Debunking the myths about deep learning

Published on: February 20, 2019

The industry has set high expectations for deep learning. However, according to Rein Lemmens, we should look at these expectations with a healthy dose of suspicion.
In years to come, deep learning is expected to create high added value for existing and new products. Indeed, a wide variety of applications have already been developed based on this technology. But will deep learning be able to live up to the expectations?

Deep learning: breakthrough or bubble?

Flemish engineering community ie-net recently organized an event around that topic. More specifically, ie-net’s expert group on industrial vision and robotics invited several speakers to have a talk about the theme ‘Deep learning: breakthrough or bubble?’ The goal of the event was not only to provide an interpretation and background on deep learning, but also to critically examine this technology.

10 myths about deep learning

In his presentation ’10 myths about deep learning’, Rein Lemmens debunks some of the most persistent misconceptions about deep learning. Will all our work soon be automated? Will artificial intelligence make superhuman intelligence possible? Even worse, is the AI apocalypse upon us?
Rein examined a number of these myths and also offered solutions to avoid potential deep learning pitfalls, all of which were underpinned by suspiciously scientific case studies from the Kapernikov portfolio.
Here is the presentation ’10 myths about deep learning’ in full: 10 myths about deep learning