A Unified MRC Framework for Named Entity Recognition

Li, Xiaoya; Feng, Jingrong; Meng, Yuxian; Han, Qinghong; Wu, Fei; Li, Jiwei

doi:10.48550/arxiv.1910.11476

Cited by 47 publications

(50 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, it is challenging to directly adapt MRC framework for NER task especially in the low-resource setting. First, for each entity class, the model needs to answer its associated natural language question and repeat this procedure until all the questions are answered (Li et al, 2019). Thus, such a method is not scalable when the number of entity classes increases and further exacerbates the imbalanced class issue compared to conventional NER framework.…”

Section: Span Detectionmentioning

confidence: 99%

“…The most state-of-the-art model for zero-shot NER is MRC-NER (Li et al, 2019) (3) Description understanding. Description understanding is a crucial step for the success of zero-shot NER.…”

Section: Zero-shot Nermentioning

confidence: 99%

“…These limitations prohibit the use of this method for NER task. MRC-NER (Li et al, 2019) formulates NER task as a machine reading comprehension task and enables zero-shot NER. However, the inference of MRC-NER for the single sentence needs to be conducted multiple times to collect results corresponding to all the entity types of interest, incurring expensive inference cost.…”

Section: Related Workmentioning

confidence: 99%

“…Moreover, to handle the challenging that several fine-grained classes are included in one entity class, we develop an entity class attention module to focus on the most relevant information in the given entity class description that corresponds to the extracted span. Compared with direct adaption of MRC framework into NER (Li et al, 2019), the proposed method can bring more than 54%, 30% and 26% improvement in average under few-shot learning, domain transfer and zeroshot learning settings respectively. Figure 1 shows an overview of the proposed framework: the span detection stage aims to identify the span of text, and entity class inference is responsible for categorizing extracted spans based on natural language description of pre-defined entity classes.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Learning from Language Description: Low-shot Named Entity Recognition via Decomposed Framework

Wang

Chu

Zhang

et al. 2021

Preprint

View full text Add to dashboard Cite

In this work, we study the problem of named entity recognition (NER) in a low resource scenario, focusing on few-shot and zero-shot settings. Built upon large-scale pre-trained language models, we propose a novel NER framework, namely SpanNER, which learns from natural language supervision and enables the identification of never-seen entity classes without using in-domain labeled data. We perform extensive experiments on 5 benchmark datasets and evaluate the proposed method in the few-shot learning, domain transfer and zero-shot learning settings. The experimental results show that the proposed method can bring 10%, 23% and 26% improvements in average over the best baselines in few-shot learning, domain transfer and zero-shot learning settings respectively.

show abstract

Section: Span Detectionmentioning

confidence: 99%

“…The most state-of-the-art model for zero-shot NER is MRC-NER (Li et al, 2019) (3) Description understanding. Description understanding is a crucial step for the success of zero-shot NER.…”

Section: Zero-shot Nermentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning from Language Description: Low-shot Named Entity Recognition via Decomposed Framework

Wang

Chu

Zhang

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The context of the review plays a significant role in deciding whether to suggest the entity or not. Considering this, the authors chose to use the Machine Reading Comprehension-based Named Entity Recognition (MRC-NER) [12] approach. This approach reformulates a standard NER task into QnA task.…”

Section: Machine Reading Comprehension -Named Entity Recognition (Mrc...mentioning

confidence: 99%

An Intelligent Recommendation-cum-Reminder System

Saxena¹,

Chaudhary²,

Maurya³

et al. 2021

Preprint

View full text Add to dashboard Cite

Intelligent recommendation and reminder systems are the need of the fast-pacing life. Current intelligent systems such as Siri, Google Assistant, Microsoft Cortona, etc., have limited capability. For example, if you want to wake up at 6 am because you have an upcoming trip, you have to set the alarm manually. Besides, these systems do not recommend or remind what else to carry, such as carrying an umbrella during a likely rain. The present work proposes a system that takes an email as input and returns a recommendation-cumreminder list. As a first step, we parse the emails, recognize the entities using named entity recognition (NER). In the second step, information retrieval over the web is done to identify nearby places, climatic conditions, etc. Imperative sentences from the reviews of all places are extracted and passed to the object extraction module. The main challenge lies in extracting the objects (items) of interest from the review. To solve it, a modified Machine Reading Comprehension-NER (MRC-NER) model is trained to tag objects of interest by formulating annotation rules as a query. The objects so found are recommended to the user one day in advance. The final reminder list of objects is pruned by our proposed model for tracking objects kept during the "packing activity. " Eventually, when the user leaves for the event/trip, an alert is sent containing the reminding list items. Our approach achieves superior performance compared to several baselines by as much as 30% on recall and 10% on precision.

show abstract