Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Knowledgeable Preference Alignment for LLMs in Domain-specific Question Answering

Yichi Zhang, Zhuo Chen, Yin Fang, Yanxi Lu, Fangming Li, Wen Zhang, Huajun Chen

TL;DR

This work tackles domain-specific QA by aligning LLM preferences with human expectations through Knowledgeable Preference Alignment (KnowPAT). It introduces two knowledge-aware preference sets (style and knowledge) and a novel adaptive alignment loss to jointly optimize user-friendly responses and prudent use of retrieved KB content. Evaluations on CPQA and RJUA-QA show KnowPAT outperforms 15 baselines across traditional and model-based metrics, with positive human judgments and robust knowledge retention. The approach offers a practical path to deploying KB-enabled LLMs in industry, balancing discourse quality and factual grounding while mitigating irrelevant or harmful outputs.

Abstract

Deploying large language models (LLMs) to real scenarios for domain-specific question answering (QA) is a key thrust for LLM applications, which poses numerous challenges, especially in ensuring that responses are both accommodating to user requirements and appropriately leveraging domain-specific knowledge bases. They are the two major difficulties for LLM application as vanilla fine-tuning falls short of addressing. Combining these requirements, we conceive of them as the requirement for the model's preference to be harmoniously aligned with humans'. Thus, we introduce Knowledgeable Preference AlignmenT (KnowPAT), which constructs two kinds of preference sets to tackle the two issues. Besides, we design a new alignment objective to align the LLM preference with different human preferences uniformly, aiming to optimize LLM performance in real-world, domain-specific QA settings. Adequate experiments and comprehensive comparisons with 15 baseline methods illustrate that our KnowPAT is a superior pipeline for real-scenario domain-specific QA with LLMs.

Knowledgeable Preference Alignment for LLMs in Domain-specific Question Answering

TL;DR

This work tackles domain-specific QA by aligning LLM preferences with human expectations through Knowledgeable Preference Alignment (KnowPAT). It introduces two knowledge-aware preference sets (style and knowledge) and a novel adaptive alignment loss to jointly optimize user-friendly responses and prudent use of retrieved KB content. Evaluations on CPQA and RJUA-QA show KnowPAT outperforms 15 baselines across traditional and model-based metrics, with positive human judgments and robust knowledge retention. The approach offers a practical path to deploying KB-enabled LLMs in industry, balancing discourse quality and factual grounding while mitigating irrelevant or harmful outputs.

Abstract

Deploying large language models (LLMs) to real scenarios for domain-specific question answering (QA) is a key thrust for LLM applications, which poses numerous challenges, especially in ensuring that responses are both accommodating to user requirements and appropriately leveraging domain-specific knowledge bases. They are the two major difficulties for LLM application as vanilla fine-tuning falls short of addressing. Combining these requirements, we conceive of them as the requirement for the model's preference to be harmoniously aligned with humans'. Thus, we introduce Knowledgeable Preference AlignmenT (KnowPAT), which constructs two kinds of preference sets to tackle the two issues. Besides, we design a new alignment objective to align the LLM preference with different human preferences uniformly, aiming to optimize LLM performance in real-world, domain-specific QA settings. Adequate experiments and comprehensive comparisons with 15 baseline methods illustrate that our KnowPAT is a superior pipeline for real-scenario domain-specific QA with LLMs.
Paper Structure (26 sections, 7 equations, 4 figures, 4 tables)

This paper contains 26 sections, 7 equations, 4 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Problem Setting
  3. Our KnowPAT Pipeline
  4. Unsupervised Knowledge Retrieve
  5. Knowledgeable Preference Set Construction
  6. Fine-tuning and Preference Alignment
  7. Experiments and Analysis
  8. Experiment Settings
  9. Dataset Information
  10. Baseline Methods
  11. Evaluation Metrics
  12. Implementation Details
  13. Main Results (Q1)
  14. Ablation Study (Q2)
  15. Case Study (Q3)
  16. ...and 11 more sections

Figures (4)

  • Figure 1: A simple case of intelligent service for cloud products. Such a simple example is meant to illustrate the importance of selective use of retrieved knowledge as MAC is a terminology in computer networking rather than a kind of computer or lipstick in the user context.
  • Figure 2: The overall architecture of KnowPAT. We design three important modules in our framework: unsupervised knowledge retrieve, knowledgeable preference set construction, and knowledgeable preference alignment.
  • Figure 3: The human evaluation results. For each competition, we randomly select 100 questions and compare the generated results of the two methods.
  • Figure 4: The commonsense ability on five domains.