Summary of Investigating Ocr-sensitive Neurons to Improve Entity Recognition in Historical Documents, by Emanuela Boros and Maud Ehrmann
Investigating OCR-Sensitive Neurons to Improve Entity Recognition in Historical Documents
by Emanuela Boros, Maud Ehrmann
First submitted to arxiv on: 25 Sep 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 The paper investigates the impact of OCR-sensitive neurons within the Transformer architecture on named entity recognition (NER) performance. By analyzing neuron activation patterns, it identifies and neutralizes OCR-sensitive neurons to improve model performance. The study uses two large language models, Llama2 and Mistral, and demonstrates the existence of OCR-sensitive regions. The results show improvements in NER performance on historical newspapers and classical commentaries. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how a type of neuron in a special kind of AI called the Transformer affects its ability to find important words (named entity recognition) when dealing with old documents that have writing errors (OCR). They test this by looking at how the neurons react to clean and messy text, and then they try to get rid of any neurons that are sensitive to these errors. By doing this, they show that it can actually help the AI do better on these kinds of texts. |
Keywords
» Artificial intelligence » Named entity recognition » Ner » Transformer
