Notable Publisher Writes A Textbook Using Machine Learning

Springer Nature scientific publisher has recently published the first entirely-machine-learning-algorithm-written book.

The book called Lithium-Ion Batteries: A Machine-Generated Summary of Current Research has 247 research pages about Li-ion batteries as well as summaries written by AI. The content of the book is about how AI is able to help researchers be aware of the newly-released study, and algorithms can review vast literature bodies and choose the most significant details by themselves.

Beta Writer, which is an algorithm created by German Goethe University’s scientists at the Applied Computational Linguistics lab, is the book's author.

Beta Writer had to read an enormous amount of lithium-ion- related articles released in academic journals of Springer and selected the most applicable information in order to write this book, reported by Gizmodo.

Additionally, Springer Nature states that it has been planning to take advantage of Beta Writer to deal with other research fields in the coming time, using the basic tech but aiming at different aspects of science. 

Sprinter wrote on the page of the book:

Depends on the requested subject matter, the algorithm will first take advantage of machine learning to research thousands of publications to make sure that only relevant ones are chosen for the book. Then comes the process of parsing, condensing, organizing of those peer-reviewed, pre-approved publications from the database of Springer Nature into logical chapters and sections. In each chapter, the algorithm automatically generates summaries. They can incorporate quoted passages which are linked to original research papers through hyperlinks for anybody who wishes to dive deeper on a certain subject.

At this point, the process is not completely autonomous as human moderators will still need to be in control for the quality of the source material. But after all, we probably do not want an AI to take over the publishing industry anyway.