EvoSkillRec automates recommender architecture discovery using LLMs and skill genomes

Researchers have developed EvoSkillRec, a novel framework designed to automate the evolution and discovery of recommender system architectures. This system decomposes recommenders into executable skills, represented as typed skill genomes, and utilizes a coupled evolution space. It combines a constrained skill space for mutation, recombination, and reuse of validated skills with an open-ended code space where LLM planners invent new modules. EvoSkillRec has demonstrated consistent effectiveness across various tasks, including CTR prediction and multi-task learning, optimizing for both predictive quality and computational efficiency. AI

IMPACT This framework could accelerate the development of more efficient and specialized recommender systems by automating architectural discovery.

RANK_REASON The cluster describes a novel research paper detailing a new framework for recommender systems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• arXiv
• click-through rate
• EvoSkillRec
• Hugging Face
• LLM

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →