KALE: Enhancing Knowledge Manipulation in Large Language Models via Knowledge-aware Learning

2026-01-12

AI Summary

This work addresses the “knowing-but-not-using” problem in large language models (LLMs), where models fail to effectively leverage their internal knowledge to generate correct answers. To mitigate this, the authors propose KALE, a novel framework that integrates knowledge graphs to extract multi-hop reasoning paths and synthesize high-quality reasoning rationales. KALE further introduces a knowledge-aware fine-tuning strategy based on KL divergence to align the model’s output distributions with and without explicit reasoning rationales, thereby encouraging the internalization of interpretable reasoning processes. Extensive experiments across eight mainstream benchmarks and six prominent LLMs demonstrate that KALE improves accuracy by an average of 4.18%, with gains as high as 11.72%, significantly enhancing the models’ ability to manipulate and apply stored knowledge.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Knowledge Representation and Reasoning: Knowledge Acquisition

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Despite the impressive performance of large language models (LLMs) pretrained on vast knowledge corpora, advancing their knowledge manipulation-the ability to effectively recall, reason, and transfer relevant knowledge-remains challenging. Existing methods mainly leverage Supervised Fine-Tuning (SFT) on labeled datasets to enhance LLMs'knowledge manipulation ability. However, we observe that SFT models still exhibit the known&incorrect phenomenon, where they explicitly possess relevant knowledge for a given question but fail to leverage it for correct answers. To address this challenge, we propose KALE (Knowledge-Aware LEarning)-a post-training framework that leverages knowledge graphs (KGs) to generate high-quality rationales and enhance LLMs'knowledge manipulation ability. Specifically, KALE first introduces a Knowledge-Induced (KI) data synthesis method that efficiently extracts multi-hop reasoning paths from KGs to generate high-quality rationales for question-answer pairs. Then, KALE employs a Knowledge-Aware (KA) fine-tuning paradigm that enhances knowledge manipulation by internalizing rationale-guided reasoning through minimizing the KL divergence between predictions with and without rationales. Extensive experiments on eight popular benchmarks across six different LLMs demonstrate the effectiveness of KALE, achieving accuracy improvements of up to 11.72% and an average of 4.18%.

Problem

• knowledge manipulation
• large language models
• known&incorrect phenomenon
• reasoning
• knowledge recall

Innovation

• Knowledge-Aware Learning
• Knowledge Graph
• Rationale Generation
• Multi-hop Reasoning
• KL Divergence Fine-tuning