Llm recommendation improves accuracy with training-free dual-profile communication
HARP: Training-Free Dual-Profile Agentic Communication for LLM-Based Recommendation
Information RetrievalArtificial Intelligence
Summary
Recommendation systems often try to guess what users like but usually only use detailed profiles for users, not items. The authors look at recommendation as a two-way chat where both users and items have descriptions that are shared and updated over time. They create HARP, a new method that adds item profiles and updates user profiles without retraining the model, improving recommendation accuracy on multiple public datasets. This approach helps recommendation systems understand both sides better and adapt without extra training.
What this means in practice
- •For software engineers: Enhance recommendation platforms by integrating symmetrical user-item profiles updated in real time without retraining underlying models.
- •For ecommerce platform developers: Improve product recommendations by adding item self-descriptions and profile-level homophily at retrieval, increasing accuracy without extra model training.$Commercial implications: Enables building more accurate recommendation engines for online stores by using training-free methods for item and user profile communication.
Authors
Zhen Tao, Xun Zhou, Xuhui Chen, Ziyue Qiao, Qingqiang Sun
Abstract
In LLM-based recommendation, memory acts as a user-level agent in a memory-augmented backbone. Yet users and items remain asymmetric: only the user is given a structured profile; it enters a bandwidth-limited reranker prompt; and it is frozen after warmup. We recast LLM-based recommendation as an agentic-communication process in four stages: BUILD a self-description, TRANSMIT it along the collaborative graph, UNDERSTAND the other side, and SYNC it from the next interaction. This formulation reveals three unresolved communication challenges: the item side never produces a comparable self-description (one-sided BUILD); profile-level signals enter only the bandwidth-limited reranker prompt rather than retrieval, where they would not compete for context tokens (prompt-trapped UNDERSTAND); and the user's description is frozen after warmup with no principled update rule (absent SYNC). We instantiate the missing stages as HARP (Homophily-aware Agentic Recommendation via dual Profiles), a training-free, three-operator addition to a frozen LLM backbone: a rule-extracted symmetric item profile (zero extra LLM calls), a channel-decomposable profile-level homophily H(phi_u, phi_i) consumed at retrieval time rather than inside the reranker prompt, and a single-pass reflective update realising in-context profile evolution without any gradient step. On four public benchmarks (Yelp, Books, MovieTV, Goodreads), HARP lifts Hit@1 over the strongest published baseline by 6.9%-24.0%, with family-wise Holm-corrected p < 1e-9 on every dataset. Ablations and a per-case attribution study show that the operators contribute complementary gains, with the retrieval-side gate as the main source; we also report the trade-off introduced by the reflective pass. We formulate agentic communication and provide a training-free instantiation. Code is available at https://github.com/Jove-Zhen/HARP.