The SDM & RL core technology cluster provides a network for researchers working on sequential decision-making problems.

We, the SDM & RL cluster, focus on exploring the fascinating field of sequential decision-making. We are interested in developing algorithms and models that enable intelligent agents to make optimal decisions in dynamic and uncertain environments. Most of us research in the direction of applied or foundational (deep) reinforcement learning, while another large subset of us works with robotics and control theory.

Goal: The main purpose of the cluster is to provide a network for practitioners and researchers in the SDM & RL field, to advance our understanding of the fundamental principles underlying intelligent, sequential decision-making, to foster scientific discussion and collaboration, and to exchange practical knowledge.

Cluster Activities

With respect to the goal of the cluster, we meet bi-weekly via zoom and typically engage in one of the following activities:

    • Research discussions: Cluster members are encouraged to present their research in SDM & RL, engaging in valuable scientific discourse.
    • Paper discussions: The cluster discusses SOTA papers from the field. Typically this involves one person presenting a summary of the paper, followed by a discussion.
    • Practical knowledge exchange: The cluster discusses implementational details, hyperparameters, and heuristics associated with the complex code underlying sequential decision-making problems and algorithms.
    • External speaker invitation: The cluster occasionally invites external (potential industrial) speakers to talk about their SDM & RL research.

Communication

The cluster communicates via the ctc_reinforcementlearning@wasp-sweden.se mailing list and the private Slack channel #cluster-sequential-decision-making-and-rl on the wasp-sweden Slack. To join this cluster, use the cluster registration form and contact the current cluster leader.

Recent Cluster Member Publications

Most of our publications can be found on DiVA (this is along query, the site will load a while). Here we feature some of the most recent publications:

  • Aligning Human Preferences With Baseline Objectives In Reinforcement Learning (2023), Daniel Marta et al.
  • A Stack-Of-Tasks Approach Combined With Behavior Trees : A New Framework For Robot Control (2022), Marco Iannotta et al.
  • A Survey Of Behavior Trees In Robotics And Ai (2022), Jonathan Styrud and Matteo Iovino et al.
  • Adaptive Control Of Data Center Cooling Using Deep Reinforcement Learning (2022), Albin Heimerson et al.
  • Empirical Analysis Of The Convergence Of Double DQN In Relation To Reward Sparsity (2022), Samuel Blad et al.
  • Evaluating Sequential Reasoning About Hidden Objects In Traffic (2022), Truls Nyberg et al.
  • Finding Critical Scenarios For Automated Driving Systems: A Systematic Mapping Study (2022), Magnus Gyllenhammar et al.
  • Hierarchical Goals Contextualize Local Reward Decomposition Explanations (2022), Finn Rietz et al.
  • Importance Sampling Cams For Weakly-Supervised Segmentation (2022), Arvi Jonnarth et al.
  • Learn2Reg: Comprehensive Multi-Task Medical Image Registration Challenge, Dataset And Evaluation In The Era Of Deep Learning (2022), Niklas Gunnarsson et al.
  • Learning Optimal Antenna Tilt Control Policies : A Contextual Linear Bandit Approach (2022), Yassir Jedra et al.
  • MPR-RL : Multi-Prior Regularized Reinforcement Learning For Knowledge Transfer (2022), Quantao Yang et al.
  • Multi-Agent Exploration With Reinforcement Learning (2022), Alkis Sygkounas et al.
  • The Magni Human Motion Dataset : Accurate, Complex, Multi-Modal, Natural, Semantically-Rich And Contextualized (2022), Yufei Zhu et al.
  • Variable Impedance Skill Learning For Contact-Rich Manipulation (2022), Alexander Dürr and Quantao Yang et al.

Cluster Members

Keywords: Robotics, Reinforcement Learning, Planning
Affiliation: Lund University

Keywords: Curiosity driven exploration
Affiliation: Örebro University
Company: Nexer

Keywords: MPC, DRL, DL
Affiliation: Chalmers University of Technology
Company: Volvo AB

Keywords: Continuous state and action spaces, symbolic and usual RL, Sim2Real, Multi Task Learning
Affiliation: Lund University

Keywords: RL, risk-sensitive RL, Safe RL, epistemic / aleatory risk
Affiliation: Chalmers University of Technology
Company: Zenseact

Keywords: representation learning, model-based RL
Affiliation: Royal Institute of Technology

Keywords: Validation
Affiliation:

Keywords: Motion modelling, temporal imaging, 3D reconstruction,
Affiliation: Uppsala University
Company: Elekta Instrument AB

Keywords: Autonomous vehicles, safety assurance, risk-aware safety assessment, quantitative safety assessment, functional safety, AV architecture
Affiliation: Royal Institute of Technology
Company: Zenseact

Keywords: RL for computer vision
Affiliation: Lund University

Keywords: Deep Reinforcement Learning, Reasoning, Planning
Affiliation: Örebro University

Keywords: (Cloud)
Affiliation: Lund University

Keywords: Generative models, RL
Affiliation: Royal Institute of Technology
Company: SEB

Keywords: Motion-planning , autonomous vehicles, planning under uncertainty, pomdps,
Affiliation: Linköping University
Company: Scania

Keywords: Behavior Trees, RL, industrial-collaborative robotics
Affiliation: Örebro University
Company: Suzuki Garphyttan

Keywords: Behavior trees, industrial-collaborative robotics, learning
Affiliation: Royal Institute of Technology
Company: ABB

Keywords: RL theory, MAB, Adaptive control
Affiliation: Royal Institute of Technology

Keywords:
Affiliation:

Keywords: Vision, Robotics, RL, Autonomous vehicles
Affiliation: Linköping University
Company: Husqvarna

Keywords: Model based, uncertainty, generative models
Affiliation: Chalmers University of Technology

Keywords: DRL, RL in games, generalization, scalability
Affiliation: Royal Institute of Technology

Keywords: Cloud Computing, decision making
Affiliation:

Keywords: RL, Causal Inference, Bandits, Optimal stopping, Healthcare
Affiliation:

Keywords: Dual control, dual estimation
Affiliation: Lund University

Keywords: Autonomous vehicles, safe decision making, formal methods, discrete event systems, reactive synthesis, automatic abstractions
Affiliation: Chalmers University of Technology
Company: Zenseact

Keywords: Robotics, model-free RL, Sim2Real
Affiliation: Chalmers University of Technology

Keywords: Safe RL
Affiliation: Royal Institute of Technology
Company: Elekta

Keywords: Meta-reinforcement learning
Affiliation: Uppsala University
Company: Ericsson

Keywords: MAB, Optimal stopping, Information theory, Value of Information
Affiliation: Royal Institute of Technology
Company: Ericsson

Keywords: Decision-making, Biology, Neuroscience, Cooperative RL
Affiliation: Royal Institute of Technology

Keywords: Risk-sensititivity, multi-agent reinforcement learning
Affiliation: Linköping University
Company: Saab AB

Keywords: DRL, HRI, Human-in-the-loop
Affiliation: Royal Institute of Technology

Keywords: Off-policy evaluation, decision-making
Affiliation: Chalmers University of Technology

Keywords: Behavior trees, industrial-collaborative robotics, learning
Affiliation: Lund University

Keywords: Autonomous vehicles, motion planning, risk-aware decision making, pomdp:s
Affiliation: Royal Institute of Technology
Company: Scania

Keywords: Multi-agent Systems, Algorithmic Game Theory, Combinatorial Optimization
Affiliation: Linköping University

Keywords: (Under-supervised)
Affiliation: Chalmers University of Technology

Keywords: DRL, Multi-objective RL, Knowledge Transfer, Explainability
Affiliation: Örebro University

Keywords: Dual control
Affiliation: Lund University

Keywords: Decision-making, Healthcare, causal inference
Affiliation: Chalmers University of Technology

Keywords: Behavior Trees, RL, Robotics, Sim2Real
Affiliation: Royal Institute of Technology
Company: ABB

Keywords: Autonomous drug design, bandits, RL, active learning/online learning,
Affiliation: Chalmers University of Technology

Keywords:
Affiliation: Örebro University

Keywords: RL, DL, Safe RL, Safety and Reliability, Goal driven RL, Validation of safety in RL models
Affiliation: Royal Institute of Technology

Keywords: Power system control, Sim2Real, safe RL, model-based RL
Affiliation: Royal Institute of Technology
Company: HitachiABB

Keywords:
Affiliation:

Keywords: Transfer learning, Sim2Real, robotic manipulation, safe RL
Affiliation: Örebro University

Keywords: Deep Reinforcement Learning, Robotic manipulation
Affiliation: Örebro University

Keywords: Vision, Localisation and Mapping, Visual SLAM, Scene representation
Affiliation: Royal Institute of Technology
Company: Univrses

Keywords: Human motion prediction, dynamics mapping
Affiliation: Örebro University

Keywords: Electric vehicles, combinatorial bandits, online learning, RL, autonomous vehicles
Affiliation: Chalmers University of Technology

Keywords: Vision, Pose Estimation, Robotics, RL, Sim2Real
Affiliation:

Current Cluster Leader

Finn Rietz

PhD Student, Center for Applied Autonomous Sensor Systems, Örebro University