Summary of Evaluating Large Language Models For Structured Science Summarization in the Open Research Knowledge Graph, by Vladyslav Nechakhin et al.
Evaluating Large Language Models for Structured Science Summarization in the Open Research Knowledge Graph
by Vladyslav Nechakhin, Jennifer D’Souza, Steffen Eger
First submitted to arxiv on: 3 May 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); Information Theory (cs.IT)
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 The paper proposes using Large Language Models (LLMs) like GPT-3.5, Llama 2, and Mistral to automatically suggest properties for structured science summaries, which enhances research findability. Current methods involve manual curation by domain experts, but this is labor-intensive and inconsistent. The study performs a comprehensive comparative analysis between manually curated properties and those generated by the LLMs through four unique perspectives: semantic alignment, fine-grained mapping accuracy, SciNCL embeddings-based cosine similarity, and expert surveys comparing manual annotations with LLM outputs. The evaluations occur within a multidisciplinary science setting. Overall, LLMs show potential as recommendation systems for structuring science, but further finetuning is recommended to improve their alignment with scientific tasks and mimicry of human expertise. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper talks about how we can make it easier to find research papers by using special computer models called Large Language Models. Right now, people have to manually add information to describe what a research paper says, which is time-consuming and not very consistent. The researchers wanted to see if these computer models could help do this job instead. They compared the suggestions from these models with how human experts would describe the same papers. They looked at four different ways to measure how good the computer models were: matching up words, mapping specific details correctly, comparing similarities between words, and asking experts what they thought of the computer-generated descriptions. The results show that these computer models have some potential for helping us find research papers, but we need to make them better first. |
Keywords
» Artificial intelligence » Alignment » Cosine similarity » Gpt » Llama
