SALT: Sales Autocompletion Linked Business Tables Dataset

2025-01-06

AI Summary

Foundational models (e.g., Transformers) struggle to effectively model multi-table relational structures in enterprise ERP systems—particularly those linked via foreign keys—limiting their practical applicability in business contexts. To address this, we introduce SALT, the first large-scale, foreign-key-aligned, and intent-annotated multi-table commercial dataset derived from real-world ERP logs, specifically designed for sales auto-completion. We propose a systematic methodology encompassing foreign-key relationship modeling, table-structure standardization, and fine-grained sales intent annotation to ensure high data quality and reproducibility. SALT bridges a critical gap in multi-table representation learning for business data, significantly improving model generalization on cross-table query understanding and sales field prediction tasks. It establishes a new benchmark for representation learning tailored to enterprise-grade structured data.

Technology Category

• Machine Learning: Statistical Relational/Logic Learning
• Data Mining & Knowledge Management: Intelligent Query Processing
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Foundation models, particularly those that incorporate Transformer architectures, have demonstrated exceptional performance in domains such as natural language processing and image processing. Adapting these models to structured data, like tables, however, introduces significant challenges. These difficulties are even more pronounced when addressing multi-table data linked via foreign key, which is prevalent in the enterprise realm and crucial for empowering business use cases. Despite its substantial impact, research focusing on such linked business tables within enterprise settings remains a significantly important yet underexplored domain. To address this, we introduce a curated dataset sourced from an Enterprise Resource Planning (ERP) system, featuring extensive linked tables. This dataset is specifically designed to support research endeavors in table representation learning. By providing access to authentic enterprise data, our goal is to potentially enhance the effectiveness and applicability of models for real-world business contexts.

Problem

• Transformer Models
• Multi-Table Data
• Business Intelligence

Innovation

• SALT Dataset
• Multi-Table Association
• Business Scenario Optimization