Joan Bruna Contact Information
+1 917 767 0617 [email protected]
Invariant Signal Representations, Pattern recognition, Harmonic analysis, Stochastic Processes, Machine Learning, Deep Learning.
Ecole Polytechnique, Palaiseau, France Ph.D. in Applied Mathematics, “Scattering Representations for Recognition”, 20082013. • Thesis Topics: Invariant signal representations, classification, pattern recognition, stochastic processes, invariance learning, differential geometry. • Adviser: Professor St´ephane Mallat. Ecole Normale Superieure, Cachan, France MSc “Math´ ematiques, Vision, Apprentissage” in Applied Mathematics, 20042005. • Image, audio and video processing, Harmonic analysis, Machine Learning, Wavelet theory. • Mention Tr`es bien. Universitat Polit` ecnica de Catalunya, Barcelona, Spain BS, MSc Telecommunications, 1999-2004, • Master Thesis developed in Nokia Denmark (Aalborg), 2003-2004. • Title: “New Modeling techniques for HSDPA/WCDMA” (A with honors). BS Mathematics, 1998-2002. • With Honors, Ranked 2nd.
UC Berkeley, Berkeley, CA Assistant Professor, Dept of Statistics,
Facebook AI Research, New York, NY Post Doctoral Fellow, AI Research,
Courant Institute, NYU, New York, NY Postdoctoral Researcher Oct 2012 - Sep 2014 • Invariance Learning, Deep learning architectures, scattering representations of stochastic processes • Applications to Speech recognition, Pattern recognition, Texture synthesis, Bioacustics. • Supervisor: Professor Yann LeCun. CSR (through acquisition), Malakoff, France Senior Reseach Consultant Oct 2010 - July 2012 • Conception and supervision of real-time algorithms for video processing applications: frame-rate conversion, 2d to 3d conversion. • Machine Learning, Statistical Model Selection, Non-supervised learning for robust motion estimation.
1 of 4
Zoran (through acquisition) , Malakoff, France Senior Reseach Engineer June 2008 - Sep 2010 • Conception, development and implementation of real-time algorithms for several video processing applications: frame-rate conversion, 2d to 3d conversion, denoising. • Geometric Spatio-temporal representations, Sparse evolutive models, multiscale representations, real-time algorithm design. Let it Wave, Palaiseau and Malakoff, France. Research Engineer April 2005 to June 2008 • Algorithmic research on Video Super-Resolution, Denoising, and Deinterlacing • Conception and development of real-time video up-converter based on spatio-temporal geometric processing.
Ranzato,M.A., Szlam, A., Bruna, J., Mathieu, M., Collobert, R. and Chopra, S., “Video (language) modeling: a baseline for generative models of natural videos”, 2015. Sukhbaatar, S., Bruna, J., Paluri, M. Bourdev, L. Fergus, R, “Training Convolutional Networks with Noisy Labels”, ICLR Worshop, 2015. Bruna, J, Chintala, S., Piantino, S., Szlam, A. and Tygert, M, “A theoretical argument for complex-valued convolutional networks”, submitted, 2015. Goroshin, R, Bruna, J., Thomson, J. Eigen, D and Lecun, Y, “Unsupervised Learning of Spatiotemporally Coherent Metrics”, ICLR Workshop, 2015. Bruna, J., Sprechmann, P., LeCun, Y, “Audio Source Separation with NMF Scattering”, ICASSP, 2015. Denton, E. Zaremba, W. Bruna, J, Fergus, R. LeCun, Y, “Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation”, NIPS, 2014. Krishnan, D., Bruna, J., and Fergus, R., “Blind Deconvolution with Re-weighted Sparsity Promotion” , 2013. Bruna, J., Mallat, S., “Audio Texture Synthesis with Scattering Moments”, preprint 2013. Bruna, J., Szlam, A., LeCun, Y. “Signal Recovery from `p Pooling Representations”, ICML, 2014. Szegedy, C. Zaremba, W., Sutskever, I. Bruna, J, Erhan, D, Goodfellow, I, Fergus, R., “Intriguing Properties of Neural Networks”, ICLR, 2014. Bruna, J., Zaremba, W., Szlam, A and LeCun, Y, “Spectral Networks and Locally connected networks on Graphs”, ICLR, Jan 2014. Bruna, J., Mallat, S., Muzy, J-F and Bacry, E., “Intermittent Process Analysis with Scattering Moments”, Annals of Statistics, 2013. Bruna, J., Szlam, A., LeCun, Y. “Learning Stable Group Invariant Representations with Convolutional Networks”, ICLR, 2013. Bruna, J., Mallat, S. “Invariant Scattering Convolution networks” , IEEE trans of PAMI, 2012. Bruna, J., Mallat, S., “Classification with Scattering Operators”, CVPR, 2011. Bruna, J., Mallat, S. “Classification with Invariant Scattering Representation”, IVMSP, 2011. Bruna, J., Mallat, S., “Geometric Models with Co-occurrence Groups”, ESANN, 2010, Brugge. Bruna, J., Mallat, S., “Super-Resolution Bandlet Upconversion for HDTV”, SMPTE , Los Angeles, 2006.
2 of 4
EP10703636: Video sequence analysis for robust motion estimation. EP10703637: Temporal video interpolation method with 2-frame occlusion handling. US 20110096227: Frame Rate Conversion with Motion estimation in a plurality of resolution levels. US 20110058106, with M.Shand: Sparse Geometry for super-resolution video processing. EP2010/068133: Cadence Detection for Interlaced Video Based on Temporal Regularity. US 20100104210, with S.Mallat, N.Laveau, C.Bernard: Method and apparatus for spatiotemporal subband video enhancement with small time-delay. US 20090278981, with S.Mallat: Method and apparatus for robust super-resolution video scaling.
Awards • ENS Cachan Scolarship, 2004-2005 • Honors Mention in Spanish Physics Olympiad, 1998 • 2nd national place in Cangur Mathematical Contest, 1996
Collegial • NIPS reviewer 2014 • EECV reviewer 2014 • JMLR reviewer 2014 • CVPR reviewer 2014 • PNAS reviewer 2013
Talks and Seminars
• Google Apr 2014 • Duke University Mar 2014 • Max-Plank Institute, Tubingen, Germany, Mar 2014. • UC Berkeley, Feb 2014 • Yale, Nov 2013. • IBM Watson Research, NY, Oct 2013. • Duke University NC, Dept of Electrical Engineering, Oct 2013. • UC Berkeley CA, Dept of Statistics, Aug 2013. Scattering Seminar. • Yale University CT, Applied Mathematics Seminar, Jan 2013. • Princeton University NJ, Applied Mathematics Seminar, Nov 2012. • Invited Talk at the Royal Society Meeting, Chicheley Hall, march 2012, Product Representations of Stationary Processes. • Seminar on Classification with Scattering Operators, CSIRO, Perth, 2011. • Talk on Scattering Operators, MSC-MOA, Toulon, 2011. • Invited talk in opening ceremony of academic year of CFIS, UPC, Barcelona, 2010. • “Super-Resolution Bandlet Upconversion for HDTV”, SMPTE conference, Los Angeles, 2006. • Spring School of combinatorics, presenting “A new overlapping theorem and applications to the Kakeya Conjecture”, Borova Lada, Czech Republic, 2003.
3 of 4
Technical Skills • Operating systems: OS X, Unix/Linux. • Programming Languages: C, Yorick, Matlab, VHDL, Java. • Embedded systems: MIPS.
Language Skills • Catalan: Mother tongue • Spanish: Bilingual • French: Bilingual • English: Bilingual • Japanese: Beginner level (2 years)
Travelling, Food, Photography, Literature, New Wave music.
4 of 4