iQorians are authors! We are seriously impressed with the amount of smart and talented iQorians we have all over the world, and now, we want you to meet them. Check out one our iQorian Authors below.
Maciej Michalewicz, Ph.D
Maciej is part of the iQor Data Solutions & Analytics team, serving as President of iQor Warsaw, leading our team of data scientists and developers. Over the lifetime of his career, he has led the development of data analysis systems, information retrieval systems, knowledge discovery and intelligent information systems. Maciej has published over 50 articles, edited/co-edited 11 books, and has spoken at a number of international conferences. Maciej possesses expert knowledge of Bayesian networks, neural networks, rough sets, fuzzy systems, fuzzy modeling, and genetic algorithms. He holds a Master of Science degree in Applied Mathematics from the Warsaw Technical University, Poland and a Ph.D. degree in Computer Science from Academy of Sciences of Russia. From 1988 to 1991, Dr. Michalewicz led the Chernobyl Disaster data analysis project to investigate the medical effects and long-term consequences of the radiation fallout.
Maciej lives in Warsaw, Poland with his wife and daughter. He spends his free time swimming, playing Chess, Bridge and reading (many) books.
Maciej, you’re a leading mathematician. Can you tell us a little more about you and your history?
I’ve always been interested in applied mathematics and science while I was studying at the Department of Mathematics at Warsaw University. Many times, mathematician and science careerists focus on proving their own theoretical solutions. I wanted to do more than that. I wanted to help solve real-world problems. I was inspired by famous mathematicians such as Kuratowki, Sierpinski, Steinhaus, Ulam, and their purpose of teaching people how to solve complex problems with practical applications.
This is what launched me into the field of mathematics and computer science, which led me on a winding path into a number of industries, including medicine, military, and government.
When we think of scientists, we drum up images of Einstein, Newton, Marie Curie, and Tesla. We don’t hear of many modern standalone names anymore. Where are all the scientists hiding?
Many people think that the art of the ‘scientist’ is a dying. It’s true that the times of Einstein and solo discoveries have faded away (with a few modern-day exceptions; Stephen Hawking being a famous example), but exciting innovations and new developments are taking shape every day within professional teams across industries. iQor is one example but teams from Tesla Motors, IBM, Johnson and Johnson, to even local and national governments are making incredible advances.
What do think is the role of today’s scientist and mathematicians?
The most difficult problems usually require very deep scientific knowledge and creativity, as well as advanced technological skills. We need to continue to bring together the theoretical thinkers and the developers, statisticians, and lovers of numbers in order to complete the full life cycle of scientific development and problem solving. From finding, refining, and understanding complex problems, to applying the right mathematical models and building a solution that is practical and easy to implement in the outside world. My motto is that ‘it’s not one person’s job to solve complex problems’ and that has never been truer today.
We’ve read you have a background in ‘fuzzy systems’, 'rough systems', and 'Bayesian networks'. New terms to us and probably many of our readers. Can you explain what they are?
Currently we are witnessing an increase in the role of IT applications that control a variety of technical devices. Often times, control mechanisms (e.g. control of conditioning units, cars, drones, etc.) can be expressed in vague concepts (near, far, warm, high, etc.). Control mechanisms are based on the processing of strict sentential logic expressed in Boolean algebra. In many applications, 'fuzzy control systems’ support the control of these devices or machines using fuzzy logic. Roughly speaking, fuzzy logic uses degrees of truth as a mathematical model within various degrees of vagueness and is used to solve various practical problems.
Rough sets (complementary and a similar concept to fuzzy sets) allow us to find and express clear borders between objects related to a vague concept. There are objects which ‘definitely belong to a vague concept’ and objects which ‘definitely do not belong to vague concept’. Bayesian networks are a very unique statistical methodology allowing us to find the statistical cause and effect that traditional or classical statistics cannot find or explain.
Can you provide us some examples of systems you have worked on?
I helped build a system known as PATHOS, for data acquisition and data retrieval in hospitals. We were able to automate the medical documentation process using semi-natural language based on medical thesauruses. At the time, this solution was groundbreaking in that it required new methodologies that my team was responsible for implementing and could be used online on a world-wide scale. The system ended up winning the Polish Academy of Sciences award.
I also helped build a system that can evaluate the suitability of candidates applying to become military pilots. I worked with some of the best AI teams from Poland and examined over 400 attributes – schools attended, exam scores, psychological and physical tests and so on – to define the right compatible traits that matched to the military program’s needs and requirements. Our system automates about half of the selection process, which reduced the number of recruit dropouts by 20%, training costs, and led to a more a selective process and more qualified pilots making the cut.
Can you tell us a little bit about what you do at iQor?
I lead our group of software engineers, technology program leaders, and scientists that form our Data Solutions Analytics (DSA) team. And we do quite a lot! To put it simply, we focus mainly on providing comprehensive analytical support. We have been recently called in to analyze a new set of training scores. iQor’s Experience Innovation Lab recently launched a multitask training model for our agents to assess how their perform in simultaneous chats. These chats are done with iQor-developed bots mimicking normal customer interactions.
What type of technology are you and your team engaged with daily?
AI, machine learning, classical statistics, and a few client-specific tools. We continuously enhance and analyze our automatic cosmetic inspection process where computer vision software automatically recognizing defective set-top box units. Another exciting tool is our predictive repair system that uses algorithms to identify areas and units where the repair process can be shortened.
Any other projects that we can look forward too?
Together with Andrzej Jankowski we are planning enhance our interaction analytics platform, VALDI, to a tool of knowledge discovery that is capable of supporting all clients’ needs, both analytical and predictive using AI and machine learning. This will enable VALDI to provide value on a worldwide scale, allowing our company to gain significant competitive advantage.
Maciej’s recent papers can be found in Intelligent Information Systems 2001, which covers progress in theory and applications of artificial intelligence, machine learning, knowledge discovery, knowledge based systems and reasoning, intelligent statistical analysis and soft computing.
About Our Data Solutions and Analytics Team
Maciej manages an experienced and enthusiastic team of software engineers, technology program leaders, and data scientists that comprise the Warsaw office of our Data Solutions Analytics (DSA) team. This group of talented individuals combine systems architecture, cloud-centric software engineering, operations research, and leading applied analytics to generate commercial value from seas of data, and synthesize innovative product solutions for iQor, our clients, and the analytics and business intelligence marketplace.