Posts By: Marc Deisenroth

J.P. Morgan PhD Fellowship to Sanket

Sanket Kamthe has been selected as a 2019 J.P. Morgan PhD Fellow. Congratulations!

3 Papers accepted at NeurIPS

Hugh and James got papers accepted at NeurIPS, which is excellent news. Very well done! Gaussian Process Conditional Density Estimation Vincent Dutordoir, Hugh Salimbeni, James Hensman, Marc P. Deisenroth Orthogonally Decoupled Variational Gaussian Processes Hugh Salimbeni, Ching-An Cheng, Byron Boots, Marc P. Deisenroth Maximizing acquisition functions for Bayesian optimization James Wilson, Frank Hutter, Marc P…. Read more »

Dr Chamberlain

Congratulations to Ben Chamberlain for defending his PhD! Ben has done some great work on network embeddings, community detection at scale and inference of latent features of social media users. The PhD hat reflects some of the work Ben has done over the last years. 2/3 of PhD committee post-viva

Steindor’s paper accepted at UAI 2018

Steindor’s paper on Meta Reinforcement Learning has been accepted at UAI 2018. Congratulations! Steindor Saemundsson, Katja Hofmann, Marc Peter Deisenroth Meta Reinforcement Learning with Latent Variable Gaussian Processes Conference on Uncertainty in Artificial Intelligence, 2018

Simon’s paper accepted at ICML 2018

Congratulation to Simon, whose paper got accepted at ICML 2018   Simon Olofsson, Marc Peter Deisenroth, Ruth Misener Design of Experiments for Model Discrimination Hybridising Analytical and Data-Driven Approaches International Conference on Machine Learning, 2018

PLOS ONE paper accepted

Congratulations to Ben Chamberlain for his accepted paper Benjamin P. Chamberlain, Josh Levy-Kramer, Clive Humby, Marc P. Deisenroth Real-Time Community Detection in Full Social Networks on a Laptop PLOS ONE, 2018

Two papers accepted at AISTATS 2018

Congratulations to Sanket and Hugh: Their papers have been accepted at AISTATS 2018 Sanket Kamthe, Marc P. Deisenroth, Data-Efficient Reinforcement Learning with Probabilistic Model Predictive Control, Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), 2018 Hugh Salimbeni, Stefanos Eleftheriadis, James Hensman, Natural Gradients in Practice: Non-Conjugate Variational Inference in Gaussian Process Models, Proceedings of the International… Read more »

New lab members

We are happy to welcome Janith and James to the lab and Imperial.

Two papers accepted at NIPS 2017

Two papers from the group got accepted at NIPS: Hugh Salimbeni, Marc P. Deisenroth, Doubly Stochastic Variational Inference for Deep Gaussian Processes, Advances in Neural Information Processing Systems (NIPS), 2017 Stefanos Eleftheriadis, Thomas F. W. Nicholson, Marc P. Deisenroth, James Hensman, Identification of Gaussian Process State Space Models, Advances in Neural Information Processing Systems (NIPS), 2017 Congratulations to Hugh… Read more »

Deep RL Survey accepted at IEEE SPM

Our Brief Survey of Deep Reinforcement Learning has been accepted for publication in the IEEE Signal Processing Magazine. A Chinese translation is also already online. Kai Arulkumaran, Marc P. Deisenroth, Miles Brundage, Anil A. Bharath: A Brief Survey of Deep Reinforcement Learning, IEEE Signal Processing Magazine, 2017

Papers accepted at ECML 2017

Congratulations to Ben Chamberlain who got two papers accepted at ECML: Benjamin P. Chamberlain, Clive Humby, and Marc P. Deisenroth: Probabilistic Inference of Twitter Users’ Age based on What They Follow CH Bryan Liu, Benjamin P. Chamberlain, Duncan A. Little, and Angelo Cardoso: Generalising Random Forest Parameter Optimisation to Include Stability and Cost  

Paper accepted at KDD 2017

Congratulations to Ben Chamberlain, whose paper Customer Life Time Value Prediction Using Embeddings has been accepted as at KDD 2017. The overall acceptance rate was 21.4%. Well done!

Papers at NIPS Workshop

Hugh and Matt present their work at the NIPS workshop on Practical Bayesian Nonparametrics: Matthew C. H. Lee, Hugh Salimbeni, Marc P. Deisenroth, Ben Glocker, Patch Kernels for Gaussian Processes in High-Dimensional Imaging Problems, NIPS Workshop on Practical Bayesian Nonparametrics, 2016 Hugh Salimbeni, Marc P. Deisenroth, Gaussian Process Multiclass Classification with Dirichlet Priors for Imbalanced Data, NIPS… Read more »

New group members

Dean Garlick and Riccardo Moriconi are joining as PhD students today. Welcome!

Dr. Calandra

Congratulations to Roberto for a successful PhD completion!

Google Scholarship to Sanket Kamthe

Sanket Kamthe has been awarded a Google EMEA Scholarship for 2016/17. The Google EMEA Scholarship is highly competitive and only awarded to 10 students per year. Scholarships are awarded based on the strength of the applicant’s academic background, leadership skills and demonstrated passion for Computer Science. Congratulations!

Ben’s Talk at PyData

Ben gave a talk at pydata on his work on Real-Time Community Detection in Large Social Networks on a Single Laptop.

ICML Workshop on Data-Efficient Machine Learning

ICML 2016 Workshop on Data-Efficient Machine Learning (24 June 2016) Website: Submission deadline: 1 May 2016 Submission website: 1. Call for Papers We invite researchers to submit their recent work on the development and analysis of methods leading towards more data-efficient machine learning. A submission should take the form of an extended abstract of 2 pages in… Read more »

Dr Carstens

Lucas defended his PhD on March 10. Congratulations, Dr. Carstens!

Microsoft PhD Scholarship

Marc Deisenroth is a recipient of a Microsoft Research PhD Scholarship. The scholarship provides four-year funding for a PhD student working in the field of machine learning. More information can be found here.

Google Faculty Research Award

Marc Deisenroth, Lecturer in Statistical Machine Learning in the Department of Computing, Imperial College London, is a recipient of a prestigious Google Faculty Research Award for 2016. The one-year award supports the work of world-class, permanent faculty members at top universities around the world with the aim of advancing cutting-edge research in computer science, engineering… Read more »

Paper accepted at ESCAPE 26

Doniyor’s paper on Bayesian optimization for biological processes has been accepted at ESCAPE 26. Congratulations! Doniyor Ulmasov, Caroline Baroukh, Benoit Chachuat, Marc P. Deisenroth, Ruth Misener: Bayesian Optimization with Dimension Scheduling: Application to Biological Systems Proceedings of the European Symposium on Computer Aided Process Engineering, June 2016

New lab members

Hugh Salimbeni and Sanket Kamthe are joining the lab. Welcome!

SYSID paper on Learning Deep Dynamical Models published

Our paper on learning deep dynamical models from image pixels will be published at SYSID. Niklas Wahlström, Thomas B. Schön, Marc P. Deisenroth Learning Deep Dynamical Models From Image Pixels IFAC Symposium on System Identification 2015 Congratulations, Niklas!

AMAI paper on Bayesian optimization for locomotion accepted

Our paper got accepted at Annals in Mathematics an AI. Congratulations, Roberto! R Calandra, A Seyfarth, J Peters, MP Deisenroth: Bayesian Optimization for Learning Gaits under Uncertainty To appear in Annals of Mathematics and Artificial Intelligence

Distributed Gaussian Processes

Gaussian process regression for really large data sets. The model is conceptually simple, and it can be applied to distributed systems. Marc Peter Deisenroth, Jun Wei Ng Distributed Gaussian Processes arXiv pre-print Abstract: We propose the generalised Bayesian Committee Machine (gBMC), a practical and scalable hierarchical Gaussian process model for large-scale distributed non-parametric regression. The… Read more »

Learning from pixels to torques

How can an autonomous agent learn low-level closed-loop controllers from video streams? Check out our paper From Pixels to Torques: Policy Learning using Deep Dynamical Models for an idea.

Imperial College JRF awarded to Marc Deisenroth

Marc Deisenroth has been awarded an Imperial College Junior Research Fellowship on Robot Learning and Control from High-Dimensional Sensory Inputs with Application to Neurotechnology, sponsored by Aldo Faisal.

NIPS-2014 Workshop on Novel Trends and Applications in Reinforcement Learning accepted

Our NIPS Workshop on Novel Trends and Applications in Reinforcement Learning is accepted. The objective of the workshop is to provide a platform for researchers from various areas (e.g., deep learning, game theory, robotics, computational neuroscience, information theory, Bayesian modelling) to disseminate and exchange ideas, evaluating their advantages and caveats. More details are here:

Best Paper Award at ICRA 2014

Our paper Multi-Task Policy Search for Robotics received the Best Cognitive Robotics Paper Award at ICRA 2014. Marc P. Deisenroth, Peter Englert, Jan Peters, Dieter Fox Multi-Task Policy Search for Robotics Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2014

Mike Stilman

The shocking news this morning was that Mike Stilman (1981-2014), a young robotics professor at Georgia Tech who just got tenured, passed away. Official Georgia Tech Statement

Contribution for AISTATS-Highlights Track

Our TPAMI paper was accepted for a long presentation at the AISTATS-Highlights track. Marc P. Deisenroth, Dieter Fox, Carl E. Rasmussen Gaussian Processes for Data-Efficient Learning in Robotics and Control IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014

Paper accepted at ICASSP 2014

Sanket’s paper has been accepted at ICASSP 2014. Congratulations! Sanket Kamthe, Jan Peters, and Marc P. Deisenroth Multi-Modal Filtering for Non-linear Estimation International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2014

Paper accepted at AISTATS 2014

Nooshin’s paper has been accepted at AISTATS 2014. Congratulations! Nooshin HajiGhassemi and Marc P. Deisenroth Approximate Inference for Long-Term Forecasting with Periodic Gaussian Processes Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), 2014

Triple Success at ICRA 2014

We got three papers accepted at ICRA 2014 (acceptance rate 48%). Well done everybody! Bastian Bischoff, Duy Nguyen-Tuong, Herke van Hoof, Andrew McHutchon, Carl E. Rasmussen, Alois Knoll, Jan Peters, and Marc P. Deisenroth Policy Search For Learning Robot Control Using Sparse Data. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA… Read more »

Paper accepted at LION

Our paper on Bayesian Gait Optimization for Bipedal Locomotion has been accepted at the Learning and Intelligent Optimization (LION) conference. Most credit goes to Roberto Calandra. Roberto Calandra, Nakul Gopalan, Andre Seyfarth, Jan Peters, and Marc P. Deisenroth. Bayesian Gait Optimization for Bipedal Locomotion. In Proceedings of the Learning and Intelligent Optimization Conference (LION), 2014.

Lucas accepted at MLSS

Lucas got accepted at the Machine Learning Summer School (MLSS 2014) in Iceland. Congratulations!

Ben Taskar

The shocking news this morning was that Ben Taskar (born 1977) passed away. Statement by the University of Washington GeekWire article

Paper accepted at PAMI

Our paper “Gaussian Processes for Data-Efficient Learning in Robotics and Control” got accepted at IEEE Transactions on Pattern Analysis and Machine Intelligence (in the special issue on Bayesian Nonparametrics). The paper is available (Open Access).

Papers on before publication?

During ICML and EWRL 2013, I tried to collect some opinions on whether putting papers on (before publication) is a good or bad thing. As far as I know, there is no analysis of this topic, and it remains highly controversial. Here are some pros and cons: Pros: Availability: Research results (even premature) are… Read more »

Tony O’Hagan interviews Dennis Lindley

Tony O’Hagan interviews Dennis Lindley for the Royal Statistical Society’s Bayes 250 Conference held in June 2013. They discuss the Bayesian paradigm. Very interesting.

RL Tutorial at MLSS in Tunisia

I gave a tutorial on Reinforcement Learning at the Machine Learning Summer School on Big Data in Hammamet (Tunisia). Here are the slides

Survey on Policy Search

We published a book with a survey on policy search for robotics. Book’s website Download pdf

Delicious Twitter Digg this StumbleUpon Facebook