Investigating new algorithms in "Learning Automaton" at Carleton University
This autumn I went to Carleton University in Canada to investigate and bring back some exiting newly evolved algorithms in "Learning Automaton" to our university. These techniques are powerful and much better than other solutions in this research field, and can be applied to problems of databases, health care and mobile radio communications.
My trip to Ottawa in Canada to visit the University of Carleton was amazing. First, let me tell you the reason for me travelling there. Chancellor's professor at Carleton University, B. John Oommen which is connected to UIA through our Artificial Intelligence Research Centre (CAIR), and some of his students have newly advanced some algorithms in “Learning Automaton” (LA). Associate Professor at UIA, Lei Jiao, is interested in this research and asked me if I could travel to Carleton University to meet with Prof. Oommen and his students to learn all about these algorithms and bring back the knowledge to UIA, so that we possibly can use it in some of UIA’s research. Lei Jiao did not have the time to travel himself, and I became the lucky one to travel in his place.
I need to admit that I was quite nervous about the task I was given. I was only going to be there for one week, what if I didn’t manage to understand the concepts in those few days? Luckily, it all went well. Prof. Oommen was a great supervisor, and together with his student, Ekaba Bisong, I worked hard to understand every detail of the algorithms and how to use them.
Roughly 30 000 students are studying at Carleton University, and my first day there was quite overwhelming. The buildings were amazing, and the lab I was working at was in the “Herzberg building”. Ekaba Bisong was a nice guy, and even though he was defending his master thesis the week I was there, he helped me with every question I had and showed me his understanding of the algorithms and how he had programmed them using Python.
The algorithms that I learned are the following: OMA, EOMA, PEOMA and TPEOMA. These are powerful algorithms for solving partitioning problems. Partitioning is a hard case to solve in “Learning Automaton” (LA), but with these algorithms, the problems are solved in much fewer iterations than other solutions in this research field. Making the OMA-algorithms the benchmark for solving equal partitioning problems with LA! Lei Jiao hopes to implement these algorithms for solving problems in mobile communications, and the possible applications for applying these algorithms are many. By using these new techniques in different applications, both the LA field in ICT and the application it is applied to, can evolve and hopefully enhance UIA’s research.
Thanks for the opportunity Lei Jiao, and for my pleasant stay Prof. Oommen and Ekaba Bisong.