Bogdan Penkovsky CV Film photography Other Logs In
Chimera expert. Artist. Adventurer.
Currently I work at the FEMTO-ST lab in Besançon, France. The main research interest is in the field of reservoir computing. The story began in May 2012, at the conference in Mellas, Crimea, Ukraine where I attended a brilliant talk by Prof. Laurent Larger. I was already familiar with artificial neural networks, and the talk had further ignited my interest. I learned that those structures can be implemented with the help of a laser or even a bucket of water. That was my first encounter with reservoir computing. Eventually, I pursued my Ph.D. tackling problems related to this fascinating subject. The official title of my thesis was "Theory and Modeling of Complex Nonlinear Delay Dynamics Applied to Neuromorphic Computing" [pdf].
I do support an idea of open source. What is appealing to me is the ‘license’ that not only allows, but encourages the constructive criticism (hence, the source is open). In my own vision, open source is sometimes a creativity and exclusivity that you can afford instead of commercial perfectness and mass production.
At present, I do adore functional programming paradigm. I heard about the paradigm and about Haskell programming language before. But I had never dealt with it until my eye occasionally caught a bulletin at UKMA, where I did my Master's degree. The author of the bulletin, printed in a monospace font, was Semen who later introduced me the world of Haskell (see also Ukrainian translation). Here, I would also like to acknowledge the CLaSH compiler, work of Christian Baaij based on the same functional programming language ideas. The compiler ultimately helped me to build a reservoir computer during my Ph.D.
My Ph.D. in three minutes
Below is my Ph.D. thesis promo recorded on the regional final of "Ma thèse en 180" contest. The speech is in French, but it will take less than 180 seconds, I promise.
L. Larger, B. Penkovsky, Y. Maistrenko, Laser chimeras as a paradigm for multistable patterns in complex systems. Nature Communications 6, 7752 (2015).
See my CV for the complete list of publications.
Multi Expression Programming is a genetic programming variant encoding multiple solutions in the same chromosome. Essentially, a chromosome is a computer program consisting of reusable subprograms. The remarkable thing about MEP algorithms is that they are programs actually writing programs.
Learning is yet another open source library to facilitate machine learning research in Haskell. It provides evaluation metrics and other reusable machine learning tools.
Cellular automata is a classical example of highly complex systems achieved when combining simple rules. For fun and glory, you may build and observe your cellular automata in FPGAs, programmable logic circuits.
Delay differential equations (DDE) is a Haskell microlibrary I have extracted after my PhD. My work involved a lot dynamical systems described by DDEs. I kept switching between several tools, still I had no fast and flexible library. So here is one now.
I wanted to see Himalayas since I was eight. Suddenly, the dream became real. Twelve years later. It was a backpack trip to India. Getting lost in a 22 million city. A chance to visit numerous Buddhist places (the residence of Dalai Lama in Dharamshala, stupas in Bodh Gaya, Rajgir, Vaishali, etc.), the Golden Temple, as well as Hindu funerals in Varanasi. Subsequently, there were Dead Sea and Jerusalem in Israel; cuisine of Gaziantep, hot Antalya, ancient mosaics in Turkey; Indonesia with its volcanoes, tropical rains, corals, too much of ayam goreng and lots of islands; Germany, Switzerland, Spain, Malaysia, Brunei, Denmark, Luxembourg, Italy, San Marino, Monaco, Greece, Albania, Cuba, Hungary. Now I stay in one of the beautiful parts of Europe, Eastern France.
During my Indonesian trip in 2013, when everyone took digital pictures, I met Oleg Shchapov. He, who looked on photography from a different perspective, was an enthusiastic film photographer. The very final push to try the film again was this article shared later by him.
My attempt to rediscover the film is motivated by the question: what are the differences between the photo media? Film provides you an artifact, a tangible result of your work. You can develop it. You can touch it. And you can loose everything by exposing undeveloped film to the light. On the other hand, you do not need to rely on your hard drive or a CD having a limited lifetime. Film, on the contrary, can live much longer. Working with a mechanical camera, you never worry about frozen batteries or wet electronics. It happened to me once when I was shooting in the cold foggy morning mountains. The battery failed and the only working camera was the old-school revueflex. Another time, in Indonesia, the digital camera went down due to the high humidity. Thus, no pictures from that day here.
Kiev 4a revueflex Minolta SRT 101 Minolta X-700 Leica R4
Last, but not least
Mango lassi, mathematics, programming and etc. The background of this page is a chimera, monstrous fire-breathing creature. Here and here is highlighted my work on chimeras in nonlinear time-delay systems. Here are the slides from the PhysCon 2017 talk.