Back to AI News
arxiv

System-One RSI: Guided Practice for Fast Interactive Agents

System-One RSI is a framework that pairs a high-frequency OODA control loop, using a decision model and decision tree, with an offline recursive improvement loop run by reasoning agents. The offline team revises prompts and decision trees, derives supervision from interaction records, and retrains or redeploys updated decision parameters, aiming to build external "muscle memory" without invoking a reasoning model at action time. Using a 27B decision model, experiments on Snake, Zenless Zone Zero, and Black Myth: Wukong report a Snake mean score rise from 1.30 to 7.35, an 8.4% median completion-time reduction in one ZZZ scenario, and improved combat survival and win rate in one Wukong scenario.

AI Brief

It offers a concrete self-improvement loop for low-latency interactive agents and reports measured gains across three game environments.

arXiv:2611.26007v1 Announce Type: new Abstract: Reasoning models provide strong multimodal task understanding but struggle with sub-second control because deliberation increases decision latency. Fast decision models can meet the response window, yet prompting and few-shot examples do not readily provide the fine-grained experience required by a new task. Inspired by coached motor learning, this paper presents System-One Recursive Self-Improvement (System-One RSI), which couples two loops. The high-frequency online inner loop is an Observe, Orient, Decide, Act (OODA) cycle: a decision model maps each visual observation and task question to structured judgments, and a decision tree converts them into timed actions without invoking a reasoning model. The offline outer loop is a recursive improvement cycle carried out by a team of reasoning agents. The team revises the decision-model prompt and decision tree, captures interaction records, derives supervision from those records, and trains or deploys updated model parameters. The improved system then returns to the environment and generates the experience for the next round, consolidating learned distinctions as external muscle memory. The framework operates with a small number of real episodes, without expert demonstrations, environment modification, sandboxing, or parallel simulation. Experiments in Snake, Zenless Zone Zero, and Black Myth: Wukong use a 27B decision model. Mean Snake score increases from 1.30 to 7.35. In an evaluated scenario of Zenless Zone Zero, median completion time decreases by 8.4%. In an evaluated combat scenario of Black Myth: Wukong, the evasive policy achieves 11 times the mean survival time of a stationary reference, while the offensive policy win rate rises from 0% to 66.7% across the evaluated configurations.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!