Adapts off-policy RL stabilizers to the available data regime: introduces WarpSAC, a regime-aware family using Sample Weight Decay plus two regime-matched variants (WarpSAC-L and WarpSAC-A) to improve sample efficiency, wall-time learning, and sim-to-real deployment.
AI Video Papers2026
Provides an end-to-end, reproducible foundation for camera‑controllable, long‑horizon video world models — converting 1.43M clips from 10 datasets into a unified canonical corpus and releasing data, pipelines, recipes, and weights. Introduces backbone‑native adaptation and a three‑stage training recipe to produce 5B–33B models that enable minute‑to‑hour rollouts after training on 5s sequences.
AI Agent Papers2026
Reconstructs executable terminal workspaces from recorded agent trajectories and synthesizes verifiable single- and multi-round coding tasks for agent training; it replays file operations, uses an LLM completion agent to fill missing files/dependencies, and verifies tasks with autogenerated test suites.
Evaluates layer dropout (stochastic depth) in LLM pretraining, finds optimal layer distributions and time schedules that save ~20–25% training FLOPs at comparable or lower validation loss, and enables zero-shot layer skipping and self-speculative decoding for up to ~1.4–1.5× inference speedup.
AI Dataset2026
Simulation-ready home dataset for embodied AI: CAD-based household scenes with configured physical properties and metadata, plus 1,000 robot trajectory episodes (RGB-D, HDF5/USDZ) for simulation training and evaluation under CC BY-NC-SA 4.0.
AI Model2026
A pretrained autoregressive foundation LLM from Yandex that uses a hybrid MoE design (80B params, 3B active) and supports up to 262,144-token context. Notably strong on Russian factual knowledge, math and coding benchmarks; intended for research and fine-tuning rather than out-of-the-box production.
Turns implemented functionality in open-source codebases into executable RL training environments by automatically generating behavioral specifications, execution-grounded tests, and post-rollout filtering; yields 5,545 verifiable coding tasks across 23 languages to train coding agents.
AI Video Papers2026
Provides WROP: a 1.5M-sample synthetic video corpus and a 300-question exam for training and evaluating object permanence and solidity in video world models. Includes 150 Blender task generators, a human Elo benchmark across 14 models, and a fine-tuned 16B continuation model (PWM-WROP).
AI Video2026
Enables native joint video-and-audio generation and training up to 2K using a dynamic block-sparse attention scheme that cuts attention cost while preserving audio–visual coupling; provides preview checkpoints, training and inference code. High-resolution training requires large multi-GPU (80GB) clusters and CUDA ≥ 12.4.
Improves test-time scaling of looped transformers by adaptively assigning extra recurrent iterations to tokens that benefit most. TaH2 is a post-training method that jointly trains an iteration decider with the backbone using lookahead depth supervision, boosting the accuracy–compute slope and peak accuracy on challenging benchmarks.
Uses large-scale, mixed-task reinforcement learning to drive self-improvement of multimodal foundation models. Key features include fully-asynchronous large-batch RL (1,568 prompts × 16 rollouts, up to 3.7B tokens/update, 1M context), groupwise agentic grading, and sparse MoE architectures.
AI Agent Papers2026
Chengyi Yang, Renxu Bao +11Affiliation: Xiamen University, Affiliation: Washington University in St. Louis +2
Converts real repositories into verifiable, sandboxed training environments with black-box tests to train coding agents on long-horizon program reproduction. Provides Program-Env for automated environment construction and Program-Agent for a role-isolated development workflow, yielding 1,648 C/C++/Go/Rust tasks and improving benchmark performance.