The authors formalized Personal-Agent Mediated Recommendation: a platform first ranks candidates using platform-local information, then a personal agent (an LLM) mediates that ranking using user-authorized cross-platform history to produce the final slate.
Personal LLM agents improve recommendations using cross-platform history
The authors formalize Personal-Agent Mediated Recommendation, introduce the MediateRec benchmark, and propose PAMO training to balance rescues against harmful overrides.
Big Tech
Yu Xia · Jiangfan Zhang · Jun Xiao · Julian McAuley · Xiangjun Fan
University of California San Diego · Meta AI
Research Digest··3 min read
The authors formalize a new recommendation paradigm in which a personal LLM agent uses cross-platform user history to mediate a platform's candidate ranking before delivering the final top-K slate.
Why this paper
From Meta AI and University of California San Diego
In one line
PAMO trains personal agents to selectively override platform rankings using cross-platform history, outperforming matched outcome-only reinforcement learning on accuracy and rescue-harm balance.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§