Connectome AI meetup #7 was held on April 5, 2019, at BINARYSTAR’s Office in Ginza, Tokyo with the goal of showcasing the benefits of OpenCog’s open-source initiative towards the development of human-level AI.
The event also gave attendees the opportunity to learn about the latest technical developments in OpenCog’s framework, including natural language understanding & generation (GHOST), knowledge-base graphic representation & rule engine (AtomSpace), uncertain inference (PLN) and how one can contribute towards the creation of Artificial General Intelligence (AGI).
Dr. Ben Goertzel, PhD, CEO of SingularityNET
Ben is the CEO of the decentralized AI network SingularityNET and the Chief Scientist of Hanson Robotics. He is one of the world’s foremost experts in Artificial General Intelligence, a subfield of AI oriented toward creating thinking machines with general cognitive capability at – and beyond – human level intelligence. He has published 20 scientific books and 140+ scientific research papers and is the main architect and designer of the OpenCog system and associated design for human-level general intelligence.
At the event, he gave us an in-depth view of OpenCog, which is a platform for Artificial General Intelligence (AGI).
Decentralised networks like blockchain need a specific kind of AI that enables the network to do its own thinking. AGI is primarily based on the intelligence of the human mind. Moreover, the AGI being developed at OpenCog is based on the new concept of “atom space” where nodes and links act as two different types of atoms; each performing different functions such as perception, cognition, prediction etc.
Two key algorithms that are used for knowledge creation are (a) MOSES Probabilistic Evolutionary Learning and (b) Probabilistic Logic Networks (PLM).
There was a question about the difficulty of training the complex AGI, to which Ben replied that while a simple AI is trained, an AGI would need to be “educated.”
After Ben’s talk, Mr. Leung Man Hin from the OpenCog team of SingularityNET gave a workshop about the algorithms of OpenCog.
Some of the key open-source algorithms for Artificial General Intelligence include ECAN, PLN, OpenPsi, and MOSES. Leung provided participants with a detailed workshop about these libraries.
AI companies require a lot of data. This is considered as one of the major problems for startups and engineers developing AI, as the lack of open data means that the technology is only fully open to a limited number of organisations and companies. OpenCog is developing a framework for the democratization of AI.
It was a great opportunity to learn from and talk with the Connectome AI team and the OpenCog team.