Researchers present a multi-agent vision-language system that helps generate written product reviews grounded in user-uploaded images and videos from e-commerce platforms. The approach is intended to make use of visual feedback showing item quality, defects, packaging, and real-world use. The work is published as an arXiv preprint.
Researchers introduce Valley3, a multimodal large language model designed for e-commerce applications across different markets. The model handles text, images, video, and audio within a single framework, aiming to combine understanding and reasoning abilities across those modalities. The paper is a revised arXiv preprint, with an updated version posted to the cs.AI category.
A newly published arXiv paper tackles the challenge of extracting product attributes from large, unstructured e-commerce catalogs. The authors present a parallel decoding approach designed to identify the attributes shoppers care about and extract them efficiently at scale.
Researchers propose RA-CoA, a training-free method for fashion image captioning that combines retrieval with a chain-of-attributes reasoning approach. The work targets e-commerce use cases, where captioning demands fine-grained visual analysis and correct domain-specific fashion terminology rather than generic scene description. It is published as an arXiv preprint in the cross-listed machine learning category.
A new arXiv paper addresses quality issues in large e-commerce catalogs, including duplicate entries and unit mismatches that frustrate shoppers. The authors introduce a positional task conditioning technique that lets a single model reason over several error types in long product listings while scaling across many product families. The approach aims to automate catalog curation more reliably for retailers with extensive inventories.
A new version of an arXiv paper (2502.01555) explores how to connect short user search queries with the correct brand entities during e-commerce product retrieval. The authors highlight the difficulty of the task, since queries average only three to four words and lack grammatical structure. The work is cross-listed in the cs.AI and cs.LG categories.
A new arXiv preprint introduces Agentic Share-of-Search, a multi-agent framework that automates how visible sellers are when AI shopping assistants mediate product discovery. The system pairs competitive visibility measurement with root-cause analysis to help sellers make decisions as LLM-driven commerce grows. It targets the emerging seller-side market created by AI assistants redirecting consumer search.