Tsinghua University recently opened an open source online dictionary on GitHub called Want Wrong, which claims to be the first open source online reverse dictionary that supports Chinese and cross-language queries.What is a reverse dictionary? Whereas a normal dictionary tells you the definition of a word, a reverse dictionary does the opposite and tells you which words match the meaning of the description you typed. The following is a screenshot of wanciwang’s online reverse dictionary, which demonstrates an example of reverse lookup. Enter “mountain is very high”, and the system will return a series of words that the model thinks express the meaning of “mountain is very high”, such as “high”, “lofty”, etc.

What can a reverse dictionary be used for?

  • Get rid of the ‘tip of the tongue’ problem — forgetting words temporarily
  • Help language learners learn and consolidate vocabulary
  • Improved quality of life for patients with selective aphasia, who can recognize and describe an object but cannot recall its name

Simply put, is now the young people will have word poor, when you word poor, this tool can be used, such as when you fluttery and indecisive, do not know which word to use, you can put it into, will get indecisive, indecisive, flustered and so on 100 words. And the part of speech and rhyme can be customized.In addition, it also supports Chinese, English, Chinese-English, English-Chinese and other different conversion methods.The overall system architecture is also simple:The core model of Wanciwang is the multi-channel reverse dictionary model proposed in a previous paper published in AAAI-20 by natural Language Processing Laboratory (THUNLP) of Tsinghua University Department of Computer Science: Multi-channel Reverse Dictionary Model [paper] [code], whose Model architecture is as follows: About Natural Language Processing and Social Human Computing Laboratory of Tsinghua University

The Laboratory of Natural Language Processing and Social Human Computing (THUNLP) of the Department of Computer Science of Tsinghua University was founded in the late 1970s. It was first engaged in the research of Chinese information processing under the leadership of Professor Huang Changning. It is the earliest and most influential research institution in the research of natural language processing in China. At the same time, it is also the affiliated unit of computational Linguistics Committee of Chinese Language Information Society of China (the national first-class society). The academic leader of the laboratory is Professor Sun Maosong, and the faculty of the laboratory also includes Professor Liu Yang and Associate Professor Liu Zhiyuan. In recent years, with the support of 973, 863, national Natural Science Foundation of China and other projects, the Laboratory conducts systematic and in-depth research work on the frontier basic topics of natural language processing with Chinese as the core, covering the core issues of computational linguistics, social computing and human computing. The teachers and students of the laboratory have published many high-level academic papers in IJCAI, AAAI, ACL, EMNLP and other international top conferences and journals, and have a long-term good cooperative relationship with CMU, NUS, Google and other international universities and enterprises. Most of the outstanding graduates work in tsinghua University, Google, Baidu, Ali, Microsoft and other famous universities and enterprises.

WantWords is developed and maintained by THUNLP. The project is supervised by Professor Sun Maosong and Associate Professor Liu Zhiyuan, and the development team includes Quan Fanchao, Zhang Lei and Yang Yanhui.Currently, Wantwords has been marked 1K on GitHub, with 56 branches (GitHub address:Github.com/thunlp/Want…

OpenSourceTop (ID: OpenSourceTop)