News

GitHub Repository for Training Environment

Source code repository for search agents’ training environment is set up. The environment provides macros definitions in C++ that make the training programs easy to read. It implements the concepts of training stages, training bots that communicate with the agents, and alphabet keys for the communication that can be shared among the bots on different stages. An examinator bot can be assigned on each stage to control the training progress of the agent.

The training environment is used in some ELEPHANT examples of how to train Machine Understanding.

Elephant.Training Is Up

Elephant is an open-source project on how to train Machine Understanding.

The role of understanding for research and learning is well delivered in the poem of John Godfrey Saxe¬†‘The Blind Man and the Elephant’:

And so these men of Indostan, disputed loud and long, 
each in his own opinion, exceeding stiff and strong, 
Though each was partly in the right, and all were in the wrong!
Elephant
When no understanding, the hypotheses about the elephant are independent and contradictory.
With the knowledge there is an elephant, all the useful hypotheses get explained.

Machine Understanding is needed to introduce common sense and general knowledge in AI. It provides the basis for explaining how different kinds of knowledge relate to each other.

Machine Understanding relates to transfer learning, also known as learning to learn.