Personalizing How Recommendation Agents Reason: Self-Evolving Per-User Skills for Recommendation Agents
Year: 2026
Type: conference-paper
Abstract: LLM-based recommendation agents personalize what they know through evolving peruser semantic memory, yet how they perform reasoning relies on a universal, static system prompt shared across all users.When a recommendation fails, the agent only updates its memory, but never examines or revises the reasoning that leads to the failure.To close this gap, we propose SAGER (Self-Evolving AGent for PErsonalized Recommendation), the first framework to equip each user with a dedicated, continuously evolving policy skill.SAGER incorporates three tailored mechanisms for recommendation scenarios: (i) a tworepresentation skill architecture constrained by a Cognitive Injection Boundary, decoupling a rich evolution substrate from a minimal inference-time injection; (ii) a contrastive differential chain-of-thought engine that diagnoses ranking errors against unchosen nearmisses under a refine-not-replace principle; and (iii) skill-augmented listwise reasoning that mitigates the Score-Tie Bottleneck, enabling the slim skill to serve as an active tie-breaker.Evaluated on four public benchmarks, SAGER achieves state-of-the-art accuracy.Furthermore, plugging its policy skill layer into three structurally distinct host agents delivers consistent 9%-12% gains on the H@1 metric.These results indicate that policy skill evolution provides architecture-agnostic improvements that are largely complementary to memory accumulation.
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
Funder: Tencent
OpenAlex ID: W7224606331