KnowMap: Efficient Knowledge-Driven Task Adaptation for LLMs

2025-06-24

AI Summary

Large language models (LLMs) suffer from reliance on static pretraining knowledge, inefficient adaptation to novel tasks, and catastrophic forgetting during fine-tuning. To address these challenges, this paper proposes KnowMap—a knowledge-driven dynamic adaptation framework. Its core innovation is a lightweight knowledge embedding model that dynamically integrates environmental observations and interaction experiences to construct and continuously update task-specific knowledge bases; LLMs are then rapidly adapted via knowledge injection rather than full-parameter fine-tuning. This approach eliminates the need for costly labeled data and parameter retraining, thereby significantly mitigating forgetting. Evaluated on the ScienceWorld benchmark, KnowMap improves GPT-4-Turbo’s task completion rate by 17.71%, demonstrating its efficacy and robustness in enhancing reasoning capabilities for open-world agent tasks through efficient, adaptive knowledge integration.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Multiagent Systems: Adversarial Agents

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

While Large Language Models (LLMs) possess significant capabilities in open-world agent tasks, they also face challenges in rapidly adapting to new, specialized tasks due to their reliance on static pre-trained knowledge. Traditional methods such as fine-tuning are often costly, data-intensive, and may lead to "catastrophic forgetting." Therefore, we present KnowMap, a novel approach that dynamically constructs a knowledge base from environmental and experiential data. KnowMap fine-tunes a small knowledge-embedding model to equip a larger LLM with valuable task-specific knowledge. Our experiments on the ScienceWorld benchmark demonstrate 17.71% improvement for the performance of gpt-4-turbo model. KnowMap not only provides an efficient and effective means for LLM task-adapting, but also highlights how integrating environmental and experiential knowledge can enhance LLMs' reasoning capabilities.

Problem

• Efficient adaptation of LLMs to new specialized tasks
• Overcoming static knowledge reliance in LLMs
• Reducing costly fine-tuning and catastrophic forgetting

Innovation

• Dynamic knowledge base construction from environment
• Fine-tunes small knowledge-embedding model for LLM
• Enhances reasoning with environmental experiential knowledge