Large Language Models aren't all that you need
Abstract
Two approaches, including a Conditional Random Fields model and a fine-tuned Large Language Model with novel techniques, are evaluated for the MultiCoNER II task, showing significant improvements in NER performance.
This paper describes the architecture and systems built towards solving the SemEval 2023 Task 2: MultiCoNER II (Multilingual Complex Named Entity Recognition) [1]. We evaluate two approaches (a) a traditional Conditional Random Fields model and (b) a Large Language Model (LLM) fine-tuned with a customized head and compare the two approaches. The novel ideas explored are: 1) Decaying auxiliary loss (with residual) - where we train the model on an auxiliary task of Coarse-Grained NER and include this task as a part of the loss function 2) Triplet token blending - where we explore ways of blending the embeddings of neighboring tokens in the final NER layer prior to prediction 3) Task-optimal heads - where we explore a variety of custom heads and learning rates for the final layer of the LLM. We also explore multiple LLMs including GPT-3 and experiment with a variety of dropout and other hyperparameter settings before arriving at our final model which achieves micro & macro f1 of 0.85/0.84 (on dev) and 0.67/0.61 on the test data . We show that while pre-trained LLMs, by themselves, bring about a large improvement in scores as compared to traditional models, we also demonstrate that tangible improvements to the Macro-F1 score can be made by augmenting the LLM with additional feature/loss/model engineering techniques described above.
Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper