Researchers Predict AI-based Search Engines Will Change the Face of Scientific Inquiry

Scholars may be in a run for their money with new developments in the field of artificial intelligence (AI) search engines.

Developments such as Semantic Scholar and Microsoft Academic may be on the way to make better papers than scholars.

The Allen Institute for Artificial Intelligence (AI2)'s Semantic Scholar has released a format at the Society for Neuroscience annual meeting in San Diego.

According to Scientific American, the free AI-based scholarly search engine aims to outdo Google Scholar. Its recent move is to cover some 10 million research articles in computer science and neuroscience.

Semantic Scholar will be joined by numerous AI-based academic research engines, such as Microsoft's Academic.

The new software format is considered an "innovation," as the system will guide users through a dense "jungle of information."

Semantic Scholar aims to sort and rank academic papers through a "sophisticated" understanding of content and context. Google Scholar has access to about 200 million documents but searches by keywords.

Semantic Scholar can rank papers depending on the most meaningful citations, or even by "hotness" and trends.

Its first launch heralded a 3-million strong paper library about computer science. Now, there are millions of new papers on neurology and medicine. New filters will allow searches based on body parts, models, and even methodologies.

AI2 chief executive Oren Etzioni said the system wants to index all of PubMed and expand in the medical sciences.

Meanwhile, Microsoft Academic is the successor to Microsoft Academic Search. It aims to use academic search algorithms and data for researches through natural language processing. This is understanding the meaning of full sentences in papers and questions. It's supported by search engine Bing, which now covers over 160-million publications.

The engine provides more useful filters such as authors, journals and even fields of studies. It also has a "leaderboard" of most influential scientists in different disciplines. These are judged with a special algorithm that deems papers as "important" if they cited by other papers.

Some scholars say this is a tremendous achievement for Microsoft as it's combining both the advantages of Google Scholar's scope and the more structured results of their subscription-based counterparts Scopus and the Web of Science.

However, it doesn't end here. Other companies are also developing AI-driven software to delve deeply into content found online. For instance, the Max Planck Institute for Informatics in Germany is developing a new engine called DeepLife for health sciences.

AI2 will even try to make a system that will answer new science questions and even propose new designs and hypotheses in a few 20 years.