Skip to main content

Data-limited Learning Workshop

Save to calendar

Nov 24

Date and time: 24 November 2023, 09:00-17:30 CET followed by dinner
Title: Data-limited Learning Workshop
Where: Digital Futures hub, Osquars Backe 5, floor 2 at KTH main campus


A maximum of 50 participants are onsite at the Digital Futures cafeteria. First come, first serve.

This workshop is by invitation only, but if you would like to join without having received an invitation, please contact Anubhab [].

Welcome to the Data-limited Learning Workshop on Friday, 24th of November!

The workshop will discuss progress and challenges related to learning with limited data access, focusing on two application areas: continuous bioprocessing and reinforcement learning in cyber-physical systems. The workshop is part of the Digital Futures project Data-limited Learning of Complex Dynamical Systems.

Modern machine learning tools often require tremendous amounts of data to learn complex dynamical systems’ behaviour or perform simple tasks. However, many real-world systems are complicated or expensive to obtain large quantities of data from. Researchers from multiple disciplines will present their research in this area, intertwined with talks from keynote speakers with expertise in the field. There will be ample opportunities for discussion and more informal activities with the possibility to mingle.

The workshop will focus on theory and practical results and combine technical discussions with informal interactions. Presentations will be held both by project participants and invited speakers. We will celebrate achievements in the field and look forward towards future problems and relevant applications for the research. Join us to gain insight into data-limited learning and discuss results and applications over lunch and coffee. The workshop will be rounded off with a dinner for interested participants.


9:00 AM – 9:15 AM: Welcome Coffee

9:15 AM – 10:00 AM: Introduction

(David Broman, Veronique Chotteau, Alexandre Proutiere, Håkan Hjalmarsson)

10:00 AM – 10:15 AM: Coffee break

10:15 AM – 11:00 AM: Keynote by Assistant Professor Matthieu Barreau, KTH

11:00 AM – 12:00 AM: Networking Activity

12:00 PM – 1:30 PM: Lunch

1:30 PM – 3:15 PM: Posters and discussion session 1


  1. Stefan Stojanovic—Spectral Entry-wise Matrix Estimation for Low-Rank Reinforcement Learning
  2. Anubhab Ghosh—DANSE: Data-Driven Non-linear State Estimation in an Unsupervised learning setup
  3. John Wikman—Memorized Rollouts for Interaction Delayed Reinforcement Learning in Near-Deterministic Environments
  4. Martin Orrje—Building Cyber-Physical Systems as Tools for Reinforcement Learning Research
  5. Oscar Eriksson—Partial Evaluation of Automatic Differentiation for Differential-Algebraic Equations Solvers
  6. Oscar Eriksson, Martin Orrje, John Wikman—End-To-End Design for Automatic Control of Cyber-Physical Systems

3:15 PM – 5:00 PM: Posters and discussion session 2


  1. Mirko Pasquini—Model-based medium optimization in continuous perfusion cultures
  2. Kévin Colin—Optimal Adaptive Exploration Strategies for Linear Quadratic Adaptive Control
  3. Lars Hummelgren—Expression Acceleration: Seamless Parallelization of Typed High-Level Languages
  4. Meeri Mäkinen—Workflow for Mining Process Relevant Knowledge from Transcriptomics
  5. Linnea Stjerna—Programming with Context-Sensitive Holes using Dependency-Aware Tuning
  6. Robert Bereza—Towards Scalable Identification of Non-linear Differential-Algebraic Equations with Process Disturbance
  7. Daniele Foffano—Conformal Off-Policy Evaluation for Markov Decision Processes

5:00 PM: Closing Remarks


Questions? Please contact Anubhab by sending an email to