CONE: Embeddings for Complex Numerical Data Preserving Unit and Variable Semantics

TL;DR

CONE is a hybrid transformer model designed to encode complex numerical data with units and semantics, significantly improving numerical reasoning in large language models across multiple domains.

Contribution

The paper introduces a novel composite embedding algorithm for numerical data, enhancing semantic preservation and reasoning capabilities in transformer models.

Findings

Achieves 87.28% F1 on DROP dataset, outperforming SOTA by 9.37%.

Demonstrates strong numerical reasoning across diverse domains.

Provides a new approach to embedding structured numerical data.

Abstract

Large pre-trained models (LMs) and Large Language Models (LLMs) are typically effective at capturing language semantics and contextual relationships. However, these models encounter challenges in maintaining optimal performance on tasks involving numbers. Blindly treating numerical or structured data as terms is inadequate -- their semantics must be well understood and encoded by the models. In this paper, we propose CONE, a hybrid transformer encoder pre-trained model that encodes numbers, ranges, and gaussians into an embedding vector space preserving distance. We introduce a novel composite embedding construction algorithm that integrates numerical values, ranges or gaussians together with their associated units and attribute names to precisely capture their intricate semantics. We conduct extensive experimental evaluation on large-scale datasets across diverse domains (web, medical, finance, and government) that justifies CONE's strong numerical reasoning capabilities, achieving an F1 score of 87.28% on DROP, a remarkable improvement of up to 9.37% in F1 over state-of-the-art (SOTA) baselines, and outperforming major SOTA models with a significant Recall@10 gain of up to 25%.