Technical Program
A preliminary version of the technical program is now available (see also program and contributions abstracts for a pdf version). The workshop videos are also available now.Thursday, 4 September 2014
08.00-08.45: Registration
Roberts Foyer, UCL Roberts Engineering Building
08.45-09.00: Welcome (video)Roberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Miguel Rodrigues and John Shawe-Taylor (UCL)Robert Calderbank and Lawrence Carin (Duke U.)
09.00-10.30: Invited Talks I
Roberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Chair: David Dunson, Duke U.Conjugate gradient iterative hard thresholding for compressed sensing and matrix completion (video, slides)
Jared Tanner, University of Oxford
Breaking the coherence barrier - A new theory for compressed sensing (video, slides)
Anders Hansen, University of Cambridge
Optimal compressive imaging for Fourier data (video, slides)
Gitta Kutyniok, Technical University of Berlin
10.30-11.00: Coffee Break
Roberts Foyer, UCL Roberts Engineering Building
11.00-13.00: Invited Talks IIRoberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Chair: Petros Dellaportas, AUEBVisual pattern encoding on the Poincaré sphere (video, slides)
Aleksandra Pizurica, Ghent University
Tracking dynamic point processes on networks (video, slides)
Rebecca Willett, University of Wisconsin-Madison
Deep Gaussian processes (video, slides)
Neil Lawrence, Sheffield University
Mondrian forests: Efficient random forests for streaming data via
Bayesian nonparametrics (video, slides)
Yee Whye Teh, University of Oxford
12.30-14.30: Lunch Break
14.30-15.30: Whiteboard Session I
Roberts Foyer, UCL Roberts Engineering Building
15.30-16.00: Coffee BreakRoberts Foyer, UCL Roberts Engineering Building
16.00-17.30: Poster SessionRoberts Foyer, UCL Roberts Engineering Building
17.30-18.30: Keynote LectureRoberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Chair: Guillermo Sapiro, Duke U.The Unreasonable Effectiveness of Deep Learning (video, slides)
Yann LeCun, Facebook and New York University
19.30-21.30: Workshop Dinner
South Cloisters, UCL
Friday, 5 September 2014
08.00-09.00: Registration
Roberts Foyer, UCL Roberts Engineering Building
09.00-10.30: Invited Talks IIIChair: Arthur Gretton, UCL
NuMax: A convex approach for learning near-isometric linear embeddings
Richard Baraniuk, Rice University
Beyond stochastic gradient descent for large-scale machine learning (video, slides)
Francis Bach, INRIA
Living on the edge: Phase transitions in convex programs with random data (video, slides)
Joel Tropp, California Institute of Techology
10.30-11.00: Coffee Break
Roberts Foyer, UCL Roberts Engineering Building
11.00-12.30: Invited Talks IVRoberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Chair: Ingrid Daubechies, Duke U.Building an automatic statistician (video, slides)
Zoubin Ghahramani, University of Cambridge
Variable selection in high dimensional convex regression (slides)
John Lafferty, University of Chicago
High-dimensional learning with deep network contractions (video, slides)
Stéphane Mallat, Ecole Normale Superieure
12.30-14.30: Lunch Break
14.30-15.30: Whiteboard Session II
Roberts Foyer, UCL Roberts Engineering Building
15.30-16.00: Coffee BreakRoberts Foyer, UCL Roberts Engineering Building
16.00-17.30: Industry Session: Big Data - Challenges and Opportunities (video)Roberts G06 Sir Ambrose Fleming LT, UCL Roberts Engineering Building
Moderators: Robert Calderbank (Duke U.) and Patrick Wolfe (UCL)Panelists: Christophe Bernard (Winton Capital), Christoph Best (Google), Thore Graepel (Microsoft Research), Gabriel Hughes (Elsevier), Yann LeCun (Facebook)
Whiteboard Session I
Hard thresholding pursuit algorithms: The greedy wayJean-Luc Bouchot, Drexel University // RWTH Aachen University
Asymptotic independence of highly coupled very high dimensional data
Erol Gelenbe, Imperial College London
A new look at mean embeddings
Steffen Grunewalder, University College London
Breaking the coherence barrier - A new theory for compressed sensing
Anders Hansen, University of Cambridge
Compressed sensing with side information
João Mota, University College London
Visual pattern encoding on the Poincaré sphere
Aleksandra Pizurica, Ghent University
Stein shrinkage for cross-covariance operators and kernel independence testing
Aaditya Ramdas, Carnegie Mellon University
The distribution of restricted least squares with a Gaussion matrix (invited)
Galen Reeves, Duke University
Whiteboard Session II
Beyond stochastic gradient descent for large-scale machine learningFrancis Bach, INRIA
Designer Bayes factorizations: Applications to tensors & networks (invited)
David Dunson, Duke University
High-dimensional change-point detection with sparse alternatives (invited)
Farida Enikeeva, University of Poitiers
Bayesian models for social interactions (invited)
Katherine Heller, Duke University
Inference in high-dimensional varying coefficient models
Mladen Kolar, University of Chicago Booth School of Business
Damian Kozbur, ETH Zurich
Fast and robust multiscale methods for high-dimensional data (invited)
Mauro Magionni, Duke University
Kernel MMD, the median heuristic and distance correlation in high dimensions
Aaditya Ramdas, Carnegie Mellon University
Poster Session
Sparse inverse covariance estimation with hierarchical matricesJonas Ballani, EPFL
On the absence of the RIP in practical CS and the RIP in levels
Alexander Bastounis, University of Cambridge
Efficient inference for joint models of LPF and spiking data
David Carlson, Duke University
Shrinkage mappings and their induced penalty functions
Rick Chartrand, Los Alamos National Laboratory
Dictionary designs for compressive sensing and distributed compressive sensing
Wei Chen, University of Cambridge
Unlocking energy neutrality in energy harvesting wireless sensor networks: An approach based on distributed compressed sensing
Wei Chen, University of Cambridge
Deep networks with adapted Haar scattering
Xiuyuan Cheng, Ecole Normale Superieure
Mathematically grounded methods for analysing time series data on animal movement
Sarah Chisholm, University College London
Orthogonal matching pursuit (OMP) to reconstruct optical coherence tomography (OCT) image
Yue Dong, University of Liverpool
Refined analysis of sparse MIMO radar
Dominik Dorsch, RWTH Aachen University
Recovery of wavelet expansion from nonuniform Fourier samples via weighted iterative hard thresholding
Jonathan Fell, RWTH Aachen University
Sparsistent additive modeling in multi-task learning
Madalina Fiterau, Carnegie Mellon University
Mladen Kolar, University of Chicago Booth School of Business
Low-complexity compressive sensing detection for spatial modulation in large-scale multiple access channels
Adrian Garcia-Rodriguez, University College London
A multiscale approach to discrete optimal transport
Sam Gerber, Duke University
Multichannel adaptive filtering in compressive domains
Karim Helwani, Huawei European Research Center
Modulator design for binary classification of Poisson measurements
Jiaji Huang, Duke University
Robert Calderbank, Duke University
Analyzing the structure of multidimensional compressed sensing problems through local coherence
Alex Jones, University of Cambridge
Robust uniform recovery of low-rank matrices from Gaussian measurements
Maryia Kabanava, RWTH Aachen University
Matrix completion on graphs
Vassilis Kalofolias, EPFL
Tensor low-rank and sparse light field photography
Mahdad Hosseini Kamal, EPFL
Coherence and sufficient sampling densities for reconstruction in compressed sensing
Franz Kiraly, University College London
Learning with cross-kernels and ideal PCA
Franz Kiraly, University College London
Modeling correlated arrival events with latent semi-Markov processes
Wenzhao Lian, Duke University
MUSIC for single-snapshot spectral estimation: Stability and super-resolution
Wenjing Liao, Duke University
Albert Fannjiang, University of California, Davis
Terahertz imaging via block based compressive sensing
Lin Liu, University of Liverpool
Sparse recovery conditions and realistic forward modeling in EEG/MEG source reconstruction
Felix Lucka, University of Munster
Fast and robust multiscale dictionary learning
Mauro Maggioni, Duke University
Distributed compressed sensing algorithms: Completing the Puzzle
João Mota, University College London
A unified algorithmic approach to distributed optimization
João Mota, University College London
Learning from negative examples for machine translation
Tsuyoshi Okita, Dublin City University
Finite dimensional FRI for reconstruction of sparse signals
Jon Onativia, Imperial College London
Pier Luigi Dragotti, Imperial College London
Supervised learning on an unsupervised atlas
Nikolaos Pitelis, University College London
Compressive classification of a mixture of Gaussians: Analysis, designs and applications
Hugo Reboredo, University of Porto--Instituto de Telecomunicações
Reconstruction of high-dimensional GMM data from low-dimensional features
Francesco Renna, University of Porto-Instituto de Telecomunicações
Classification of high-dimensional data from low-dimensional features in the presence of side information
Francesco Renna, University of Porto-Instituto de Telecomunicações
Order statistics of exponential random variables with imperfect measurement and unknown Gaussian disturbance for resource allocation compression models
Ramiro Samano Robles, Instituto de Telecomunicações (IT)/ Research Centre of Real Time and Embedded Computer Systems (CISTER)
On asymptotic sparsity in compressed sensing
Bogdan Roman, University of Cambridge
Variational Bayesian inference for sparse matrix factorization
Evangelos Roussos, University of Oxford
Sparse estimation with generalized Beta mixture and the Horseshoe prior
Zahra Sabetsarvestani, Amirkabir University of Technology
Portfolio optimization via manifold learning
Alireza Samani, Duke University
Adaptive MCMC with kernel embeddings
Dino Sejdinovic, Gatsby Unit, University College London
Learning features for classification
Jure Sokolic, University College London
Classification of signals with mismatched MAP classifier
Jure Sokolic, University College London
Achieving compressed sensing physical system via random demodulation
Pingfan Song, Harbin Institute of Technology
Low-rank tensor recovery via Theta bodies
Zeljka Stojanac, University of Bonn
Simple consistent distribution regression on compact metric domains
Zoltan Szabo, Gatsby Unit, University College London
Analysis of brain states from multi-region LFP time-series
Kyle Ulrich, Duke University
Nonlinear information-theoretic compressive measurement design
Liming Wang, Duke University
Semi-deterministic sensing matrices by partial randomly phase modulated unit-norm tight frames
Peng Zhang, Imperial College London
Compressed sensing non-uniformly sparse signals: An asymptotically optimal power allocation
Xiaochen Zhao, Imperial College London
Wei Dai, Imperial College London
Block-structured sparse tensor decomposition for classification of multi-dimensional data
Syed Zubair , University of Surrey
Wenwu Wang, University of Surrey
Jonathon Chambers, Loughborough University