Proposes Recuris, a recursive Experiential-Working Memory architecture that separates Working Memory (task progress) from Experiential Memory (skills) and uses a Meta-Agent to validation-gate localized skill updates, enabling bounded recursive skill evolution for long-horizon agents.
Proposes treating game development as a recursive data engine and introduces RLHEV (Reinforcement Learning with Human-Engine Verification) to combine dense engine checks (collision, physics, navigability) with human acceptance feedback, producing trajectory data and rewards for post-training world models.
Provides a streaming dual-brain memory for real-time speech agents: an informational left brain for factual retrieval and an affective right brain for persona/emotion, achieving high top-5 accuracy while keeping retrieval latency within VAD budgets (~134 ms).
Synthesizes, repairs, and self-evolves task-adaptive agent harnesses on demand for off-the-shelf LLM agents, using a trainable harness-intelligence model that distills signals from past configurations. Demonstrates consistent performance gains across benchmarks and model families by producing four-module, composable harnesses.
Analyzes how to generate useful interaction data for LLM agents and proposes the ACE lens — Accuracy, Complexity, divErsity — while factorizing agentic data as (E, q, τ, v). Surveys verification, difficulty calibration, and coverage strategies and outlines implications for training and benchmarks.
Decides when past post-training updates should be reused for autonomous LLM adaptation by introducing Boundary-Calibrated Intervention Transfer (BCIT). BCIT binds effects to source context, checks applicability and hard conflicts, and runs bounded trials to obtain current-state evidence—reducing harmful updates and improving equal-budget final-model quality.
Rewrites physical scenes as executable world programs (e.g., MuJoCo scene descriptions) and uses an agentic abductive loop to propose, execute, render, verify, and iteratively refine those programs from videos or text. Verified executable worlds supply scalable physical supervision for training vision–language models.
Autonomous multimodal GUI agent that executes natural-language interface tasks across mobile apps, web domains, and desktop OS. Expands environment coverage (170+ multilingual apps, 4,000+ web domains), uses function-grounded task generation and keypoint-based multi-model verification to produce reliable RL rewards for real-world deployment.
Trains LLM agents to proactively edit and manage their working context for long-horizon tasks using an expanded toolset (planning, long-term memory, soft offloading) and a fine-grained RL algorithm that identifies critical edits and assigns action-level credit. Improves accuracy while keeping contexts compact on long-context QA and deep search.
Provides ~483K agent instruction‑tuning trajectories for supervised fine‑tuning, including tool calls, environment feedback, errors/retries and verification across search, code, office and general agent workflows; static snapshots for SFT and mix‑ratio studies.
Turns each research paper into a training environment to generate verifiable research plans by synthesizing questions from goals/background and deriving evaluation criteria from methods/experiments. Key features: four-stage extraction that reduces criterion leakage to 3.7%, a two-stage rubric-centered training (self-distillation then GRPO), and the PaperGym-20k corpus with two held-out benchmarks.
Wraps vision–language–action policies into executable skills that are runtime-validated, executed as bounded low-level action chunks, outcome-verified, and logged as structured trajectories. A fixed skill interface enables swapping or adapting low-level VLA policies and provides component-level supervision for training and optional online adaptation.