Perceiving, Imagining, & Reasoning Like Humans
Vicarious, an Artificial Intelligence (AI) start up company has attracted investments totaling over $56 million to date from venture capital firms and private investors including Elon Musk, Jeff Bezos, Mark Zuckerberg and others. Elon Musk, in a recent interview, talked about the potential threat of super intelligent computers one day outsmarting human creators.
Vicarious is attempting to build software that has the ability to think and learn as human beings do, using computational constants found in the brain. In order to accomplish this the company’s mission is to “build the next generation of AI algorithms.”
Replicating The Human Neocortex
The aim is to create software that does what human cortexes do.
The ultimate goal is for machines to crunch information and perform complex tasks such as finding answers to serious problems humanity currently faces such as eradicating diseases like cancer, anticipating adverse weather events and helping doctors diagnose illness.
The company admits such software is years off, possibly decades, but it is starting with the basics, making sure it’s artificial intelligence technology is “seeing” objects, shapes, textures, etc.
Vicarious identifies a number of areas they are “pushing the boundaries” on:
- generalizing from a limited number of training examples using deep neural networks, and machine learning algorithms.
- unsupervised learning and supervised learning with large amounts of data.
- understanding how the neocortex is a source of inductive biases and constraints.
- understanding the network structure and architecture of the brain and its role in human thought and generalizations.
- creating generative models that infer, detect, recognize and segment information and use structured probabilistic models to interpret information and causes.
The following video shows Vicarious co-founder Scott Phoenix.