Saikat Chatterjee

I am an associate professor in Dept of Information Science and Engineering, School of Electrical Engg and Computer Science, KTH, and a Fellow of Digital Futures. I also have a visiting researcher position at Karolinska Institute and Karolinska Hospital in the area 'AI for Health Care'. I also spend time in Oslo University Hospital, Oslo, Norway.

Signal procesingand machine learning- these two vast fields are highly interconnected. My research mainly is in these two fields. In few words, the research interests are processing and analysis of signals and/or data for inferences such as detection, estimation, classification, prediction, learning, and finally decision. Please note that I get (easily) excited by the potential and/or prospects of algorithm design, analysisand applications. Therfore, please moderate your expectations. To mention some topics:

• Signal modeling,such as sparsity, compressive sensing, dynamical systems.
• Statistical signal processing, statistical machine learning, deep learning.
• Speech,audio, image processing.
• Medical data analytics (AI for Healthcare).
• Perception for autonomous systems.
• Distributed machine learning
• Explainable machine learning (XML) or Explainable AI (XAI).

Explainable machine learning is close to my heart having a signal processing background. Medical data analysis is slowly becoming a passion due to inherent challenges and societal importance.

Acknowldgement to Research Funding bodies

I am fortunate to receive support of several funding bodies (perhaps mostly tax-payers money). My group is generously funded by (1) SSF - Swedish Foundation for Strategic Research (https://strategiska.se)  (2) Region Stockholm. (3) European Union (4) Digital Futures (https://www.digitalfutures.kth.se/) (5) WASP (https://wasp-sweden.org/)  (6) Companies like Ericsson, Scania, Saab, etc.

Present group members (PhDs and Postdocs)

I am currently supervisor / co-supervisor of the following scholars.

• Anubhab Ghosh (PhD; my role: main supervisor; co-supervisor: Mikael Skoglund @ KTH)
• Gustav Noren (PhD; my role: main supervisor; co-supervisors: Magnus Jasson @ KTH and Torbjorn Blomqvist @ Saab)
• Fredrik Cumlin (PhD; my role: main supervisor; co-supervisor: Magnus Jansson @ KTH)
• Dr. Yogesh Todarwal (Postdoc; my role: main supervisor; co-supervisor: Sebastiaan Meijer @ KTH and Eric Herlenius @ Karolinska)
• Dr. Xinqi Bao (Postdoc; my role: main supervisor; co-supervisors: Martina Scolamiero @ KTH and Sara Garcia Ptacek @ Karolinska)
• Dr. Satarupa Chakrabarti (Postdoc; my role: co-supervisor; main supervisor: Arvind Kumar @ KTH)
• Dr. Kaushiki Roy (Postdoc; my role: main supervisor; co-supervisor: Joakim Jalden @ KTH)

Past group members

I was supervisor/ co-supervisor of the following PhD and Postdoc scholars.

• Sandipan Das (PhD; my role: main supervisor; co-supervisors: Magnus Jansson and Maurice Fallon @ Oxford); graduated in 2024. Thesis title: "State estimation with auto-calibrated sensor setup".
• Antoine Honore (PhD; my role: main supervisor, co-supervisors: Mikael Skoglund and Eric Herlenius @ Karolinska); graduated in 2023. Thesis title: "Perspectives of Deep Learning for Neonatal Sepsis Detection".
• Mostafa Shayan (Postdoc; my role: co-supervisor, main supervisor: Sara Garcia Ptacek @ Karolinska)
• Alireza M. Javid (PhD; my role: main supervisor, co-supervisor: Mikael Skoglund); graduated in 2021. Thesis title: "Neural Network Architecture Design: Towards Low-complexity and Scalable Solutions".
• Xinyue Liang (PhD; my role: main supervisor, co-supervisor: Mikael Skoglund); graduated in 2021. Thesis title: "Decentralized Learning of Randomization-based Neural Networks".
• Ahmed Zaki (PhD; my role: co-supervisor, main supervisor: Lars K. Rasmussen); gradutaed in 2018. Thesis title: "Cooperative Compressive Sampling".
• Arun Venkitaraman (PhD; my role: co-supervisor, main supervisor: Peter Händel); graduated in 2018. Thesis title: "Graph Signal Processing Meets Machine Learning".
• Dave Zachariah (PhD; my role: co-supervisor, main supervisor: Magnus Jansson); graduated 2013. Thesis title: "Estimation for Sensor Fusion and Sparse Signal Processing".
• Amirpasha Shirazinia (PhD; my role: co-supervisor, main supervisor: Mikael Skoglund); graduated 2014. Thesis title: "Source and Channel Coding for Compressed Sensing and Control".
• Dennis Sundman (PhD; my role: co-supervisor, main supervisor: Mikael Skoglund); graduated 2014. Thesis title: "Greedy Algorithms for Distributed Compressed Sensing".

Master Thesis Opportunity

I am willing to supervise interesting master thesis projects, preferably within the above mentioned research topics, but not limited to them. Interested candidates may contact either by a mail or to drop in my office for an informal discussion.

Collaborations

I am open to collaborations. You are welcome to approach me. Few points below.

• KTH has an outstanding enviroment and many of my colleagues are experts, with whom I collaborate in a regular basis. You may directly approach them, or approach me to bring them on board.
• I have established collaborations with several institutions across the world.
• Fortunately I enjoy close cooperations with several companies and medical institutions, where I try to place my students for thesis and jobs.
• From this year 2022, supported by a generous grant from SSF, I spend time in Karolinska University Hospital and Karolinska Institutet for medical data analysis.
• I also like to provide consultancy in signal processing, machine learning and data science. Please note that the consultancy will be only offered if KTH has no competing interest.

What I do wish or like

I love my primary jobs - Teaching and Research. In research, I likealgorithm design and mathematicalanalysis. While I like various application scenarios,Medical Data Analysisin the larger area AI for Health Care is slowly becoming an objective, perhaps because Covid has taught us a meaning of life. A new wish - research fordefence of Sweden, Nordics or EU, given the sudden war situations. While I get (easily) excited by machine learning and sigal processing, I try to avoid 'extraordinary wishful claims', but search for theoretical motivations and analytical foundations. Naturally, we proceed withexplainable AI (XAI) and/orexplainable machine learning (XML).