Summary of Improving Entity Recognition Using Ensembles Of Deep Learning and Fine-tuned Large Language Models: a Case Study on Adverse Event Extraction From Multiple Sources, by Yiming Li et al.
Improving Entity Recognition Using Ensembles of Deep Learning and Fine-tuned Large Language Models: A Case Study on Adverse Event Extraction from Multiple Sources
by Yiming Li, Deepthi Viswaroopan, William He, Jianfu Li, Xu Zuo, Hua Xu, Cui Tao
First submitted to arxiv on: 26 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 investigates the performance of traditional deep learning models and large language models (LLMs) in extracting adverse event (AE) information from text data related to COVID-19 vaccines. The study focuses on understanding the effectiveness of these models in extracting three types of entities: “vaccine”, “shot”, and “ae” from reports, posts, and tweets. The authors explore and fine-tune various LLMs, including GPT-2, GPT-3.5, GPT-4, and Llama-2, as well as traditional deep learning models like RNN and BioBERT. The study also evaluates the impact of ensembling these models on performance. The results show that the ensemble model achieves the highest performance in extracting AE-related information with a micro-average F1 score of 0.903. This study contributes to the advancement of biomedical natural language processing, providing valuable insights into improving AE extraction from text data for pharmacovigilance and public health surveillance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how well computers can understand information about bad reactions to COVID-19 vaccines from texts online. It compares two types of computer programs: traditional deep learning models and large language models (LLMs). These models are used to find specific words like “vaccine”, “shot”, and “bad reaction” in reports, posts, and tweets. The study finds that combining the best of both worlds – using a mix of traditional deep learning models and LLMs – gives the most accurate results. This research helps improve computer programs that analyze text data to monitor vaccine safety and public health. |
Keywords
» Artificial intelligence » Deep learning » Ensemble model » F1 score » Gpt » Llama » Natural language processing » Rnn
