Description
Leader of the group Machine Learning: Peter Grunwald.
Our research group focuses on how computer programs can learn from and understand data, and then make useful predictions based on it. These algorithms integrate insights from various fields, including statistics, artificial intelligence and neuroscience.
Machinelearning applications are increasingly part of every aspect of life, from speech recognition on cell phones to illness prediction in healthcare. One common problem is extremely polluted data, for which no single model can provide adequate explanations. At CWI we address this issue with statistical machine learning based on combining predictions from different models and experts in order to achieve reliable conclusions.
We also study how networks of neurons in the brain process information, and how modern deeplearning methods can benefit from neuroscience. We develop novel neural networks, like Deep Adaptive Spiking Neural Networks, and also theoretical models of neural learning and information processing in biology. Applications of our work range from lowenergy consumption neural machine learning to neuroprosthetics, to increased insight into the question of how the brain works.
News
Sander Bohté, professor by special appointment of Cognitive Neurobiology at UvA
Sander Bohté (1974) has been named professor by special appointment of Cognitive Neurobiology, specialising in Computational Neuroscience, at the University of Amsterdam (UvA)'s Faculty of Science. The chair was designated on behalf of the Science Plus Foundation (Stichting Bèta Plus). He will be combining his professorship with his position as senior researcher at CWI's Machine Learning group.
Sterkenburg argues universal method of prediction is impossible
Is it possible to formulate a universal method of prediction? PhD candidate Tom Sterkenburg of CWI’s Machine Learning group argues in his thesis that it is not. He will defend his thesis on 18 January 2018.
CWI participates in new NWO Perspectief programme
In the coming years almost a hundred researchers are going to develop innovative technologies together with industry and social organisations. That will happen in six new Perspectief programmes, which have been given the green light by NWO, Netherlands Organisation for Scientific Research, on 21 November 2017. CWI's Machine Learning group participates in the programme Efficient Deep Learning Systems.
CWI researchers selected as ACM Future of Computing Academy members
CWI researchers Tim Baarslag and Wouter Koolen have been selected as members of the of the ACM Future of Computing Academy (FCA).
Current events
ML Seminar: Thomas Moerland (Delft University)
 20180904T11:00:00+02:00
 20180904T12:00:00+02:00
ML Seminar: Thomas Moerland (Delft University)
Start: 20180904 11:00:00+02:00 End: 20180904 12:00:00+02:00
Everyone is welcome to attend the ML seminar 'Monte Carlo Tree Search for Asymmetric Trees' given by Thomas Moerland.
Abstract:
We present an extension of Monte Carlo Tree Search (MCTS) that strongly
increases its efficiency for trees with asymmetry and/or loops.
Asymmetric termination of search trees introduces a type of uncertainty
for which the standard upper confidence bound (UCB) formula does not
account. Our first algorithm (MCTST), which assumes a nonstochastic
environment, backsup tree structure uncertainty and leverages it for
exploration in a modified UCB formula. Results show vastly improved
efficiency in a wellknown asymmetric domain in which MCTS performs
arbitrarily bad. Next, we connect the ideas about asymmetric termination
to the presence of loops in the tree, where the same state appears
multiple times in a single trace. An extension to our algorithm
(MCTST+), which in addition to nonstochasticity assumes full state
observability, further increases search efficiency for domains with
loops as well. Benchmark testing on a set of OpenAI Gym and Atari 2600
games indicates that our algorithms always perform better than or at
least equivalent to standard MCTS, and could be firstchoice tree search
algorithms for nonstochastic, fullyobservable environments.
ML Seminar: Jaron Sanders (Delft University)
 20180920T10:00:00+02:00
 20180920T11:00:00+02:00
ML Seminar: Jaron Sanders (Delft University)
Start: 20180920 10:00:00+02:00 End: 20180920 11:00:00+02:00
We have the pleasure to announce our CWI Machine Learning seminar with Jaron Sanders with the title: Optimal Clustering Algorithms in Block Markov Chains
This paper considers cluster detection in Block Markov Chains (BMCs). These Markov chains are characterized by a block structure in their transition matrix. More precisely, the n possible states are divided into a finite number of K groups or clusters, such that states in the same cluster exhibit the same transition rates to other states. One observes a trajectory of the Markov chain, and the objective is to recover, from this observation only, the (initially unknown) clusters. In this paper we devise a clustering procedure that accurately, efficiently, and provably detects the clusters. We first derive a fundamental informationtheoretical lower bound on the detection error rate satisfied under any clustering algorithm. This bound identifies the parameters of the BMC, and trajectory lengths, for which it is possible to accurately detect the clusters. We next develop two clustering algorithms that can together accurately recover the cluster structure from the shortest possible trajectories, whenever the parameters allow detection. These algorithms thus reach the fundamental detectability limit, and are optimal in that sense.
Members
Associated Members
Publications

Sander Bohté  Bionieuws  020618. (2018). Sander Bohté  Bionieuws  020618.

Grünwald, P.D. (2018). Safe probability. Journal of Statistical Planning and Inference, 195, 47–63. doi:10.1016/j.jspi.2017.09.014

Eén alarmerend bericht, en iedereen heeft het idee: het gaat héél slecht  NEMO Kennislink  26022018. (2018). Eén alarmerend bericht, en iedereen heeft het idee: het gaat héél slecht  NEMO Kennislink  26022018.

Sterkenburg, T.F. (2018, January 18). Universal prediction : a philosophical investigation.

Leitao Rodriguez, A, Oosterlee, C.W, Ortiz Gracia, L, & Bohte, S.M. (2018). On the datadriven COS method. Applied Mathematics and Computation, 317, 68–84. doi:10.1016/j.amc.2017.09.002

van Gerven, M, & Bohte, S.M. (2017). Editorial: Artificial neural networks as models of neural information processing. Frontiers in Computational Neuroscience (Vol. 11). doi:10.3389/fncom.2017.00114

Kaufmann, E, & KoolenWijkstra, W.M. (2017). MonteCarlo tree search by best arm identification. In Advances in Neural Information Processing Systems (pp. 4898–4907).

Science Café over kunstmatige neurale netwerken  Gelderlander ed. Nijmegen  021117. (2017). Science Café over kunstmatige neurale netwerken  Gelderlander ed. Nijmegen  021117.

Sterkenburg, T.F. (2017). A generalized characterization of algorithmic probability. Theory of Computing Systems, 61(4), 1337–1352. doi:10.1007/s0022401797749

Borovykh, A, Bohte, S.M, & Oosterlee, C.W. (2017). Conditional time series forecasting with convolutional neural networks. In Lecture Notes in Computer Science/Lecture Notes in Artificial Intelligence (pp. 729–730).
Software
Squint: Experimenting in Prediction with Expert Advice problems
Squint provides a codebase to perform numerical proofofconcept experiments in learning theory, particularly in Prediction with Expert Advice problems, a core problem in learning theory.