The 8th Annual Henry Taub TCE Conference | Deep Learning: Theory & Practice

  • Date/Time
    Date(s) - 14/06/2018
    All Day

    Kogan Auditorium, Meyer Building (EE)

    Categories No Categories



    The TCE center, jointly with the new Technion Machine Learning Center, will host this year the 8th annual Henry Taub International Conference

    Deep Learning: Theory & Practice

    June 14th, 2018

    ​Kogan Auditorium, Meyer Building (EE) Technion, Haifa


    The 8th annual TCE conference will focus on theory and practice in Deep Learning: Theory & Practice. Why is it working so well, and how can we improve it in various domains, such as vision, language, and audio.

    Registration open!

    In the past years, deep learning has been revolutionizing the field of machine learning and the high tech industry, leading to breakthrough results in various domains. Despite this impressive success, there are still major open questions, such as:

    • Modern deep learning models achieve "surprisingly good" performance in supervised learning tasks. However, it is not yet clear why, and how can we further improve these models.
    • How do we reduce the need for labeled data using techniques such as active learning and semi-supervised learning?
    • How can we make reinforcement learning work in real-life applications?
    • How do we effectively combine into our models all forms of prior knowledge, such as domain expertise, world knowledge, or information from other tasks (including "transfer learning" and "lifelong learning")?
    • How can we reliably estimate the uncertainty of the model and its predictions?
    • How to enable efficient hyper-parameter tuning and architecture selection?
    • How can we improve the resource efficiency and parallelization of existing models without degrading their accuracy?
    • In certain challenging tasks, such as text understanding, or multi-modal tasks, deep learning still has a relatively weak performance. How can we improve this?

    There is a growing need to develop useful theory and effective practical methods to facilitate scientific progress and help practitioners assimilate deep learning techniques in their applications. In this workshop, invited researchers and practitioners will present recent progress and challenges related to both theory and practice of deep learning.


    Conference Chairs:

    Daniel Soudry, EE Technion

    Ran El-Yaniv, CS Technion

    Confirmed speakers include:

    • Lior Wolf, Tel Aviv University, Israel
    • Michal Irani, Weizmann Institute of Science, Israel
    • Nathan Srebro, Toyota Technological Institute at Chicago, USA
    • Uri Shalit, Technion, Israel
    • Yoav Goldberg, Bar Ilan University, Israel
    • Zachary Chase Lipton, Carnegie Mellon University, USA