Two entries from the group into the 2014 Engineering Department Photo Competition are featured in The Telegraph.

The article features Yarin Gal’s second place prize-winning photo of an extrapolated image of Van Gogh’s starry night and David Duvenaud’s photo of a draw from a deep Gaussian process.

The article in the telegraph can be found at the following link:

http://www.telegraph.co.uk/technology/11228471/In-Pictures-Cambridge-University-Engineering-Photo-Competition-Winners.html?frame=3105262

The deadline for the Cambridge-Tuebingen PhD Fellowships is approaching!  The deadline is on the 2nd December 2014, and interviews will take place from 7th-8th January 2015.

For more information on the fellowships, which are supported by Facebook and Amazon, can be found here.

A new MPhil programme has been created jointly by the Machine Learning Group and the Speech Group at the Engineering Department at the University of Cambridge.

The MPhil (Masters of Philosophy) in Machine Learning, Speech, and Language Technologies is a 12 month full-time programme, which will see its first students starting in October 2015.

For more information on the programme, on applying, and for application/funding deadlines, please see the programme webpage: http://www.mlsalt.eng.cam.ac.uk/

Congratulations to David Knowles for winning this year’s Lindley Prize for his paper “Fixed-Form Variational Posterior Approximation through Stochastic Linear Regression”.  The prize is awarded every four years by the International Society for Bayesian Analysis (ISBA).  From the IBSA homepage: “Award winning papers will present research in Bayesian statistics that is is judged important, timely and notably original; truly innovative work will be judged more highly than successful development of ideas previously exposed.”

More on the Lindley Prize can be found here.

Sara Wade’s PhD Thesis “Bayesian nonparametric regression through mixture models” has been selected as a finalist for the Savage Award in the Theory & Methods category by the International Society for Bayesian Analysis (ISBA).  Four finalists were selected.  Bios of the finalists and the abstracts of their theses may be found in the ISBA bulletin here.

More info on the Savage Award may be found here.

Congratulations to David Knowles for being awarded the ISBA@NIPS travel award for the best invited Bayesian paper.

After a careful consideration by the ISBA Program Council the ISBA@NIPS Travel Award for best invited paper has been granted to Dr. David Knowles. In his young career, Dr. Knowles has provided important interdisciplinary contributions both in the Bayesian methodology and in the development of advanced computational methods. The ISBA Program Council greatly acknowledges his work and congratulates him with this award.

Ten new papers from the group are to appear at the 2014 conference on Advances in Neural Information Processing Systems (NIPS 2014), to be held in December in Montreal, Canada.

The list of papers are:

• Gaussian Process Volatility Model.  Yue Wu, Jose Miguel Hernandez-Lobato, and Zoubin Ghahramani.
• General Table Completion using a Bayesian Nonparametric Model.  Isabel Valera and Zoubin Ghahramani.
• Predictive Entropy Search for Efficient Global Optimization of Black-box Functions.  Jose Miguel Hernandez-Lobato, Matthew Hoffman, and Zoubin Ghahramani.  Selected for spotlight presentation.
• Tree-structured Gaussian Process Approximations.  Thang Bui and Richard Turner.  Selected for spotlight presentation.
• Variational Gaussian Process State-Space Models.  Roger Frigola, Yutian Chen, and Carl E. Rasmussen.
• Distributed Variational Inference in Sparse Gaussian Process Regression and Latent Variable Models.  Yarin Gal*, Mark van der Wilk*, Carl E. Rasmussen.  (* Joint first authors.)
• Gibbs-type Indian buffet processes.  Creighton Heaukulani and Daniel M. Roy.
• Mondrian Forests: Efficient Online Random Forests.  Balaji Lakshminarayanan, Daniel M. Roy, and Yee Whye Teh.
• Probabilistic ODE Solvers with Runge-Kutta Means.  Michael Schober, David Duvenaud, and Philipp Hennig.  Selected for oral presentation.
• A Filtering Approach to Stochastic Variational Inference.  Neil M. T. Houlsby and David Blei.

For the most up to date versions of the papers, visit the authors’ webpages, which may be found through our group members page.

We are delighted to announce that Facebook has donated $200,000 to support our unique Cambridge-Tuebingen PhD in Machine Learning. The Cambridge-Tuebingen PhD Programme is a joint programme in Machine Learning between the University of Cambridge and the Max Planck Institute for Intelligent Systems in Tuebingen. We received well over 100 high quality applications from around the world, interviewed about 12, and three students will be starting in the programme in October 2014: Shixiang (Shane) Gu studied Engineering Science at the University of Toronto, where he did his undergraduate thesis with Geoff Hinton on distributed training of neural networks. He then worked as an R&D Engineer at Panasonic Silicon Valley Lab before joining our programme. Adam Scibior has a BA in Computer Science from Cambridge and a BA and MSc in Physics from Posnan, and has worked on Monte Carlo methods and Probabilistic Programming. Mateo Rojas-Carulla received an MSc in Mathematical Engineering at Ecole Nationale des Ponts and a Distinction in Part III Mathematics at Cambridge, and has been working at Cantab Research. We are thrilled to welcome all three to our programme!

A first version of the Automatic Statistician is online at http://www.automaticstatistician.com.  The website describes the project, links to example automatic data analyses and has a simple demo where you can upload your own data sets.

Congratulations to Carl Rasmussen on his promotion to the senior academic post of Professor in the Department of Engineering!

The announcement in the reporter can be found here.