Daily Papers of 2026-04-17

  1. HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds 110 upvotes, #1 of 2026-04-17
  2. DR^{3}-Eval: Towards Realistic and Reproducible Deep Research Evaluation 35 upvotes, #2 of 2026-04-17
  3. How to Fine-Tune a Reasoning Model? A Teacher-Student Cooperation Framework to Synthesize Student-Consistent SFT Data 34 upvotes, #3 of 2026-04-17
  4. RAD-2: Scaling Reinforcement Learning in a Generator-Discriminator Framework 28 upvotes, #4 of 2026-04-17
  5. Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems 24 upvotes, #5 of 2026-04-17
  6. GlobalSplat: Efficient Feed-Forward 3D Gaussian Splatting via Global Scene Tokens 24 upvotes, #5 of 2026-04-17
  7. HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System 20 upvotes, #7 of 2026-04-17
  8. ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack 19 upvotes, #8 of 2026-04-17
  9. UniDoc-RL: Coarse-to-Fine Visual RAG with Hierarchical Actions and Dense Rewards 15 upvotes, #9 of 2026-04-17
  10. LeapAlign: Post-Training Flow Matching Models at Any Generation Step by Building Two-Step Trajectories 12 upvotes, #10 of 2026-04-17
  11. Boosting Visual Instruction Tuning with Self-Supervised Guidance 11 upvotes, #11 of 2026-04-17
  12. KV Packet: Recomputation-Free Context-Independent KV Caching for LLMs 10 upvotes, #12 of 2026-04-17
  13. Switch-KD: Visual-Switch Knowledge Distillation for Vision-Language Models 9 upvotes, #13 of 2026-04-17
  14. Representations Before Pixels: Semantics-Guided Hierarchical Video Prediction 8 upvotes, #14 of 2026-04-17
  15. OneHOI: Unifying Human-Object Interaction Generation and Editing 8 upvotes, #14 of 2026-04-17
  16. Reinforcement Learning via Value Gradient Flow 7 upvotes, #16 of 2026-04-17
  17. TRACER: Trace-Based Adaptive Cost-Efficient Routing for LLM Classification 7 upvotes, #16 of 2026-04-17
  18. Don't Retrieve, Navigate: Distilling Enterprise Knowledge into Navigable Agent Skills for QA and RAG 7 upvotes, #16 of 2026-04-17
  19. LongAct: Harnessing Intrinsic Activation Patterns for Long-Context Reinforcement Learning 7 upvotes, #16 of 2026-04-17
  20. Cross-Tokenizer LLM Distillation through a Byte-Level Interface 6 upvotes, #20 of 2026-04-17
  21. Towards Autonomous Mechanistic Reasoning in Virtual Cells 6 upvotes, #20 of 2026-04-17
  22. RadAgent: A tool-using AI agent for stepwise interpretation of chest computed tomography 6 upvotes, #20 of 2026-04-17
  23. MM-WebAgent: A Hierarchical Multimodal Web Agent for Webpage Generation 6 upvotes, #20 of 2026-04-17
  24. SuperLocalMemory V3.3: The Living Brain -- Biologically-Inspired Forgetting, Cognitive Quantization, and Multi-Channel Retrieval for Zero-LLM Agent Memory Systems 5 upvotes, #24 of 2026-04-17
  25. Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes 5 upvotes, #24 of 2026-04-17
  26. C2: Scalable Rubric-Augmented Reward Modeling from Binary Preferences 4 upvotes, #26 of 2026-04-17
  27. Three-Phase Transformer 3 upvotes, #27 of 2026-04-17
  28. An Optimal Transport-driven Approach for Cultivating Latent Space in Online Incremental Learning 2 upvotes, #28 of 2026-04-17
  29. Model Capability Dominates: Inference-Time Optimization Lessons from AIMO 3 2 upvotes, #28 of 2026-04-17

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.