Recent instruction-tuned large language models (LLMs) have demonstrated remarkable performance on various downstream tasks, including named entity recognition (NER). However, previous approaches often generate incorrect predictions, particularly regarding entity boundaries and types. Many of these errors can be corrected to match the ground truth by revising the entity boundaries and/or types. In this paper, we propose a self-correcting instruction tuning approach that simultaneously learns to perform NER and correct errors through natural language instructions. Self-correcting instruction tuning requires only a standard annotated NER dataset. Supervision for self-correction can be automatically generated from error patterns observed in LLMs fine-tuned solely on NER tasks. We conducted extensive experiments on eight NER datasets with two LLMs to validate the effectiveness of the proposed approach. The results demonstrate that the proposed approach enhances NER performance by effectively correcting prediction errors and substantially reducing false positives. We further analyze the self-correction behavior to better understand how the models improve performance.

Much research has reported the training data of summarization models are noisy; summaries often do not reflect what is written in the source texts. We propose an effective method of curriculum learning to train summarization models from such noisy data. Curriculum learning is used to train sequence-to-sequence models with noisy data. In translation tasks, previous research quantified noise of the training data using two models trained with noisy and clean corpora. Because such corpora do not exist in summarization fields, we propose a model that can quantify noise from a single noisy corpus. We conduct experiments on three summarization models; one pretrained model and two non-pretrained models, and verify our method improves the performance. Furthermore, we analyze how different curricula affect the performance of pretrained and non-pretrained summarization models. Our result on human evaluation also shows our method improves the performance of summarization models.

This paper describes our model for the reading comprehension task of the MRQA shared task. We propose CLER, which stands for Cross-task Learning with Expert Representation for the generalization of reading and understanding. To generalize its capabilities, the proposed model is composed of three key ideas: multi-task learning, mixture of experts, and ensemble. In-domain datasets are used to train and validate our model, and other out-of-domain datasets are used to validate the generalization of our model’s performances. In a submission run result, the proposed model achieved an average F1 score of 66.1 % in the out-of-domain setting, which is a 4.3 percentage point improvement over the official BERT baseline model.

We describe here our system and results on the FEVER shared task. We prepared a pipeline system which composes of a document selection, a sentence retrieval, and a recognizing textual entailment (RTE) components. A simple entity linking approach with text match is used as the document selection component, this component identifies relevant documents for a given claim by using mentioned entities as clues. The sentence retrieval component selects relevant sentences as candidate evidence from the documents based on TF-IDF. Finally, the RTE component selects evidence sentences by ranking the sentences and classifies the claim simultaneously. The experimental results show that our system achieved the FEVER score of 0.4016 and outperformed the official baseline system.