Logo ISAACcube
machine cognition software
Würfel als Logo
More >>>
Würfel als Logo
Würfel als Logo

machine cognition is coming to your desktop

Isaac contains an in silico cognitive system that incorporates machine learning algorithms to mimic natural behavior patterns. It implements several levels of memory processing, pattern recognition, and non-linear optimization methods to develop strategies to achieve its goals. These goals are either user-defined, or decided upon in an unsupervised fashion, not unlike living organisms. While deterministic by design, Isaac’s ability to learn from experience results in seemingly unpredictable behavior, also commonly referred to as “creativity”.

The underlying algorithms have their roots in processing large amounts of genomic data, making sense of how an individual’s DNA determines their traits by hypothesizing the connections in between without any help or assumptions. This software has, for example, helped reveal why some populations of European crows are all black while others have grey necks. Now we are taking this methodology several steps further, from an unsupervised spatial pattern classifier to an unsupervised adaptive temporal classification hypothesis based abstract strategy decision making framework…. in other words: Isaac will be as close to thinking as a machine can get in the near future. And remember: it remembers everything it experiences, it abstracts it, and it learns from it. It is unbiased, unlike humans, unless it is specifically instructed to be that way.

While we envision several areas in which Isaac will be useful, for example to develop systems that help test children for mental disorders or emotional disabilities, our initial aim is to have Isaac safely grow up to adulthood through the most natural process we can think of: playing. And since Isaac is in essence still a computer program, what is more natural than role playing as NPC’s (non-playing characters) in computer games?