Computational Learning TheoryModule description #41014 / Version 1
- 3 CP
- Portfolio examination
- Graded
- Winter semester
- English
This module description is not valid in the current semester (Winter semester 2026/27).
| Module / Version | #41014 / #1 |
|---|---|
| Validity | Winter semester 2020/21 – Summer semester 2021 |
| Available language(s) | English |
| Credit points | 3 |
| Person responsible for the module | Opper, Manfred |
| Grading | Graded |
| Exam format | Portfolio examination |
| Language(s) of instruction | English |
Affiliation
| Faculty | Fakultät IV |
|---|---|
| Institute | Institut für Softwaretechnik und Theoretische Informatik |
| Academic chair | 34351600 FG Künstliche Intelligenz |
Contact
| Office | MAR 4-2 |
|---|---|
| Contact person | Lippke, Cordula |
| E-mail address | manfred.opper@tu-berlin.de |
| Website | http://www.ki.tu-berlin.de/menue/lehre |
Learning outcomes
Students have a deep and profound understanding of the theoretical and especially statistical background of machine learning. This is achieved by the independent study of a current research topic in the field including rigorous mathematical derivations using literature and by the independent preparation of a seminar presentation.
Content
The seminar deals with the theory of machine learning from a (mathematically)
rigorous statistical perspective. The so-called PAC (probably almost correct)
learning framework focusses on the question how well a machine learning algorithm which was trained on random independent data can make predictions on new instances which are generated from the same but unknown probability distribution.
The seminar will be based on the recent textbook “Understanding Machine Learning-From theory to Algorithms” by Shai Shalev-Schwartz and Shai Ben-David.
After explaining the statistical learning framework, the seminar will discuss uniform convergence results and the VC dimension. It will also discuss theoretical aspects of different machine learning models, such as convex learning, support vector machines, boosting, online learning and stochastic gradient descent. It will cover regularisation, model selection and dimensionality reduction. Finally, also more advanced topics such as
Rademacher complexities will be included.
rigorous statistical perspective. The so-called PAC (probably almost correct)
learning framework focusses on the question how well a machine learning algorithm which was trained on random independent data can make predictions on new instances which are generated from the same but unknown probability distribution.
The seminar will be based on the recent textbook “Understanding Machine Learning-From theory to Algorithms” by Shai Shalev-Schwartz and Shai Ben-David.
After explaining the statistical learning framework, the seminar will discuss uniform convergence results and the VC dimension. It will also discuss theoretical aspects of different machine learning models, such as convex learning, support vector machines, boosting, online learning and stochastic gradient descent. It will cover regularisation, model selection and dimensionality reduction. Finally, also more advanced topics such as
Rademacher complexities will be included.
Module components
Compulsory area
The following courses are mandatory for this module:
| Courses | Type | Number | Frequency | Language | Hours per week | Course catalog |
|---|---|---|---|---|---|---|
| Computational Learning Theory | Seminar | 3435 L 10792 | Winter semester | [en] | 2 | ISIS |
Workload and credit points
Computational Learning Theory (Seminar)
| Workload description | Multiplier | Hours | Total |
|---|---|---|---|
| Attendance | 15 | 2 h | 30 h |
| Pre/post processing | 15 | 4 h | 60 h |
| 90 h | |||
The workload of the module adds up to 90 hours. The module therefore comprises 3 credit points.
Description of teaching and learning methods
The seminar includes skill adaptation training in current literature as well as the development of a talk.
Requirements for participation and examination
Desirable prerequisites for participation in the courses:
Requirements: good knowledge in Mathematics (linear algebra, analysis, stochastics) as well as programming experience.
Mandatory requirements for the module examination registration:
This module has no examination requirements.
Completion of the module
| Grading | Graded |
|---|---|
| Exam format | Portfolio examination |
| Type of portfolio examination | 100 points in total |
| Language | English |
Test elements
| Name | Points/Weight | Category | Duration/Extent |
|---|---|---|---|
| (Deliverable assessment) presentation | 50 | oral | 45 minutes |
| (Learning process review) consultation | 50 | oral | 20 minutes |
Duration of the module
The following number of semesters is estimated for taking and completing the module: 1 semester(s).
This module can be started in the following semesters: Winter semester.
Maximum number of participants
The maximum capacity of participants is 16.
Registration procedures
Die Anmeldung zur Lehrveranstaltung erfolgt über die ISIS-Seite. Dies ersetzt nicht die Prüfungsanmeldung in QISPOS bzw. beim Prüfungsamt.
Recommended reading, lecture notes
Lecture notes
unavailable
Electronic lecture notes
unavailable
Recommended literature
No recommended literature given
Assigned programs
This module is not used in any program.
Miscellaneous
M.Sc. Computer Engineering (StuPO 2015)
M.Sc. Computer Science (StuPO 2015)
M.Sc. Elektrotechnik (StuPO 2015)
M.Sc. Computer Science (StuPO 2015)
M.Sc. Elektrotechnik (StuPO 2015)