In the recent years we have been observing a
few significant trends in robotics emerging
from the economic and social environment subjected to the fourth industrial
revolution [1], DARPA Grant Challenge Projects (DRC) [2], Internet of Things
(IoT), IBM’s Watson [3], to name only a
few.
Based on my experience, the deep learning platforms might be partially in responsible for
Kirobo and many other projects. Deep learning provides very convenient (re-)
packaging of existing AI algorithms to the point in which the economics of
software development have changed. AI is much more economically feasible now.
One particular aspect, that is still difficult to handle by
machine learning – based AI is common
knowledge. Common knowledge in the AI jargon means a knowledge base of
simple inferences e.g. if you drop your phone on the floor it
will break. There are millions of similar inferences that impact our everyday
lives. Douglas Lenat and his CYC project [4] attempted to solve this problem by
manually creating rules for a system capable of reasoning about common
knowledge. After a few decades of painstaking development, CYC might be getting
ready for primetime.
The recent creation of Kirobo shows that small robots
connected through smartphones, hotspots
or other networks to powerful cloud computing platforms will be increasingly
important. Small robots equipped with vision systems and other sensors will
become an efficient user interface
platform for personal assistant systems capable of speech and gesture
recognition, spatial awareness, interconnectivity with other devices [5].
From the interface design point of view, attaching a likable personality to the information stream allows
for reduced information load. There
have been many attempts to create effective user interfaces with a personal touch and machine learning in the
background. The Ms. Dewey search engine
project by Microsoft [6], although discontinued, was one of the most ambitious attempt
to create a user interface with a distinctive and pleasant personality.
The research behind advanced user interfaces is very
significant and mature. Now, there are massive computing resources that can be supporting
such systems. It feels that Ray Kurzweil with his singularity and transhumanism
predictions [7] appear increasingly on target.
[1] The Fourth
Industrial Revolution, Klaus Schwab, World Economic Forum: https://www.amazon.com/dp/1524758868/ref=cm_sw_r_tw_dp_x_f79DybHYSEQP3
[4] www.cyc.com