HaiAI123

Curated Global AI Tools Directory

Industry update
Flat

RoboScript enables one-shot robot learning with 20-point success boost

11/13/2026 — 11/13, 08:19·1 sources (all)·1 reports (all)

Key takeaways

  • Researchers introduced RoboScript, a method that lets robots learn tasks from a single demonstration plus simulator practice. It achieved 83…

Background & analysis

Researchers released a paper on November 9, 2026, introducing RoboScript, a system that lets a robot learn a task from a single demonstration followed by practice in a simulator, all without retraining the underlying model. According to a November 13, 2026 report on DEV Community, RoboScript achieved 83.4% average success across the 50 tasks in the RoboTwin 2.0 robot benchmark, outperforming the strongest trained robot policy by more than 20 points.

Instead of baking what it has learned into model weights, RoboScript records the steps and the role of each object in an imperative task memory, which is why the authors say the approach can carry over to new settings. The available material covers the paper's release and these benchmark results only, and does not describe any later deployment or engineering work built on top of the method.

AI-generated from 1 reports · updated 12 hours ago

updream
24-hour heatHeat index 69 · peak 97 · 11h ago
24 hours agonow

Reports on this story headlines open the original

Today
  1. Researchers present RoboScript, a system where a robot learns each task from a single demonstration plus simulator practice, reaching 83.4% average success across 50 RoboTwin 2.0 tasks and beating the strongest trained policy by more than 20 points without changing the model's weights. What it learns is stored as instruction-level task memory about steps and object roles, so it carries over to new scenes. The paper was posted on Nov. 9.

    DEV Community · AIAI score 59

Other stories people are talking about

How is heat calculated?About the method

Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.

This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.

Surge
Discussion rising fast
New
First report within 6 hours
Rising
Still gathering discussion

Back to AI industry updates →