Technology

Meet GLiNER: A generalist AI model for named entity recognition (NER) using a bidirectional transformer

A key element of Natural Language Processing (NLP) applications is Named Entity Recognition (NER), which recognizes and classifies named entities such as names of people, places, data, and organizations in text. While certain entity types limit the effectiveness of traditional NER models, they also limit their adaptability to new or diverse data sets.

On the other hand, ChatGPT and other Large Language Models (LLMs) provide greater flexibility in entity recognition by allowing the extraction of arbitrary entities from plain text statements. However, these models are less useful in resource-limited situations because they are often large and have high computational costs, especially when accessed via APIs.

In recent research, a compact NER model called GLiNER was developed to address these issues. GLiNER processes text in both forward and reverse directions simultaneously because it uses a bidirectional transformer encoder. Compared to LLMs like ChatGPT, which use a sequential token generation approach, this bidirectional processing has the advantage of being more efficient while allowing entity extraction.

The team said that instead of large autoregressive models, they used smaller bidirectional language models (BiLM) such as BERT or deBERTa. Instead of viewing Open NER as a generation task, this approach defines it as a task of matching entity type embeddings with text span representations in latent space. This method solves scalability issues with autoregressive models and enables bidirectional context processing for richer representations.

After extensive testing, GLiNER performed well in a number of NER benchmarks, particularly excelling in zero-shot evaluations. In zero-shot scenarios, the model’s generalization and adaptability to a variety of datasets was demonstrated by evaluating it on entity types for which it was not explicitly trained.

In these experiments, GLiNER consistently outperformed both ChatGPT and fine-tuned LLMs, demonstrating its effectiveness in real-world NER applications. The model outperformed ChatGPT in eight out of ten untrained languages, demonstrating its robustness to languages ​​not encountered during training. This shows the effectiveness and versatility of this method in practical NER applications.

In summary, GLiNER provides a compressed and effective solution that strikes a balance between flexibility, performance and resource efficiency, making it a promising approach to NER. Its exceptional zero-shot performance on multiple NER benchmarks is due to its bidirectional transformer architecture, which enables parallel entity extraction, thus improving speed and accuracy compared to typical LLMs. This study emphasizes the importance of building tailored models for specific NLP tasks to meet the needs of resource-constrained environments while maintaining good performance.


Visit the Paper. All credit for this research goes to the researchers of this project. Also don’t forget to follow us Twitter. Join our… Telegram channel, Discord channelAnd LinkedIn Grupp.

If you like our work, you will love ours Newsletter..

Don’t forget to join our 41k+ ML SubReddit


Tanya Malhotra is a final year student at the University of Petroleum & Energy Studies, Dehradun, studying BTech in Computer Science with a specialization in Artificial Intelligence and Machine Learning.
She is a data science enthusiast with good analytical and critical thinking and a keen interest in learning new skills, leading groups and managing work in an organized manner.




Source link