World Tasks 6 upvotes, #22 of 2026-11-09
  25. PREEN: Exemplar-Based Color Style Transfer with Smartphone Editor Presets 5 upvotes, #25 of 2026-11-09
  26. Knowing When to Teach: Learning from Teachable Prefixes in Multilingual Reasoning Distillation 5 upvotes, #25 of 2026-11-09
  27. Coding or Cheating? An Investigation of Frontier Coding LLMs 5 upvotes, #25 of 2026-11-09
  28. Sketchpad: certified exact arithmetic with a small vision-language model 4 upvotes, #28 of 2026-11-09
  29. IWIBench: Evaluating Gradient-Free Model Driven Intra-weight Injection 4 upvotes, #28 of 2026-11-09
  30. OmniLatent: Unified Latent for 3D Part Modeling 4 upvotes, #28 of 2026-11-09
  31. Learning to evolve by matching velocities 2 upvotes, #31 of 2026-11-09
  32. Where to Connect? Boosting MLLMs via Dynamic Gated Pathways across ALL ViT and LLM Layers 2 upvotes, #31 of 2026-11-09
  33. Static Recovery Is Not Dynamic Stability: Dynamics-Aware Benchmarking of Protein Motif Scaffolding 2 upvotes, #31 of 2026-11-09
  34. ODE Policy: Robot Learning via Action Dynamics, Not Actions 1 upvotes, #34 of 2026-11-09

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.