Are you fed up with literature review?
Literature review is a very important task for researchers. However, for researchers who are new to the research field or whose first language is not English, literature search can sometimes be a difficult task. Most of the existing article search services use "keyword search", which means that you need to select the important words in your research by yourself, which requires in-depth knowledge of your research and sometimes language skills. I am sure that most of you know this difficulty. That's why I launched Paper Mininig, a next generation paper search service! Unlike conventional systems, you don't need to select keywords yourself. Unlike conventional systems, you do not need to select keywords yourself, but only register the Title and Abstract information of papers in your research field. You can register papers in 24 different languages, including English, Japanese, and Chinese! And best of all, Paper Mining is free.
By using artificial intelligence (AI), which has developed rapidly in recent years, we can analyze your research field and find papers on arXiv.org that contain the information you are looking for. Although I used the vague term "AI" earlier, it is actually a deep learning model that can rapidly convert sentences into numerical vectors (768-dimensional vectors). These are commonly referred to as Sentence BERT or Sentence Vector Generation Models. Here is a simple video of Sentence BERT.
And we used the SPECTER model, which can convert article information into vectors with more correct and potential meanings, among the sentence vector generation models. By generating sentence vectors based on the article information you have registered, similar sentence vectors can be found in the database and output. Here is an image movie of the system.
If you don't know how it works, there is nothing to worry about. All you need is the ability to press Ctrl + C and Ctrl + V!
What Paper Mininig can do for you.
As mentioned above, you can use the Title and Abstract registered in Copy And Paste to find similar papers. Currently, the search target is only the arXiv, but we hope to make it possible to search for papers in various journals in future versions. In addition to just searching, the application also has various other functions. For example, what I like is the email notification function. If you have a paper registered, it will send you an email notification when a new paper with a high degree of similarity is posted. Other features include a summary function and a system that allows you to talk to the AI about the contents of your paper.
Have a question for us?
If you have any questions for us, please use the contact form in the header after logging in. We are also currently developing a survey form. This survey will be used to help us upgrade Paper Mining. We would be very grateful if you could give us some feedback, no matter how trivial. Even comments such as ,The design is in poor taste! and other comments are also welcome.