Jörg Bornschein

Jörg Bornschein

eMail: jb@capsec.org
GitHub: jbornschein
My Google Scholar page

I'm a CIFAR Global Scholar in the program for
Neural Computation and Adaptive Perception.

I work as a Postdoctoral researcher in Yoshua Bengios
machine learning group at the University of Montreal.

Research Interests

Unsupervised Learning; Efficient large scale / parallel machine learning;
Computational Neuroscience

Research Projects


Peer Reviewed Publications

Journal Papers, Conference Papers and Book Chapters

  1. Training Bidirectional Helmholtz Machines
    Bornschein, Jörg and Shabanian, Samira and Fischer, Asja and Bengio, Yoshua
    In Review
  2. Reweighted Wake-Sleep
    Bornschein, Jörg and Bengio, Yoshua
    International Conference on Learning Representations (ICLR 2015)
  3. Nonlinear Spike-And-Slab Sparse Coding for Interpretable Image Encoding
    Shelton, Jacquelyn A. and Sheikh, Abdul-Saboor and Bornschein, Jörg and Sterne, Philip and Lücke, Jörg
    PLoS ONE
  4. Are V1 Simple Cells Optimized for Visual Occlusions? A Comparative Study
    Bornschein, Jörg and Henniges, Marc and Lücke, Jörg
    PLoS Computational Biology
  5. Why MCA? Nonlinear Sparse Coding with Spike-and-Slab Prior for Neurally Plausible Image Encoding
    Sterne, Philip and Bornschein, Joerg and Sheikh, Abdul-saboor and Luecke, Joerg and Shelton, Jacquelyn A
    Advances in Neural Information Processing Systems (NIPS 2012)
  6. Select and Sample - a Model of Efficient Neural Inference and Learning
    Shelton, Jacquelyn A and Sheikh, Abdul S and Berkes, Pietro and Bornschein, Joerg and Lücke, Jörg
    Advances in Neural Information Processing Systems (NIPS 2011)
  7. A Gabor Wavelet Pyramid-based Object Detection Algorithm
    Sato, Yasuomi D and Jitsev, Jenia and Bornschein, Joerg and Pamplona, Daniela and Keck, Christian and von der Malsburg, Christoph
    Advances in Neural Networks (ISNN 2011)
  8. FPGA Coprocessor for Simulation of Neural Networks Using Compressed Matrix Storage
    Bornschein, Jörg
    System and Circuit Design for Biologically-Inspired Intelligent Learning
  9. The Maximal Causes of Natural Scenes are Edge Filters
    Puertas, Jose and Bornschein, Jörg and Lücke, Jörg
    Advances in Neural Information Processing Systems (NIPS 2010)
  10. Binary Sparse Coding
    Henniges, Marc and Puertas, Gervasio and Bornschein, Jörg and Eggert, Julian and Lücke, Jörg
    Latent Variable Analysis and Signal Separation

Conference Abstracts and Workshop Papers

  1. Towards Biologically Plausible Deep Learning
    Bengio, Yoshua and Lee, Dong-Hyun and Bornschein, Jörg and Lin, Zhouhan
    arXiv preprint arXiv:1502.04156
  2. Combining Feed-Forward Processing and Sampling for Neurally Plausible Encoding Models.
    Lücke, J. and Shelton, J. A. and Bornschein, J. and Sterne, P. and Berkes, P. and Sheikh, A.-S.
    Proc. Computational and Systems Neuroscience (COSYNE)
  3. Combining Approximate Inference Methods for Efficient Learning on Large Computer Clusters
    Dai, Zhenwen and Shelton, Jacquelyn A. and Bornschein, Jörg and Shelton, Abdul S. and Lücke, Jörg
    NIPS workshop Big Learning: Algorithms, Systems, and Tools for Learning at Scale, Sierra Nevada, Spain
  4. The Maximal Causes of Binary Data
    Bornschein, Jörg and Shelton, Jacquelyn A. and Sheikh, Abdul S. and Lücke, Jörg
    Front. Comput. Nerosci: Bernstein Conference on Computational Neuroscience
  5. Sparse Codes of V1 Simple-Cells and the Emergence of Globular Receptive Fields - a Comparative study
    Bornschein, Jörg and Henniges, Marc and Puertas, Gervasioi and Lücke, Jörg
    Proc. of Computational and Systems Neuroscience (COSYNE)
  6. Approximate EM Learning on Large Computer Clusters
    Bornschein, Jörg and Dai, Zhenwhen and Lücke, Jörg
    NIPS Workshop: Learning on Cores, Clusters and Clouds
  7. Binary Hidden Variables and Sparse Sensory Coding
    Bornschein, Jörg and Henniges, Marc and Puertas, Gervasio and Lücke, Jörg
    Front. Comput. Nerosci: Bernstein Conference on Computational Neuroscience
  8. Applications of Non-linear Component Extraction to Spectrogram Representations of Auditory Data
    Bornschein, Jörg and Lücke, Jörg
    Front. Comput. Nerosci: Bernstein Conference on Computational Neuroscience
  9. Movement Planning, Motor Control and Transfer Across Movement Categories
    Otte, E. and Bornschein, J.
    5th. Intl. Conference on Development and Learning (ICDL)