DarwinAI: “AI Building AI” to Solve Deep Learning’s Complexity Challenges

An Intellyx Brain Candy Brief

Deep Learning, a subset of Machine Learning that uses multiple layers of processing in what are called Deep Neural Networks (DNN) to create complex algorithms, has been both one of the most powerful approaches to AI and one of the most difficult, expensive, and time-consuming methods to use. Still, its unique capabilities have made it the best choice for things like computer vision, speech recognition, etc. For industries and use cases that require this type of complex AI, this has created a conundrum: how to make the application of deep neural networks easier, faster, and less expensive. Until now, this question has had few answers.

Upstart DarwinAI thinks that it has one. Rooted in decades of academic research, the company has introduced a platform that it believes will help transform the application of deep learning and deep neural networks in use cases such as autonomous vehicles, consumer electronics, and intelligent hardware chips. Its approach, which it calls Generative Synthesis, effectively applies AI to deconstruct and optimize deep learning models.

By doing so, the company claims that it can help organizations significantly increase the speed and effectiveness of the resulting DNNs, which in turn, makes it possible to apply them in places where organizations could not otherwise deploy them because of resource constraints. The company, therefore, claims to help clients facilitate the deployment of AI at the edge. This optimization also has the effect of accelerating the development process of these DNNs, sometimes reducing development time from years to weeks. Beyond this optimization, however, the deconstruction process also has the effect of overcoming one of the most significant challenges with deep learning: the black box problem. Because the platform exposes the layers of cognition in order to optimize it, it also can help explain how deep learning models make decisions and recommendations, which can help organizations both produce better models and comfort uneasy would-be users of this powerful AI.

Copyright © 2019 Intellyx LLC. Intellyx advises companies on their digital transformation initiatives, and helps vendors communicate their agility stories. At the time of this writing, none of the organizations mentioned in this article are Intellyx customers. To be considered for a Brain Candy article, email us at pr@intellyx.com.

SHARE THIS: