---
title: "CogniLife: Continuous Cognitive Distillation for Deployed Multi-Expert Agent Marketplaces"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.04483v1"
permalink: "https://feed.astralor.com/items/itm_2faa4e10e41cf66ae3d916ff"
published: "2026-11-03T08:13:34.000Z"
score: 59
channel: "ai"
tags: ["LLM","AI Agent","CogniLife","Continuous Cognitive Distillation","Deployed Multi-Expert Agent","Marketplaces Multi-expert","Multi-expert","Creator Profiles"]
---

# CogniLife: Continuous Cognitive Distillation for Deployed Multi-Expert Agent Marketplaces

## At a glance

Multi-expert agent marketplaces -- where creators publish domain-specific AI experts for end users to consult -- are an emerging deployment pattern for language technology. A published expert can receive many requests y…

## Key points

- Multi-expert agent marketplaces -- where creators publish domain-specific AI experts for end users to consult -- are an emerging deployment pattern for language technology.
- A published expert can receive many requests yet little evidence about its creator, who remains active elsewhere: the cold-expert problem.
- We frame continuous cognitive distillation as converting attributed creator observations into revisable decision rules, communication patterns, and domain procedures.

## Full text

Multi-expert agent marketplaces -- where creators publish domain-specific AI experts for end users to consult -- are an emerging deployment pattern for language technology. A published expert can receive many requests yet little evidence about its creator, who remains active elsewhere: the cold-expert problem. We frame continuous cognitive distillation as converting attributed creator observations into revisable decision rules, communication patterns, and domain procedures. CogniLife proposes a cross-expert attribution design for this problem, with Creator Profiles for transferable creator patterns and Expert Journals for listing-specific knowledge; this design is not yet implemented or evaluated. A deployed platform partially realizes the surrounding lifecycle through creator-authorized authoring, capability-owned proposal filtering, and reversible consolidation of eligible owner-scoped memory. On ProfyBench-v1, an 85-task offline benchmark built from replay seeds and controlled internal tests, the Profy Full configuration as a whole achieves a mean overall score of 4.65/5 from the primary LLM judge, compared with 4.01 for Raw LLM and 4.28 for Expert Prompt, with the largest gains in completeness, specificity, and missing-information tasks. These configuration-level scores do not test cold-expert mitigation, cross-expert transfer, or longitudinal distillation.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_2faa4e10e41cf66ae3d916ff