Course Identification

Practical Deep Learning for Science

Lecturers and Teaching Assistants

Prof. Eilam Gross
Dr. Jonathan Shlomi, Dr. Michael Pitt

Course Schedule and Location

Second Semester
Wednesday, 13:15 - 16:00, Weissman, Auditorium

Field of Study, Course Type and Credit Points

Physical Sciences: Lecture; Elective; Regular; 4.00 points
Chemical Sciences: Seminar; Elective; 4.00 points







Language of Instruction


Attendance and participation


Grade Type

Numerical (out of 100)

Grade Breakdown (in %)


Evaluation Type

Final assignment

Scheduled date 1


Estimated Weekly Independent Workload (in hours)



 This course is aimed at students who have no experience with Machine Learning and would like to understand what it is about and be able to apply Neural Network in their research.

 At the end of the course you should be able to apply Deep Learning to your own research based on worked out examples. 

Topics include:

  1. Convolutional Neural Networks (CNN) 

  2. RNNs+Graphs

  3. Attention Models

  4. Generative Models and Advresarial Training

  5. AutoEncoders Anomaly Detection .

Each lecture will be divided into two parts:

  1. Some overview of the subject

  2. Hand on training

This is not a theoretical course of Machine Learning, its more like a hands-on course taught by users of machine learning.

The course is inspired and based on the fastAI web course:

The NN online book:

with some inspiration and touch of DEEP LEARNING by Goodfellow, Bengio and Courville.

There is no point in taking the course if you do not aim to exercise. To that end you are asked to open a cloud account BEFORE coming to the first lecture. 

There are a few possibilities.

For the first lecture you can open a google cloud account (recommended) or use Collab . To open a Google Cloud account please follow the instructions here:

Note,  when opening a cloud account with Google you get a credit of $300 which is at least 3 times more than the cost of the usage needed for the course.

Last but not least, the course assumes some basic familiarity with coding in python. It is advised to acquire some basic experience with python. For example you can follow .

   The course will be taught  by Eilam Gross with an extended teaching assistance by Sanmay Ganguly, Michael Pitt and Jonathan Shlomi.  


Learning Outcomes

Upon successful completion of the course the students will be able to:

  • Set up a deep learning based analysis
  • Get an idea what the buzz is all about

Reading List