According to artificial intelligence research by MIT Technology Review, the era of deep learning may end soon.
Almost everything you hear about artificial intelligence today is thanks to deep learning. This category of algorithms works by using statistics to find patterns in data, and it has proved immensely powerful in mimicking human skills such as our ability tosee and hear. To a very narrow extent, it can even emulate our ability to reason. These capabilities power Google’s search, Facebook’s news feed, and Netflix’s recommendation engine, transforming industries like health care and education. However, deep learning actually represents a small process in our quest to replicate our own intelligence.
Pedro Domingos, professor of computer science at the University of Washington and author of The Master of Algorithms, notes that the spike and fall of different techniques has long characterized AI research. In the MIT Journal of Technology, this process was reviewed on the occasion of one of the largest open source scientific databases called arXiv. We followed the words mentioned over the years to see how the field has evolved, with summaries of 16,625 articles in the “artificial intelligence” section until November 18, 2018.
Number of Studies Downloaded from ArXiv
(All articles available in the “artificial intelligence” section until 18 November 2018)
The analysis found three main trends: a shift towards machine learning in the late 1990s and early 2000s, an increase in the popularity of neural networks that began in the early 2010s, and growth in reinforcement learning over the past few years. But let's look at artificial intelligence that covers all of this:
The studies examined show that in 2018, the number of studies on artificial intelligence, or rather the number of articles written, reached its peak with 3697. This number was only 7 in 1993. After a strange decline after 2013, it started to rise again as of 2016.
However, this research signals that a change will occur soon. When the study findings are examined, the transformation of artificial intelligence (every ten years), which we use intensively today, has come to an end by 2020. This indicates that the deep learning system, which started to rise as of 2010, will soon be replaced by another system. It seems that following the evolution of artificial intelligence, which is expected to dominate all professions, will be important to make a difference.
