Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TC-RAG:Turing-Complete RAG's Case study on Medical LLM Systems

Xinke Jiang, Yue Fang, Rihong Qiu, Haoyu Zhang, Yongxin Xu, Hao Chen, Wentao Zhang, Ruizhe Zhang, Yuchen Fang, Xu Chu, Junfeng Zhao, Yasha Wang

TL;DR

The TC-RAG is introduced through rigorous proof, a novel framework that addresses challenges of domain-specific Large Language Models by incorporating a Turing Complete System to manage state variables, thereby enabling more efficient and accurate knowledge retrieval.

Abstract

In the pursuit of enhancing domain-specific Large Language Models (LLMs), Retrieval-Augmented Generation (RAG) emerges as a promising solution to mitigate issues such as hallucinations, outdated knowledge, and limited expertise in highly specialized queries. However, existing approaches to RAG fall short by neglecting system state variables, which are crucial for ensuring adaptive control, retrieval halting, and system convergence. In this paper, we introduce the TC-RAG through rigorous proof, a novel framework that addresses these challenges by incorporating a Turing Complete System to manage state variables, thereby enabling more efficient and accurate knowledge retrieval. By leveraging a memory stack system with adaptive retrieval, reasoning, and planning capabilities, TC-RAG not only ensures the controlled halting of retrieval processes but also mitigates the accumulation of erroneous knowledge via Push and Pop actions. In the case study of the medical domain, our extensive experiments on real-world healthcare datasets demonstrate the superiority of TC-RAG over existing methods in accuracy by over 7.20\%. Our dataset and code have been available at https://https://github.com/Artessay/SAMA.git.

TC-RAG:Turing-Complete RAG's Case study on Medical LLM Systems

TL;DR

The TC-RAG is introduced through rigorous proof, a novel framework that addresses challenges of domain-specific Large Language Models by incorporating a Turing Complete System to manage state variables, thereby enabling more efficient and accurate knowledge retrieval.

Abstract

In the pursuit of enhancing domain-specific Large Language Models (LLMs), Retrieval-Augmented Generation (RAG) emerges as a promising solution to mitigate issues such as hallucinations, outdated knowledge, and limited expertise in highly specialized queries. However, existing approaches to RAG fall short by neglecting system state variables, which are crucial for ensuring adaptive control, retrieval halting, and system convergence. In this paper, we introduce the TC-RAG through rigorous proof, a novel framework that addresses these challenges by incorporating a Turing Complete System to manage state variables, thereby enabling more efficient and accurate knowledge retrieval. By leveraging a memory stack system with adaptive retrieval, reasoning, and planning capabilities, TC-RAG not only ensures the controlled halting of retrieval processes but also mitigates the accumulation of erroneous knowledge via Push and Pop actions. In the case study of the medical domain, our extensive experiments on real-world healthcare datasets demonstrate the superiority of TC-RAG over existing methods in accuracy by over 7.20\%. Our dataset and code have been available at https://https://github.com/Artessay/SAMA.git.
Paper Structure (56 sections, 4 theorems, 10 equations, 3 figures, 4 tables, 1 algorithm)

This paper contains 56 sections, 4 theorems, 10 equations, 3 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Reasoning and Planning Capabilities
  4. Prelinmary
  5. Definition 1. (Stack of Memory).
  6. Definition 2. (Meta-Actions).
  7. Task Definition.
  8. TURING-COMPLETE
  9. Step 1: Construction of the Turing System.
  10. Step 2: Configuration Mapping
  11. Step 3: Simulation of Computation Steps
  12. Step 4: Preservation of Acceptance and Rejection
  13. Memory Stack and State Monitor
  14. Stack Memory and Composed Actions
  15. Stack Status and Memory Management
  16. ...and 41 more sections

Key Result

Lemma 1

If $c_1\vdash_T c_2$ in $T$, then $h(c_1)\vdash_{\textsc{Tc}}^* h(c_2)$ in Tc, where $*$ denotes one or multiple steps of derivation, $\vdash$ is the action shift operator.

Figures (3)

  • Figure 1: Overall framework of Tc--Rag.
  • Figure 2: Case Study of ReACT-based and Tc--Rag.
  • Figure 3: Hyper-parameter study with the different threshold $\sigma$ for $cppl$ (Left) and $uct$ (Right).

Theorems & Definitions (8)

  • Lemma 1
  • Proof 4.1
  • Lemma 2
  • Proof 4.2
  • Lemma 3
  • Proof 8.1
  • Lemma 4
  • Proof 8.2