Summary of Urbankgent: a Unified Large Language Model Agent Framework For Urban Knowledge Graph Construction, by Yansong Ning et al.
UrbanKGent: A Unified Large Language Model Agent Framework for Urban Knowledge Graph Construction
by Yansong Ning, Hao Liu
First submitted to arxiv on: 10 Feb 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents UrbanKGent, a unified large language model agent framework for constructing urban knowledge graphs (UrbanKGC). The authors address the limitations of manual effort in traditional UrbanKGC methods by developing a knowledgeable instruction set for relational triplet extraction and knowledge graph completion. They also propose a tool-augmented iterative trajectory refinement module to refine trajectories distilled from GPT-4. Through hybrid instruction fine-tuning with augmented trajectories on Llama 2 and Llama 3 family, the authors obtain UrbanKGC agent families consisting of UrbanKGent-7/8/13B versions. The paper demonstrates that UrbanKGent family can significantly outperform baselines in UrbanKGC tasks and surpass state-of-the-art LLM, GPT-4, with lower cost. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Urban knowledge graphs help gather important information from different urban data sources for various city-related projects. However, building these graphs still requires a lot of manual effort. This paper introduces UrbanKGent, a new way to construct these graphs using large language models. The authors developed special instructions to tell the model what to do and how to refine its ideas. They tested their approach on two real-world datasets and compared it to other methods. The results show that UrbanKGent is better than previous approaches at constructing urban knowledge graphs, especially when it comes to creating more connections between pieces of information. |
Keywords
» Artificial intelligence » Fine tuning » Gpt » Knowledge graph » Large language model » Llama
