eC-Tab2Text: Aspect-Based Text Generation from e-Commerce Product Tables

2025-02-20

AI Summary

The e-commerce domain lacks fine-grained, attribute-aware table-to-text generation datasets, hindering large language models (LLMs) from producing high-quality, factually consistent, and user-intent-aligned product reviews. Method: We introduce eC-Tab2Text—the first e-commerce-specific, fine-grained table-to-text dataset enabling aspect-based (attribute-level) controllable generation. We propose an LLM-based supervised fine-tuning framework jointly optimized using general table-to-text evaluation metrics and domain-specific criteria: correctness, faithfulness, and fluency. Contribution/Results: Experiments demonstrate that our approach significantly outperforms general-purpose baselines in both automated and human evaluations. It achieves marked improvements in contextual accuracy and factual consistency, thereby addressing critical data and task gaps in e-commerce text generation. eC-Tab2Text establishes a new benchmark for attribute-aware, intent-adaptive review generation and supports rigorous evaluation of domain-specific generation capabilities.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large Language Models (LLMs) have demonstrated exceptional versatility across diverse domains, yet their application in e-commerce remains underexplored due to a lack of domain-specific datasets. To address this gap, we introduce eC-Tab2Text, a novel dataset designed to capture the intricacies of e-commerce, including detailed product attributes and user-specific queries. Leveraging eC-Tab2Text, we focus on text generation from product tables, enabling LLMs to produce high-quality, attribute-specific product reviews from structured tabular data. Fine-tuned models were rigorously evaluated using standard Table2Text metrics, alongside correctness, faithfulness, and fluency assessments. Our results demonstrate substantial improvements in generating contextually accurate reviews, highlighting the transformative potential of tailored datasets and fine-tuning methodologies in optimizing e-commerce workflows. This work highlights the potential of LLMs in e-commerce workflows and the essential role of domain-specific datasets in tailoring them to industry-specific challenges.

Problem

• Generating product reviews from tables
• Enhancing e-commerce with LLMs
• Creating domain-specific datasets for AI

Innovation

• Domain-specific dataset creation
• Fine-tuned Large Language Models
• Structured data to text conversion