EL2805 Reinforcement Learning

Starting end of this month, I will be teaching the course EL2805 Reinforcement Learning together with Prof. Alexandre Proutiere and Damianos Tranos.

The course provides an in-depth treatment of the modern theoretical tools used to devise and analyse RL algorithms. It includes an introduction to RL and to its classical algorithms such as Q-learning, and SARSA, but further presents the rationale behind the design of more recent algorithms, such as those striking optimal trade-off between exploration and exploitation. The course also covers algorithms used in recent RL success stories, e.g., deep RL algorithms.

You can find my notes here.

ERNSI '18

This week, I attended the annual workshop of the European Research Network on System Identification (ERNSI) that was held at Pembroke College, Cambridge, U.K. There were many interesting talks and discussions. I presented a poster on method of moment identification of hidden Markov models.

STHLM TECH '18

Next week, I’m attending the STHLM TECH FEST 2018, which brings together 14 Tech Conferences, 50+ events, and totaling over 36,000 expected guests. Looking forward to many exciting discussions and talks – see you there.

CDC'18

I’m glad to announce that our paper Inverse Filtering for Linear Gaussian State-Space Models has been accepted for presentation at the 57th IEEE Conference on Decision and Control (CDC’18), that will be held 17-19 December in Miami Beach, FL, USA.

The paper extends the results in our 2017 NIPS paper (’‘Inverse Filtering for Hidden Markov Models’’) to linear Gaussian state-space models.

ICML'18

This week, I’m attending the 35th International Conference on Machine Learning (ICML 2018) in Stockholm.

If you are also attending, be sure to check out the paper Bayesian Model Selection for Change Point Detection and Clustering, which was written and will be presented by two of my collegues at the department: Othmane Mazhar and Cristian Rojas.

SYSID'18

I’m a volunteer and part of the crew organizing the 18th IFAC Symposium on System Identification (SYSID 2018), which this year will be held at KTH.

Reglermöte 2018

I presented ideas on how to extend the results in our last year’s NIPS paper (’‘Inverse Filtering for Hidden Markov Models’’) to linear Gaussian state-space models at the Swedish Control Conference (Reglermöte).

The work has been submitted to CDC’18, and is currently under review.

MATLAB EXPO

I’m attending the MATLAB EXPO 2018 in Stockholm tomorrow – looking forward to the interesting talks on:

• Keynote: Are You Ready for AI? Is AI Ready for You?
• Unlocking the Power of Machine Learning
• Demystifying Deep Learning
• Deploying Deep Learning Networks to Embedded GPUs and CPUs

and others.

STAR Meeting

As a step in our continued collaboration with abdominal aortic aneurysm (AAA) researchers at KI, I attended the Stockholm Aneurysm Research Group (STAR) meeting. It featured interesting talks by, amongst others, Prof. Christian Gasser and Joy Roy, who presented and discussed current trends in AAA research.

Licentiate Seminar

I presented, and successfully defended, my licentiate thesis Hidden Markov models: Identification, control and inverse filtering (available here) today.

I would like to thank Prof. Eric Moulines (“…undeniably one of the best international experts in Machine learning, Mathematical and Computational statistics, as well as Non-Linear Time series analysis and Markov Chains.”) for being the special reviewer of my thesis.

Our discussion was very constructive, and I received a lot of valuable feedback.