An app combines artificial intelligence to help users look up words, recite words and test vocabulary.
The mainstream mode of vocabulary learning software currently on the market is as follows: the software presents a word along with multiple definitions, and the user needs to choose the correct meaning corresponding to the word. Essentially, this trains users to translate words into their corresponding Chinese meanings when encountered. However, after reviewing relevant materials, we found that this is not an optimal way to learn English.
Learning vocabulary should involve understanding the connotations of words, rather than mechanically translating words into definitions. For example, consider the word "brand" with the definitions: n. a type of product made by a particular company and sold under a particular name; v. to stamp; to make memorable. In fact, "brand" refers to the action of "stamping," and when we say something is stamped in our minds, it actually means to remember. When we talk about putting a mark on something, this mark can be understood as a brand.
Before initiating this project, we conducted thorough market research and identified its feasibility and potential demand. Comments in blog posts discussing how to learn English, for example, revealed dissatisfaction with current learning methods and a desire for more suitable approaches among users.
Against this backdrop, we propose a new learning model that combines large-scale natural language models with complex neural networks. Users no longer need to memorize words and their multiple definitions forcibly but can learn the essence of words in an easy and efficient manner. After long-term learning using this model, users can overcome the cumbersome process of translating words into Chinese first and then piecing them together into sentences.
PRs accepted.