QuanTA: Efficient High-Rank Fine-Tuning of LLMs with Quantum-Informed Tensor Adaptation

TL;DR

QuanTA introduces a quantum-inspired tensor adaptation method for fine-tuning large language models, enabling high-rank adjustments without inference overhead, and significantly improving reasoning capabilities and scalability.

Contribution

The paper presents a novel quantum-inspired tensor adaptation technique that surpasses low-rank methods, supported by theoretical proofs, and enhances fine-tuning efficiency and performance.

Findings

QuanTA outperforms traditional methods in reasoning tasks.

It requires fewer trainable parameters for comparable performance.

The approach is scalable and compatible with existing fine-tuning algorithms.

Abstract

We propose Quantum-informed Tensor Adaptation (QuanTA), a novel, easy-to-implement, fine-tuning method with no inference overhead for large-scale pre-trained language models. By leveraging quantum-inspired methods derived from quantum circuit structures, QuanTA enables efficient high-rank fine-tuning, surpassing the limitations of Low-Rank Adaptation (LoRA)--low-rank approximation may fail for complicated downstream tasks. Our approach is theoretically supported by the universality theorem and the rank representation theorem to achieve efficient high-rank adaptations. Experiments demonstrate that QuanTA significantly enhances commonsense reasoning, arithmetic reasoning, and scalability compared to traditional methods. Furthermore, QuanTA shows superior performance with fewer trainable parameters compared to other approaches and can be designed to integrate with existing fine-tuning algorithms for further improvement, providing a scalable and efficient solution for fine-tuning large language models and advancing state-of-the-art in natural language processing.