Marketing

DS405B Practical Deep Learning with Visual Data

Lecturer: Dr. Aseem Behl
Course description:

DS405B

Language: English
Recommended for this semester or higher: 1
ECTS-Credits: 6
Course can be taken as part of following programs/modules: For more information please refer to alma.
Prerequisite for:

---

Prerequisites: Prior experience with programming in Python ideally through an introductory Python course like DS400 Data Science Project Management, DS404 Data Science with Python or DS405 Machine Learning Applications in Business and Economics
Limited attendance: 24
Course Type: Lecture
Date:

Wednesdays from 8am - 10am c.t. - room Hörsaal 05 Neue Aula
(Beginning of the first lecture April 17, 2024)

Registration: Limited to 24 participants. Registration open for all (no first-come, first-served): Registration will open on February 1, 2024 in ALMA - end of registration time: April 14, 2024. If the number of applications (limited to 24 participants) exceeds the number of places available, we will randomly select from all applications. Preferred access for students from the M.Sc. Data Science in Business and Economics. Notice of admission or non-admission on Monday, April 15 by email.
Downloads: ILIAS
Method of Assessment: Written Exam or Presentation or Assignments or Term Paper
Content: Deep learning has become widely successful in tackling fundamental tasks arising in computer vision, language processing and robotics. Many of these fundamental tasks are relevant to a much broader array of applications in business and scientific domains. This module starts with a broad view of machine learning and neural networks, and it subsequently covers the theory of neural networks in the context of practical examples and implementation of deep learning methods with the help of prominent frameworks in Python. The focus will be on applications deriving business intelligence from visual data, however, several concepts learned in the module can be applied to other data modalities like tabular or textual data.
Objectives: After this module, students can develop an understanding of how neural network models work and how to implement neural network architectures in Python with the help of deep learning frameworks. They can exploit image datasets to reliably train and debug modern deep learning techniques for applications in business and economics. They can appreciate the effectiveness of deep learning as a tool in their machine learning toolbox.
Literature: There is no required textbook for this module. Some lectures may recommend readings from the following books: 
1.    Neural Networks and Deep Learning by Michael Nielsen
2.    Deep Learning by Ian Goodfellow and Yoshua Bengio and Aaron Courville
3.    Dive into Deep Learning by Aston Zhang, Zachary C. Lipton, Mu Li, Alexander J. Smola