NLPCC 2021 will follow the NLPCC tradition of holding several shared tasks in natural language processing and Chinese computing. This year’s shared tasks focus on both classical problems and newly emerging problems, including Argumentative Text Understanding for AI Debater, Few-shot Learning for Chinese Language Understand Evaluation with the Pre-training Language Model and Sub Event Identification.

Participants from both academia and industry are welcomed. Each group can participate in one or multiple tasks and members in each group can attend the NLPCC conference to present their techniques and results. The participants will be invited to submit papers to the main conference and the accepted papers will appear in the conference proceedings published by Springer LNCS.

The top 3 participating teams of each task will be certificated by NLPCC and CCF-NLP. If a task has multiple sub-tasks or tracks, then only the top 1 participating team of each sub-task/track will be certificated.

There are three shared tasks in this year’s NLPCC conference and the detailed description of each task can be found in the task guidelines released later. Here we only give a brief overview of each task.

◇ Task 1 - Argumentative Text Understanding for AI Debater (AIDebater)

Argument and debate are fundamental capabilities of human intelligence, essential for a wide range of human activities, and common to all human societies. With an aim of developing an autonomous debating system, we make an initial step to understand argumentative text of debating in this shared task, including three tracks, namely, supporting material identification (track 1), argument pair identification from online forum (track 2) and argument pair extraction from peer review and rebuttal (track 3). In track 1, we present the fundamental scenario of supporting material identification for a given debating topic. We then move to the understanding of dialogical argumentative text in two domains, i.e., online debating forum and scientific paper review process. We provide three datasets in this task, one for each track.

Organizer: Fudan University, Alibaba Group.

Contact: aidebater@163.com

◇ Task 2 - FewCLUE: Few-shot Learning for Chinese Language Understanding Evaluation with the Pre-trained Language Model

Pretrained language models have greatly improved the performance of various NLP tasks among many fields. With the help of many labelled data, people could just make a little bit fine-tuning of a pre-trained language model to get a great performance for downstream tasks compared with the age without a pre-trained language model. However, humanity could learn a new concept with only a few samples. Few-shot learning is the one to study how a machine could learn a new concept with only a few data.

The goal of this task is to explore the capability of a pretrained language model conditioned on a few training data for the downstream task. CLUE designed this task based on CLUE benchmark, wishes to promote more research and applications in Chinese NLP few-shot learning.

Organizer: CLUE benchmark (www.CLUEbenchmarks.com)

Contact: cluebenchmark@163.com and Junyi Li (4ljy@163.com)

◇ Task 3 - Automatic Information Extraction – Sub Event Identification (AutoIE 2)

Sub-events identification is a very fundamental problem in the field of information extraction, especially in emergency situations (e.g., terrorist attacks). Events usually evolve rapidly and therefore successive sub-events occur. How to efficiently identify the target sub-events from large volume of related data is very challenging. The goal of this task is to build an IE system that can quickly adapt to a new occurring sub-event. Specifically, given a large event-related corpus and a few labelled seed data, the task aims to build an IE system capable of identifying the target sub-events. Besides machine learning model designing, selecting data from the unlabeled corpus and annotating them is also allowed, but the size of the labelled data is fixed. How to select the best data to annotate is also an important step in this task. The task setting is very practical and thus the proposed solutions may generalize well in real world applications.

Organizer: TSinghua SIGS

Contact: Xingyu Bai (bxy20@mails.tsinghua.edu.cn) and Xuefeng Yang (yang0302@e.ntu.edu.sg)

◇ Task 1 - Argumentative Text Understanding for AI Debater (AIDebater)

Registration website: http://www.fudan-disc.com/sharedtask/AIDebater21/index.html

◇ Task 2 - FewCLUE: Few-shot Learning for Chinese Language Understanding Evaluation with the Pre-trained Language Model

Registration website: https://www.cluebenchmarks.com/NLPCC.html

◇ Task 3 - Automatic Information Extraction – Sub Event Identification (AutoIE 2)

Please fill out the Shared Task 3 Registration Form (Word File) and send it to the following registration email.

Registration Email: bxy20@mails.tsinghua.edu.cn

2021/04/10：announcement of shared tasks and call for participation;

2021/04/10：registration open;

2021/04/30：release of detailed task guidelines & training data;

2021/05/30：registration deadline;

2021/06/05：release of test data;

2021/06/15：participants’ results submission deadline;

2021/06/25：evaluation results release and call for system reports and conference paper;

2021/07/15：conference paper submission deadline (only for shared tasks);

2021/07/30：conference paper accept/reject notification;

2021/08/14：camera-ready paper submission deadline.

The evaluation papers are English only. The papers will be in the proceedings of the NLPCC-2021 conference (for English) which will be published as a volume in the Springer LNAI series (EI & ISTP indexed, English papers). Submissions should follow the LNCS formatting instructions. The maximum paper length is 12 pages, including references; The submissions must therefore be formatted in accordance with the standard Springer style sheets ([LaTeX][Microsoft Word]). Manuscripts should be submitted electronically through the submission website (https://www.softconf.com/nlpcc/eval-2021). Email submissions will not be accepted. Submissions should be in PDF format.

Xingyu Bai, TSinghua SIGS

Lidong Bing, Alibaba Group

Yunbo Cao, Tencent

Liying Cheng , Alibaba Group

Weigang Guo, TSinghua SIGS

Ruidan He , Alibaba Group

Junyi Li, CLUE benchmark

Yinzi Li, Fudan University

Qin Liu, Fudan University

Chenhui Shen, Alibaba Group

Zhongyu Wei, Fudan University

Huilin Xu, CLUE benchmark

Liang Xu, CLUE benchmark

Xuefeng Yang, TSinghua SIGS

Yujiu Yang, TSinghua SIGS

Jian Yuan, Fudan University

Jiajun Zhang, Institute of Automation Chinese Academy of Sciences