Den här sidan är utskriven från Högskolan i Halmstads webbplats (www.hh.se). Texten uppdaterades senast den 2018-08-20. Besök webbplatsen om du vill vara säker på att läsa den senaste versionen.
Slawomir Nowaczyk was already a merited academic when joining Halmstad University eight years ago. Having finished his undergraduate and master studies in Poland, he moved on to earn a PhD in machine learning at Lund University and then returned to Poland to complete a Post doc.
– In Poland, academic collaborations with the industry are very rare, and that is what attracted me to Halmstad. Here, I was connected to Volvo immediately and throughout the years I have had the opportunity to initiate and take part in several collaboration projects with different companies, says Slawomir Nowaczyk.
– A major benefit with the strong focus on collaborations outside the academic world is the access that it gives to “real problems” that the industry faces every day.
– Being able to talk to people and understand the challenges they encounter when conceptualising, designing, creating and maintaining different products and services is a great inspiration. Helping them to come up with ideas and solutions is one of my main missions. What the industry faces every day is always a lot more complicated and intricate than the ideas I come up with sitting on my own in my office. There is always an entangled context that the solution needs to fit into, which poses additional constraints on what is feasible and what is not.
Slawomir Nowaczyk spends on average two days a week at Volvo in Gothenburg and is currently involved in a number of research projects concerning trucks and buses, in particular building self-monitoring systems for them. The collaboration started with Redi2Service project in 2010, in cooperation with Volvo Technology, focusing on diagnostics and prediction of maintenance as well as business models and new services that can develop from modern technical solutions. Current projects include ARISE and HEALTH, all funded by Vinnova FFI programme.
– I’m mainly looking at the data and aim to give the vehicles self-monitoring capabilities. Today’s trucks have a lot of digital sensors that collect data, and computers that process it. Many different components interact with each other in complex and often unpredictable ways. One of the main challenges is how to make these computer systems learn to predict failures before they happen, and make them figure out for themselves when parts need to be replaced.
When Slawomir Nowaczyk started to study computer science at master level he quickly got into machine learning. The idea of a computer learning new things, not only mindlessly repeating actions it was told to perform, appealed to him and has been the core of his research throughout his academic career.
– I want to learn about taking advantage of the ubiquitous data that is being collected nowadays, and what you need in order to provide computers with structures that allow them to generalise and build new knowledge. Gaining knowledge is something we do naturally as human beings but we don’t know exactly how we do it. Now we are trying to equip our computer systems with similar competences.
– It’s very exciting not only because of all the new things such intelligent computers will be able to do, but also due to all the new things we can learn about the nature of our own thinking processes.
One of the biggest challenges in machine learning, according to Slawomir Nowaczyk, is to add some of the core human characteristics to a computer – for example curiosity.
– In the fifties, artificial Intelligence researchers believed that playing chess is the ultimate intelligence test; once the computer can beat any human player, we will have achieved truly intelligent machines. But there is a lot more to real intelligence than just the mathematical or logical reasoning, he says and continues:
– If you put a person into a new environment, that person will not stand still, he will move around and interact with that environment. That is a characteristic that computers lack – but in order to be useful in the future they definitely need that kind of a trait. We can never predict everything that computers will encounter. They have to be able to identify things as interesting and try to deal with different novel situations on their own.
Slawomir is currently supervising six PhD students.
– Being an advisor to so many bright young people who are just now learning what does it mean to be a scientist is very rewarding, but also very challenging at the same time. Each of them is engrossed in their own project, with their own goals and their own dilemmas and frustrations – and it’s up to us, the senior researchers, to show them how it all fits together.
At CAISR, a lot of effort is made on aligning research and finding similarities between different approaches leading to very fruitful interdisciplinary exchange.
– We try to come up with methods and solutions that can work not only for a single problem, but for a whole bunch of them. We try to identify the general principles within each particular project we work on. In that way, the algorithms we create can hopefully be used in many different ways. This is something I find very interesting and want to go deeper into.
Collaborating with the industry has provided Slawomir Nowaczyk with an extended network of academics and professionals, and this has given him insights – for example that many researchers face similar problems.
– The problems Volvo has with their trucks bear many similarities to the problems facing hospitals with their patients – once you look at data and large numbers, instead of what exactly those numbers mean, he says and describes a current project concerning hypertension patients.
– In this project we are trying to predict which patients will not fulfill their prescriptions and then we try to support them in order to improve medication compliance. It’s a very important thing that will pay off in the future.
– Hypertension, at least in the milder forms, is example of disease that doesn’t affect your daily life all that much today, but will do so in 5-10 years. The first phase of the project is to analyse why different patients don’t take their medicine.
According to Slawomir Nowaczyk, there are a number of reasons. Some forget it, some mix up which pills to take at what time, some patients don’t like the side effects. There is plenty of research on different possible reasons, so there is no need to start from scratch, and the project will build on that foundation.
In the future he wants to continue looking into similarities across different domains.
– My drive is to solve problems and come up with new solutions. Every time I solve a particular problem I get to discover new applications to which this solution can be adapted – or I get new ideas for improvements to address issues I did not think about before. Probably the most concrete goal in my professional life is to actually make a real impact towards building an artificial truly intelligent system, says Slawomir Nowaczyk.
Text: HANNA JOHANSSON
Photo: JOACHIM BRINK
This text was originally written for CAISR:s annual report.
Title: Docent, Doctor, Associate professor in data mining.
Topics in focus: Data Mining, knowledge representation, joint human machine learning, self-organising anomaly detection, big data.