Amazon’s online book store contains thousands of auto-generated ebooks.
Shady authors often hack together new texts from other books with hopes of making quick money from the effort.
Unfortunately for the fake writers, the texts are usually unreadable. As a result, they never make as much money as they’d hoped.
All that is about to change.
A scholarly publisher – Springer Nature – just published its first book generated by a machine learning algorithm.
Since the book is reportedly as dull as ever, we can safely say that it’s no John Grisham page-turner. However, the researchers did not design the algorithm to replace authors like Stephen King or J.K. Rowling in the first place.
As you can tell from its title – Lithium-Ion Batteries: A Machine-Generated Summary of Current Research – the ebook is missing a snappy dialogue and a thrilling plot. In other words, Hollywood’s studios won’t be using this tech any time soon.
So, why did the AI researchers at the Applied Computational Linguistics (ACoLi) lab, Goethe University in Frankfurt, Germany, develop the algorithm, you ask?
According to Gizmodo, the scientists created an ebook using machine learning to make sense of the extensive study on lithium-ion batteries.
How a Machine Wrote a Book On Lithium-ion Batteries
The lithium-ion battery has become a popular field of study. Within the past three years, there have been over 53,000 publications on the battery.
With so much attention on this field, staying on top of the research is almost impossible. That’s why Springer Nature developed a machine-generated publication to turn the firehose of data into a manageable trickle.
Here is how it works.
First, it analyses thousands of publications under the subject matter to select only the relevant ones. Then, it parses, condenses and organizes pre-approved, peer-reviewed papers from Springer Nature’s online database into logical and consistent chapters.
That means each chapter or section focuses on a different aspect of the battery research. Also, the algorithm automatically generates summaries for the articles in each chapter.
Think of the result as not just any book, but a Reader’s Digest for researchers and scientists. Rather than reading hundreds of thousands of documents, you can simply read a machine-generated 180 pages to remain up to date.
The algorithm is not limited to innovations in lithium-ion battery field alone. Researchers can conveniently tweak it to any scientific research.
Fantastic, this is very useful for the scientific community. Machine learning will shoulder the burden but probably involved significant limitation.