📑Alibaba: "오디오, 이제 말로 편집합시다"

이 호에서 다룬 논문

• RL

  AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing

  Alibaba (Tongyi Lab), SJTU · Speech Foundation Model, Instruction Editing, Rectified-Flow

• AGENTS

  NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness

  Huawei (Noah's Ark Lab) · Recursive Self-Improvement, Agentic Post-Training, Routing Harness

• MULTIMODAL

  SenseNova-U1.5: Towards Native Unified Visual Intelligence

  SenseTime · Unified Multimodal, Encoder-free, Native 4K

• AGENTS

  T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks

  Tencent · Terminal Agent, Long-Horizon RL, Mixture-of-Experts

• AI RESEARCH

  Large Discrete Policy: Advancing Explicit Behavior Modeling with Stochastic Iterative Scoring

  NVIDIA, Fudan University · Behavior Policy, Discrete Action, Autonomous Driving

• REASONING

  An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics

  NVIDIA · Olympiad Math, Test-time Compute, Nemotron

• REASONING

  MindTopo: Can Foundation Models Reason in Topological Space

  Microsoft Research, Stanford · Topological Reasoning, Spatial Intelligence, MLLM Benchmark

• REASONING

  Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning

  Cohere · Multilingual Reasoning, Data Mixing, In-Language

• AGENTS

  Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails

  Salesforce AI Research · Agent Harness, On-Policy Correction, Enterprise Agents

• AGENTS

  ExecCritic: Learn to Test, Test to Improve for Coding Agents

  Microsoft Research · Coding Agent, Test Generation, Role-specific RL