Speakers

Ali Eslami

Google Deepmind

Babak N. Araabi

University of Tehran

M. A. Sadeghi

University of Tehran

Mohammad Rastegari

Allen Institute for A.I.

Reza Abbasi-Asl

UC Berkeley

Hamed Pirsiavash

University of Maryland

M. R. Abolghasemi

IPM, UT

Ahmad Kalhor

University of Tehran

Reshad Hosseini

University of Tehran

Sajjad Torabian

Stanford, Luisville

Mohsen Moosavi

EPFL

Sajjad Mozaffari

University of Tehran

M. M. Derakhshani

University of Tehran

Yasser Souri

University of Bonn

Schedule

DAY 1
    • 1
      August
    • Introduction to Machine Learning

      Wednesday, 08:30 AM -10:00 AM
    • Babak N. Araabi

    + Intro to Workshop.

    + Supervised Vs Unsupervised Learning.

    + Classification Vs Regression.

    + How to Evaluate our model? (Train/Test split, Cross Validation)

    • 1
      August
    • Machine Learning Industry

      Wednesday, 10:00 AM -10:30 AM
    • Hamid Mahini (Tap30)
    • 1
      August
    • Introduction to Deep Learning

      Wednesday, 11:00 AM - 12:00 PM
    • Mohammad Amin Sadeghi

    + How perception and SVM work?

    + Multi-layer network and activation functions.

    + How gradient descent works?

    + What is Backpropagation?

    + A few examples.

    • 1
      August
    • Aplications of Deep Learning on Network Traffic Identification

      Wednesday, 12:00 PM -12:30 PM
    • Vahid Ghanbari (Sahab Pardaz)
    • 1
      August
    • Introduction to Keras framework

      Wednesday, 1:30 PM - 3:00 PM
    • Sajjad Mozaffari

    + Why Keras?

    + Installation Guide.

    + Keras Models.

    + Keras Preprocessing.

    + Keras Optimization.

    • 1
      August
    • Hands-on: Classification with a simple network

      Wednesday, 3:30 PM - 05:30 PM
    • Sajjad Mozaffari
DAY 2
    • 2
      August
    • Convolutional Neural Networks

      Thursday, 8:30 AM - 9:15 AM
    • Ahmad Kalhor

    + What is CNN?

    + Image classification examples.

    • 2
      August
    • Recurrent Neural Networks

      Thursday, 9:15 AM - 10:00 AM
    • Reshad Hosseini

    + What is RNN?

    + Gated RNNs

    + Time series prediction

    • 2
      August
    • Variational AutoEncoder

      Thursday, 10:00 AM - 10:30 AM
    • Yasser Souri (U of Bonn)
    • 2
      August
    • Deep Learning: from your fingertips to the center of your mind

      Thursday, 11:00 AM - 11:30 AM
    • Mohammad Rastegari (Allen Institute for A.I.)
    • 2
      August
    • Self-supervised learning for visual recognition

      Thursday, 11:30 AM - 12:30 PM
    • Hamed Pirsiavash (University of Maryland)
    • 2
      August
    • CNNs and RNNs in Keras

      Thursday, 1:30 PM - 3:00 PM
    • M. M. Derakhshani / S. Mozaffari

    + Why CNNs?

    + CNN layers (Pooling, Conv, Batch Norm, Dropout)

    + Data augmentation

    + Deep CNNs

    + Transfer learning

    + Recurrent layers

    + Sequence to Sequence Models

    • 2
      August
    • Hands-on: Classification with CNNs and RNNs

      Thursday, 3:30 PM - 5:00 PM
    • M. M. Derakhshani / S. Mozaffari
DAY 3
    • 3
      August
    • Generative Adversarial Networks

      Friday, 8:30 AM - 9:30 AM
    • Mohammad Amin Sadeghi

    + Generative vs discriminative models

    + Adversarial Training

    + Generative Adversarial Network

    • 3
      August
    • Toward Highly Interpretable Convolutional Neural Networks: Stability and Compression

      Friday, 9:30 AM - 10:00 AM
    • Reza Abbasi-Asl (UC Berkeley)
    • 3
      August
    • Multivariate Pattern Analysis

      Friday, 10:00 AM - 10:30 AM
    • Sajjad Torabian (Stanford, Luisville)
    • 3
      August
    • Neural Sence Representation and Rendering

      Friday, 11:00 AM - 12:00 PM
    • Ali Eslami (Google Deepmind)
    • 3
      August
    • Geometric robustness of deep networks

      Friday, 12:00 PM - 12:30 PM
    • Mohsen Moosavi (EPFL)
    • 3
      August
    • Deep Neural Networks and Neuroscience-Inspired Computer Vision

      Friday, 1:30 PM - 2:00 PM
    • Mohammad Reza Abolghasemi (UT)
    • 3
      August
    • GANs in Keras

      Friday, 2:00 PM - 3:00 PM
    • Mohammad Mahdi Derakhshani

    What is the usage of GANs?

    Gan Training Paradigm in Keras

    Upsamling vs Deconvolution Layers

    • 3
      August
    • Hands-on: Image Generation with GANs

      Friday, 3:30 PM - 5:30 PM
    • Mohammad Mahdi Derakhshani

ABOUT DEEP LEARNING SUMMER SCHOOL

Deep neural networks can learn multiple layers of representation at different levels of abstraction. Over the past decade, deep learning has dramatically improved the state-of-the-art in multiple areas of machine learning. UTDLSS'2018 will cover both the foundations and applications of deep neural networks.

At UTDLSS, several research experts will cover fundamentals of Machine Learning and Deep Learning. We will have hands-on sessions so that participants could have first hand experience with developing deep learning models in Python. Finally, we will have several speakers discussing their latest Deep Learning research results from global research centers including DeepMind, Berkeley, Stanford, UIUC, and EPFL.

UTDLSS is aimed at industry professionals, researchers, graduate students, and senior undergraduate students who want to develop practical insight into deep learning. UTDLSS will offer attendance certificate to participants.

Participants will need to bring their laptops in order to participate in hands-on sessions. They will need to install Anaconda and Keras (instructions will be given). We will try to offer electricity outlets but it is strongly advised to fully charge your laptop beforehand. There will be 2 hours of hands-on training each day.

Address


School of Electrical and Computer Engineering, University College of Engineering

University of Tehran, North Kargar st.

Tehran, Iran


        UTDeepLearning2018        mderakhshani@ut.ac.ir


Sponsors


َACM University of Tehran Iran Knowledge Tap30 Sahab Pardaz Co