Gå til hovedinnhold
Jump to main content

Award-winning research from WISENET

Recently, the WISENET Center at UiA was awarded a prize for its project during the prestigious IEEE Data Science and Learning Workshop 2021. 

This article is more than two years old, and may contain outdated information.

WISENET Center at UiA has received an award.
WISENET Center at UiA has received an award. This was awarded to Professor Baltasar Enrique Beferull-Lozano (pictured), along with PhD-student Rohan Money and postdoctoral fellow Joshin Krishnan.


  • IEEE Data Science and Learning Workshop is an annual conference organized by IEEE signal processing society (SPS). 
  • It aims to bring together researchers in academia and industry to share the most recent and exciting advances in data science and learning theory and applications. 
  • This year PhD-student Rohan Money’s paper, co-authored with Dr. Joshin Krishnan and Professor Baltasar Beferull-Lozano received “The Best Student Paper Runner-Up Award” for the paper  “Online Non-linear Topology Identification from Graph-connected Time Series” at the prestigious science workshop.

Awarded by experts

What does it mean to get this award? 

“It is highly competitive to get a paper accepted as people from many top universities submit their work in the workshop. It is a very proud moment to get this award in such a competitive conference. The award was based on the score given by internationally recognized expert reviewers; the papers with highest scores are rewarded with the best paper awards”, says professor and Director of WISENET Center, Baltasar Beferull-Lozano. 

Professor and Director of WISENET Center, Baltasar Enrique Beferull-Lozano.

Professor and Director of WISENET Center, Baltasar Enrique Beferull-Lozano. 

- Can you elaborate about your research? 

“Our WISENET Center is working on data analytics and machine learning for various applications, including smart water networks, Aquaponics, oil and gas plants and more. The common thing about these large-scale systems is that they contain many sensors to measure various parameters (such as temperature and pressure) and controllers for governing the system. Our focus area of research is how to extract meaningful information and perform different types of intelligent tasks from the vast amount of data generated by the sensors. For example, in an oil and gas plant, the oil is extracted from the well and processed in separators (separating oil, water, and gas)."

"If an event occurs in some of the wells, it will take some time to reflect the corresponding 1 effect in the separators. Suppose we can identify the sensors in the well that are causally related to the sensors in the separator; this would allow to predict an event happening in the separator based on information from sensors in the well. Our research is in line with this direction, and we have come up with an algorithm that is able to capture such causal relationships in complex and large sensor networks. Even though we have designed this algorithm and applied it using real data from the Lundin Edvard Grieg OG Platform, our work is quite general, that is, it can be applicable to complex finance engineering problems, medical signal processing”, he says. 

Complex and dynamic systems

- What were your main findings in your research in this paper? 

Postdoctoral Fellow, Joshin Parakkulangarayil Krishnan.

Postdoctoral Fellow, Joshin Parakkulangarayil Krishnan.

“We came up with an algorithm that captures causal relationships between sensors in a network monitoring industrial processes. Note that such large-scale systems are very complex and dynamic. The relationships between sensors are usually captured in the form of what is called a “topology”, which can be also interpreted (understood) by human operators (explainable machine learning), and can vary depending on the various control actions happening in the system."

"A significant contribution is that our algorithm can track these changes in a completely online fashion, which means that as soon as new data is available, the algorithm will update the current topology or graph of relationships (dependencies). These dependencies can be in fact exploited further to perform various intelligent tasks, such as prediction of important events, correction of sensor data and control. Notice that due to the complexity of these dependencies among sensor variables, usually it is not possible for human operators or engineers to capture these (sometimes hidden) causal interactions, therefore, our algorithm provides a higher level of intelligence for the operation of complex processes”, Beferull-Lozano says. 

Targets top quality research

- What does it mean to WISENET Center and UiA to get this award? 

PhD-student Rohan Thekkemarickal Money.

PhD-student Rohan Thekkemarickal Money.

“It is indeed a proud moment to receive an Award like this in such a competitive international conference, where the best researchers in the world are publishing their research results in these areas. This means that we have done something well. In fact, the WISENET Center always targets top-quality research and focuses only on publishing on the top venues to present our work. This award will probably give much-deserved recognition to WISENET Center internationally and will further motivate us to continue our hard work and top-quality research. In addition, we would like to thank Lundin Norway for providing us with data from the oil and gas Edvard Grieg plant to do our research”, says Baltasar Beferull-Lozano. 

Manager Kim Alexander Jørgensen at Lundin Energy Norway adds: 

“There has been a very good collaboration between WISENET and Lundin Energy Norway. We are looking forward to see the system being further tested in Lundin headquarters with the real-time data from the Edvard-Grieg O&G platform”, Jørgensen says.