Teaching
Contact
Head of Group
Office
Teaching
Semester:
Summer term
2022
2.01.488 Advanced Automata Theory -
Event date(s) | room
- Mittwoch, 20.4.2022 10:15 - 11:45 | A05 1-160
- Mittwoch, 20.4.2022 14:15 - 15:45
- Mittwoch, 27.4.2022 10:15 - 11:45 | A05 1-160
- Mittwoch, 27.4.2022 14:15 - 15:45
- Mittwoch, 4.5.2022 10:15 - 11:45 | A05 1-160
- Mittwoch, 4.5.2022 14:15 - 15:45
- Mittwoch, 11.5.2022 10:15 - 11:45 | A05 1-160
- Mittwoch, 11.5.2022 14:15 - 15:45
- Mittwoch, 18.5.2022 10:15 - 11:45 | A05 1-160
- Mittwoch, 18.5.2022 14:15 - 15:45 | A02 2-239
- Mittwoch, 25.5.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 25.5.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 1.6.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 1.6.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 8.6.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 8.6.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 15.6.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 15.6.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 22.6.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 22.6.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 29.6.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 29.6.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 6.7.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 6.7.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 13.7.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 13.7.2022 14:15 - 15:45 | A03 2-209
- Mittwoch, 20.7.2022 10:15 - 11:45 | A03 2-209
- Mittwoch, 20.7.2022 14:15 - 15:45 | A03 2-209
Description
In this course, we study advanced topics of automata on finite words and introduce automata on infinite words. Moreover, we investigate the relationship of automata to logic and computer-aided verification. We will also study how automata can be used to provide formal guarantees for (recurrent) neural networks. In particular, we will consider the following topics:
* Automata over finite trees
* Learning of finite automata
* Various types of automata over infinite words (e.g., Büchi, Parity, Muller, and Rabin), their properties, and their relationship to each other
* Connection of automata and logic, specifically Linear Temporal Logic, first-order logic, and monadic second-order logic
If time permits, we will also study infinite games as a mechanism for the automated synthesis of reactive systems.
* Automata over finite trees
* Learning of finite automata
* Various types of automata over infinite words (e.g., Büchi, Parity, Muller, and Rabin), their properties, and their relationship to each other
* Connection of automata and logic, specifically Linear Temporal Logic, first-order logic, and monadic second-order logic
If time permits, we will also study infinite games as a mechanism for the automated synthesis of reactive systems.
lecturer
SWS
--
Semester:
Summer term
2022
2.01.495 Verification of Neural Networks -
Event date(s) | room
- Mittwoch, 20.4.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 27.4.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 4.5.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 11.5.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 18.5.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 25.5.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 1.6.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 8.6.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 15.6.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 22.6.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 29.6.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 6.7.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 13.7.2022 16:15 - 17:45 | A03 2-209
- Mittwoch, 20.7.2022 16:15 - 17:45 | A03 2-209
Description
The exceptional performance of deep neural networks in areas such as perception and natural language processing has made them an integral part of many real-world AI systems, including safety-critical ones such as medical diagnosis and autonomous driving. However, neural networks are inherently opaque, and numerous defects have been found in state-of-the-art networks.
In this seminar, we will study various methods for proving the reliability of deep neural networks. To this end, we will work with the book "Introduction to Neural Network Verification" by Aws Albarghouthi and select current research papers. Please note that this seminar will focus on formal methods, including topics related to logic and automated reasoning. It is not about deep learning.
In this seminar, we will study various methods for proving the reliability of deep neural networks. To this end, we will work with the book "Introduction to Neural Network Verification" by Aws Albarghouthi and select current research papers. Please note that this seminar will focus on formal methods, including topics related to logic and automated reasoning. It is not about deep learning.
lecturer
SWS
2
Lehrsprache
englisch
Semester:
Summer term
2022
2.01.494 Applied Verification Lab: Neural Networks -
Event date(s) | room
- Mittwoch, 20.4.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 27.4.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 4.5.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 11.5.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 18.5.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 25.5.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 1.6.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 8.6.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 15.6.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 22.6.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 29.6.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 6.7.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 13.7.2022 12:15 - 13:45 | A03 2-209
- Mittwoch, 20.7.2022 12:15 - 13:45 | A03 2-209
Description
The exceptional performance of deep neural networks in areas such as perception and natural language processing has made them an integral part of many real-world AI systems, including safety-critical ones such as medical diagnosis and autonomous driving. However, neural networks are inherently opaque, and numerous defects have been found in state-of-the-art networks.
In this lab, we will apply various methods for proving the reliability of deep neural networks. In particular, we will use state-of-the-art tools, such as Crown, ERAN, Marabou, and Planet, and apply them to examples from the neural network verification competition.
In this lab, we will apply various methods for proving the reliability of deep neural networks. In particular, we will use state-of-the-art tools, such as Crown, ERAN, Marabou, and Planet, and apply them to examples from the neural network verification competition.
lecturer
SWS
--
Lehrsprache
englisch