Memorizing Who, Knowing How: Benchmarking Personal and Procedural Memory for Persistent Assistants

Key points

• Memory systems for persistent personal agents must move beyond isolated memories, maintaining both personal memory about whom they serve and procedural memory about how to act.1
• Crucially, in harness-based personal assistance, procedures or skills are inherently user-conditioned: personal context determines which procedures apply, how they are instantiated, and when they should change.1
• More broadly, across existing methods, relation-related failures increase from early to late sessions as interaction history grows (59.8% to 72.9%), revealing a continual-maintenance challenge: writing correct memories is not enough if their evolving content and organization cannot be preserved for later retrieval.1
• To address this, we propose PraxisMem, which uses privileged joint-state supervision to internalize the ability to maintain memory content and organization into the memory manager, with location-aware GRPO and dynamic residual-routed OPSD.1

Sources (1)

• [1]Memorizing Who, Knowing How: Benchmarking Personal and Procedural Memory for Persistent Assistants
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 11:33 AM

  We introduce PraxisMemBench, a harness-based benchmark for persistent personal assistance across daily-life and work trajectories, where personal and procedural memory evolve together and later tasks depend on both.

  Memory systems for persistent personal agents must move beyond isolated memories, maintaining both personal memory about whom they serve and procedural memory about how to act.

Extractive summary: sentences quoted from the sources.