Learns a discrete “physical language” from unlabeled videos and uses a reason-then-render pipeline: predict compact state-transition tokens, then decode them into future video. Separates dynamics inference from pixel synthesis to improve physical fidelity, controllable simulation, and zero-shot motion transfer.
AI Agent Papers2026
Evaluates vision-language model judges on computer-using agent (CUA) trajectories to measure verifier reliability. Provides OSReward-Hard and OSReward-Multi challenge sets, the OS-Shepherd-100K reasoning-annotated corpus, and trained OS-Shepherd reward models that match commercial judges at ~30–60× lower cost.
AI Model2026
An OpenAI-compatible LLM checkpoint optimized for agentic and long-context scenarios, shipping DSpark speculative decoding and vLLM/SGLang deployment recipes; tailored for code-agent and multi-step reasoning workloads and released under MIT.
AI Agent Papers2026
Reformulates long-horizon agent execution as explicit task-state management: a manager defines bounded subtasks, fresh-context executors run them, and read-only auditors verify outcomes. Shows large performance gains on WeaveBench, Terminal-Bench and OSWorld.
Supervises audio reasoning by generating per-sample, audio-grounded rubrics that evolve with model rollouts and serve as reinforcement-learning rewards, improving perception and adaptive multi-step reasoning while avoiding reward saturation.
Analyzes why supervised fine-tuning (SFT) causes severe task conflicts under multi-stage multi-task training while reinforcement learning (RL) enables stable coexistence, attributing the effect to sparse, near-orthogonal RL parameter updates and proposing Parallel-RL to decouple multi-task training.
AI Model2026
A 20B ternary-weight Mixture-of-Experts reasoning LLM optimized for on-device and low-memory inference—delivers high throughput (200+ tok/s on M4) and an extremely long 131k-context for math/logic benchmarks, but is a preview with limited agentic fine-tuning.
Detects and filters spurious token-level teacher supervision in on-policy distillation by estimating input-groundedness and removing high-impact misleading updates, improving OPD on both LLM and VLM benchmarks.
Evaluates how large language models fabricate user attributes in personalization and whether model self-monitoring is a reliable signal. Introduces MirageBench (150 personas, 6 personalization tasks, judge-validated faithfulness taxonomy) and a 12-model leaderboard revealing pervasive over-inference and a 'Self-Monitoring Inversion'.
AI Model2026
A 27B-parameter causal language model with a native vision encoder for image/video+text understanding, long-horizon agentic tasks, and tunable thinking-mode reasoning. Native 262,144-token context (extensible to 1,000,000) and production-focused inference recipes.
AI Video Papers2026
Evaluates VLMs' ability to form global spatial awareness from long-horizon egocentric video. Introduces GST-Bench: a VQA benchmark with human-verified questions from 6,790 minutes of synthetic video, reveals a large gap (best zero-shot 42.68 vs human 79.08) and provides GST-Train dataset.
Generates retrieval-centric Chain-of-Thought (RC-CoT) over initially retrieved candidates to improve unified multimodal retrieval via reranking or full-corpus re-retrieval with a dual-mode embedder. Trains an embedder–adviser framework (UniME-R1) using mined hard negatives, supervised learning, and retrieval-oriented reinforcement learning.