Titlebook: Chinese Computational Linguistics; 18th China National Maosong Sun,Xuanjing Huang,Yang Liu Conference proceedings 2019 Springer Nature Swi

Blazon

Sharing Pre-trained BERT Decoder for a Hybrid Summarizationelected sentence by an abstractive decoder. Moreover, we apply the BERT pre-trained model as document encoder, sharing the context representations to both decoders. Experiments on the CNN/DailyMail dataset show that the proposed framework outperforms both state-of-the-art extractive and abstractive models.

CRP743

Conference proceedings 2019text classification and summarization, knowledge graph and information extraction, machine translation and multilingual information processing, minority language processing, language resource and evaluation, social computing and sentiment analysis, NLP applications..

gnarled

嬉耍

Testing the Reasoning Power for NLI Models with Annotated Multi-perspective Entailment Datasetsed) models have achieved prominent success. However, rare models are interpretable. In this paper, we propose a Multi-perspective Entailment Category Labeling System (METALs). It consists of three categories, ten sub-categories. We manually annotate 3,368 entailment items. The annotated data is use

mortuary

Enhancing Chinese Word Embeddings from Relevant Derivative Meanings of Main-Components in Characters basic unit, or directly use the internal structure of words. However, these models still neglect the rich relevant derivative meanings in the internal structure of Chinese characters. Based on our observations, the relevant derivative meanings of the main-components in Chinese characters are very h

构成

Association Relationship Analyses of Stylistic Syntactic Structuresationships of linguistic features, such as collocation of morphemes, words, or phrases. Although they have drawn many useful conclusions, some summarized linguistic rules lack physical verification of large-scale data. Due to the development of machine learning theories, we are now able to use compu

打包

Ergots

SPER

BB-KBQA: BERT-Based Knowledge Base Question Answering real-world systems. Most existing methods are template-based or training BiLSTMs or CNNs on the task-specific dataset. However, the hand-crafted templates are time-consuming to design as well as highly formalist without generalization ability. At the same time, BiLSTMs and CNNs require large-scale

cauda-equina

Reconstructed Option Rereading Network for Opinion Questions Reading Comprehension question referring to a related passage. Previous work focuses on factoid-based questions but ignore opinion-based questions. Options of opinion-based questions are usually sentiment phrases, such as “Good” or “Bad”. It causes that previous work fail to model the interactive information among passa