Lectures are in room 1140 (ground floor) of Pavillon Aisenstadt (2920 chemin de la tour). First day of lectures: Monday August 3 Last day of lectures: Wednesday August 12 There will be lectures during the week-end, i.e., a total of 10 days of lectures. The planned daily schedule is as follows: 8:30-9:00 refreshments 9:00-10:30 lecture 1 10:30-11:00 coffee break 11:00-12:30 lecture 2 12:30-2:00 lunch break 2:00-3:30 lecture 3 3:30-4:00 coffee break 4:00-5:30 posters, deep learning programming tutorials, exercises, free time Day 1, Mon, Aug 3 Pascal Vincent: Intro to ML (9:00-10:30) Yoshua Bengio: Theoretical motivations for Representation Learning & Deep Learning (11:00-12:30) Leon Bottou: Intro to multi-layer nets (2:00-3:30) Posters Aug 3 (4:00-5:30): Guillaume Alain (Université de Montréal): “Distributed Dropout in Asynchronous Stochastic Gradient Descent” Amjad Almahairi (Université de Montréal): “Dynamic Knowledge Distillation” Christof Angermueller (University of Cambridge): “Multi-task Deep Neural Networks for predicting CpG methylation from high-dimensional biological data” Gunes Baydin (National University of Ireland Maynooth): “DiffSharp: Automatic Differentiation Library” Lucas Beyer (RWTH Aachen University): “Torch7 + Theano = Beacon8” Parminder Bhatia (Georgia Tech): “Sentiment Analysis of Short Text - Sentence & User Representation Learning” Yuri Burda (University of Toronto): “Improved training of variational autoencoders” Heeyoul Choi (Samsung Advanced Institute of Technology): “Improvements on LSTM” Pradeep Dasigi (Carnegie Mellon University): “Ontologically grounded selectional preference autoencoders for modeling events” Emily Denton (New York University): “Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks” Michael Eickenberg (Université de Paris-Sud): “Convolutional Nets Map the Architecture of the Human Visual System” Pablo Fonseca (University of Campinas): “A CNN approach to Breast Density Classification” Mercedes Garcia (Université du Maine): “Class-based Continuous Space Language Model” Ross Goroshin (New York University): “Learning to Linearize under Uncertainty” Day 2, Tue, Aug 4 Hugo Larochelle: Neural nets and backprop (9:00-10:30) Leon Bottou: Numerical optimization and SGD, Structured problems & reasoning (11:00-12:30) Hugo Larochelle: Directed Graphical Models and NADE (2:00-3:30) Intro to Theano (4:00-5:30) Day 3, Wed, Aug 5 Aaron Courville: Intro to undirected graphical models (9:00-10:30) Honglak Lee: Stacks of RBMs (11:00-12:30) Pascal Vincent: Denoising and contractive auto-encoders, manifold view (2:00-3:30) GPU Programming (4:00-5:30) Day 4, Th, Aug 6 Roland Memisevic: Visual features (9:00-10:30) Honglak Lee: Convolutional networks (11:00-12:30) Graham Taylor: Learning similarity (2:00-3:30) Posters Aug 6 (4:00-5:30): Shixiang Gu (University of Cambridge): “Neural Adaptive Sequential Monte Carlo” Philip Haeusser (Technische Universität München): “FlowNet: Learning Optical Flow with Convolutional Networks” Anna Huang (Harvard University): “ChordRipple: Radical Chord Recommendations with chord2vec” Jiwoong Im (Université de Montréal): “Conservativeness of untied auto-encoder” Vamsi Ithapu (University of Wisconsin Madison): "Regularizing Deep networks via Matrix Completion” Rico Jonschkowski (Technische Universität): “Learning State Representations with Robotic Priors” Mikael Kageback (Chalmers University of Technology): “Neural context embeddings for automatic discovery of word senses” Kazuya Kawakami (Carnegie Mellon University): “Learning to Representing Words in Context from Parallel Data for Improved Supersense Tagging” Suyoun Kim (Carnegie Mellon University): “Environmental Noise Embeddings for Robust Speech Recognition” Kishore Reddy Konda (Goethe University Frankfurt): “Dropout as data augmentation” Iryna Korshunova (Ghent University): “Epileptic Seizure Prediction using Convolutional Neural Networks” David Krueger (Université de Montréal): “Speech Synthesis with Conditional Grammar Cells” Angeliki Lazaridou (University of Trento): “Unveiling the Dreams of Word Embeddings: Towards Language-Driven Image Generation” Kisuk Lee (Massachusetts Institute of Technology): “Recursive 2D-3D Convolutional Networks for Neuronal Boundary Detection” Conrado Miranda (University of Campinas): “Multi-Objective Optimization for Self-Adjusting Weighted Gradient in Machine Learning Tasks” Day 5, Fri, Aug 7 Chris Manning: NLP 101 (9:00-10:30) Graham Taylor: Modeling human motion, pose estimation and tracking (11:00-12:30) Chris Manning: NLP / Deep Learning (2:00-3:30) Theano programming, datasets with Fuel (4:00-5:30) Day 6, Sat, Aug 8 Ruslan Salakhutdinov: Deep Boltzmann Machines (9:00-10:30) Adam Coates: Speech recognition with deep learning (11:00-12:30) Ruslan Salakhutdinov: Multi-modal models (2:00-3:30) Posters Aug 8 (4:00-5:30): Olof Mogren (Chalmers University of Technology): “Extractive Summarization by Aggregating Multiple Similarities” Tsendsuren Munkhdalai (Chungbuk National University): “DeepText: End-to-end biomedical event extraction via deep learning and recursive projection model” Natalia Neverova (INSA-Lyon / University of Guelph): “Hand pose estimation by deep transductive learning” Minh Nguyen (Rice University): “A Probabilistic Theory of Deep Learning” Makoto Otsuka (Castalia Co., Ltd.): “Dynamic Boltzmann machines with homeostatic spike-timing dependent plasticity” Lionel Pigou (Ghent University): “Beyond Temporal Pooling: Recurrence and Temporal Convolutions for Gesture Recognition in Video” Dan Rosenbaum (The Hebrew University): “The Return of the Gating Network: generative models and discriminative training in image priors” Levent Sagun (New York University): “Explorations on high dimensional landscapes” Joao Felipe Santos (Institut National de la Recherche Scientifique): “Using convolutional and recurrent layers in deep neural networks for spectral speech enhancement” Irina Sergienya (University of Munich): “Learning Better Embeddings for Rare Words Using Distributional Representations” Soeren Kaae Soenderby (University of Copenhagen): “Transformer Networks for Classifying Sequences” Ruben Villegas (University of Michigan): “Improving Object Detection with Deep Convolutional Networks via Bayesian Optimization and Structured” Weizhong Yan (GE Global Research Center): “Deep Learning for Industrial Asset Condition Monitoring” Wenpeng Yin (University of Munich): “Learning Word Meta-Embeddings by Combining Diverse Embedding Versions” Asja Fischer (Université de Montréal): “Difference Target-Propagation” Day 7, Sun, Aug 9 Ian Goodfellow: Structure of optimization problems (9:00-10:30) Adam Coates: Systems issues and distributed training (11:00-12:30) Ian Goodfellow: Adversarial examples (2:00-3:30) Day 8, Mon, Aug 10 Phil Blunsom: From language modeling to machine translation (9:00-10:30) Richard Socher: Recurrent neural networks (11:00-12:30) Phil Blunsom: Memory, Reading, and Comprehension (2:00-3:30) Theano debugging tools, conv. nets (4:00-5:30) Day 9, Tue, Aug 11 Richard Socher: DMN for NLP (9:00-10:30) Mark Schmidt: Smooth, Finite, and Convex Optimization (11:00-12:30) Roland Memisevic: Visual Features II (2:00-3:30) Theano implementation of iterations (scan), RNNs (4:00-5:30) Day 10, Wed, Aug 12 Mark Schmidt: Non-Smooth, Non-Finite, and Non-Convex Optimization (9:00-10:30) Aaron Courville: VAEs and deep generative models for vision (11:00-12:30) Yoshua Bengio: Generative models from auto-encoders (2:00-3:30) Theano programming: overflow session (4:00-5:00) Farewell party (5:00-) registration form |