Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments

Key points

• Systems such as ACE and Dynamic Cheatsheet keep it as a playbook: lessons the agent writes after each task and reads in its prompt.1
• A key challenge is to decide which lessons improve later tasks.1
• We propose Randomized Block Curation (RBC), which measures this effect with a randomized experiment on practice tasks.1
• Across two memory systems, three models and three benchmarks, RBC raises average success to 66.9, against 55.5 with the whole playbook, 54.1 with no playbook and 53.8 when the agent's own judgment picks the lessons to keep.1

Sources (1)

• [1]Reliable Memory from Unstable Growth: Curating Self-Written Agent Playbooks with Randomized Experiments
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 02:06 PM

  Persistent textual memory lets language-model agents reuse lessons from past tasks to improve on future multi-step tasks without retraining.

  Systems such as ACE and Dynamic Cheatsheet keep it as a playbook: lessons the agent writes after each task and reads in its prompt.

Extractive summary: sentences quoted from the sources.