Language-Based User Profiles for Recommendation

arXiv:2402.15623v1 Announce Type: cross
Abstract: Most conventional recommendation methods (e.g., matrix factorization) represent user profiles as high-dimensional vectors. Unfortunately, these vectors lack interpretability and steerability, and often perform poorly in cold-start settings. To address these shortcomings, we explore the use of user profiles that are represented as human-readable text. We propose the Language-based Factorization Model (LFM), which is essentially an encoder/decoder model where both the encoder and the decoder are large language models (LLMs). The encoder LLM generates a compact …

abstract arxiv cs.cl cs.hc cs.ir cs.lg explore factorization human interpretability language matrix profiles recommendation text type vectors