TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

robotics

The Robot Remembered the Scene but Mishandled the Update

MEMOBench labels storage, update and compression separately across 4,200 checkpoints; its strongest memory baseline averaged 31.9 percent success.

Published Updated Story ID: mp-2026-09-09-018
Read the complete editionStory JSON

Summary

MEMOBench labels storage, update and compression separately across 4,200 checkpoints; its strongest memory baseline averaged 31.9 percent success.

The suite contains 30 history-dependent manipulation tasks, 1,500 expert demonstrations and checkpoints from 84 templates. High storage scores often coexisted with weak updating and compression, separating memory errors from manipulation failure.

Why it matters

MEMOBench labels storage, update and compression separately across 4,200 checkpoints; its strongest memory baseline averaged 31.9 percent success.

Limits and context

No additional limitation was separately recorded.

Key claims

  1. MEMOBench labels storage, update and compression separately across 4,200 checkpoints; its strongest memory baseline averaged 31.9 percent success.

    Evidence: source-2026-09-09-020

Sources

  1. arXiv preprint 2609.07047arXiv · primary research

Corrections

No corrections have been recorded for this story.