Search result: Catalogue data in Spring Semester 2018

Computer Science Master Information
Focus Courses
Focus Courses in Software Engineering
Focus Core Courses Software Engineering
NumberTitleTypeECTSHoursLecturers
263-2925-00LProgram Analysis for System Security and Reliability Information W5 credits2V + 1U + 1AM. Vechev
AbstractThe course introduces modern analysis and synthesis techniques (both, deterministic and probabilistic) and shows how to apply these methods to build reliable and secure systems spanning the domains of blockchain, computer networks and deep learning.
Objective* Understand how classic analysis and synthesis techniques work, including discrete and probabilistic methods.

* Understand how to apply the methods to generate attacks and create defenses against applications in blockchain, computer networks and deep learning.

* Understand the state-of-the-art in the area as well as future trends.
ContentThe course will illustrate how the methods can be used to create more secure and reliable systems across four application domains:

Part I: Analysis and Synthesis for Computer Networks:
1. Analysis: Datalog, Batfish
2. Synthesis: CEGIS, SyNET (Link)
3. Probabilistic: (PSI: Link), its applications to networks (Bayonet)

Part II: Blockchain security
1. Introduction to space and tools.
2. Automated symbolic reasoning.
3. Applications: verification of smart contracts (Link)

Part III: Security and Robustness of Deep Learning:
1. Basics: affine transforms, activation functions
2. Attacks: gradient based method to adversarial generation
3. Defenses: affine domains, AI2 (Link)

Part IV: Probabilistic Security:
1. Enforcement: PSI + Spire.
2. Graphical models: CRFs, Structured SVM, Pseudo-likelihood.
3. Practical statistical de-obfuscation: DeGuard: Link, JSNice: Link, and more.

To gain a deeper understanding, the course will involve a hands-on programming project.
Focus Elective Courses Software Engineering
NumberTitleTypeECTSHoursLecturers
263-2812-00LProgram Verification Information Restricted registration - show details
Number of participants limited to 30.
W5 credits2V + 1U + 1AA. J. Summers
AbstractA hands-on introduction to the theory and construction of deductive software verifiers, covering both cutting-edge methodologies for formal program reasoning, and a perspective over the broad tool stacks making up modern verification tools.
ObjectiveStudents will earn the necessary skills for designing and developing deductive verification tools which can be applied to modularly analyse complex software, including features challenging for reasoning such as heap-based mutable data and concurrency. Students will learn both a variety of fundamental reasoning principles, and how these reasoning ideas can be made practical via automatic tools.

Students will be gain practical experience with reasoning tools at various levels of abstraction, from SAT and SMT solvers at the lowest level, up through intermediate verification languages and tools, to verifiers which target front-end code in executable languages.

By the end of the course, students should have a good working understanding and experience of the issues and decisions involved with designing and building practical verification tools, and the theoretical techniques which underpin them.
ContentThe course will be organized around building up a "tool stack", starting at the lowest-level with background on SAT and SMT solving techniques, and working upwards through tools at progressively-higher levels of abstraction. The notion of intermediate verification languages will be explored, and the Boogie (Microsoft Research) and Viper (ETH) languages will be used in depth to tackle increasingly ambitious verification tasks.

The course will intermix technical content with hands-on experience; at each level of abstraction, we will build small tools on top which can tackle specific program correctness problems, starting from simple puzzle solvers (Soduko) at the SAT level, and working upwards to full functional correctness of application-level code. This practical work will include three mini-projects (each worth 10% of the final grade) spread throughout the course, which count towards the final grade. An oral examination (worth 70% of the final grade) will cover the technical content covered.
Lecture notesSlides and other materials will be available online.
LiteratureBackground reading material and links to tools will be published on the course website.
Prerequisites / NoticeSome programming experience is essential, as the course contains several practical assignments. A basic familiarity with propositional and first-order logic will be assumed.

Courses with an emphasis on formal reasoning about programs (such as Formal Methods and Functional Programming) are advantageous background, but are not a requirement.
263-2300-00LHow To Write Fast Numerical Code Information Restricted registration - show details
Does not take place this semester.
Number of participants limited to 84.

Prerequisite: Master student, solid C programming skills.
W6 credits3V + 2UM. Püschel
AbstractThis course introduces the student to the foundations and state-of-the-art techniques in developing high performance software for numerical functionality such as linear algebra and others. The focus is on optimizing for the memory hierarchy and for special instruction sets. Finally, the course will introduce the recent field of automatic performance tuning.
ObjectiveSoftware performance (i.e., runtime) arises through the interaction of algorithm, its implementation, and the microarchitecture the program is run on. The first goal of the course is to provide the student with an understanding of this interaction, and hence software performance, focusing on numerical or mathematical functionality. The second goal is to teach a general systematic strategy how to use this knowledge to write fast software for numerical problems. This strategy will be trained in a few homeworks and semester-long group projects.
ContentThe fast evolution and increasing complexity of computing platforms pose a major challenge for developers of high performance software for engineering, science, and consumer applications: it becomes increasingly harder to harness the available computing power. Straightforward implementations may lose as much as one or two orders of magnitude in performance. On the other hand, creating optimal implementations requires the developer to have an understanding of algorithms, capabilities and limitations of compilers, and the target platform's architecture and microarchitecture.

This interdisciplinary course introduces the student to the foundations and state-of-the-art techniques in high performance software development using important functionality such as linear algebra functionality, transforms, filters, and others as examples. The course will explain how to optimize for the memory hierarchy, take advantage of special instruction sets, and, if time permits, how to write multithreaded code for multicore platforms. Much of the material is based on state-of-the-art research.

Further, a general strategy for performance analysis and optimization is introduced that the students will apply in group projects that accompany the course. Finally, the course will introduce the students to the recent field of automatic performance tuning.
Seminar in Software Engineering
NumberTitleTypeECTSHoursLecturers
263-2100-00LResearch Topics in Software Engineering Information Restricted registration - show details
Number of participants limited to 22.
W2 credits2ST. Gross
AbstractThis seminar introduces students to the latest research trends that help to improve various aspects of software quality.

Topics cover the following areas of research: Compilers, domain-specific languages, concurrency, formal methods, performance optimization, program analysis, program generation, program synthesis, testing, tools, verification
ObjectiveAt the end of the course, the students should be:

- familiar with a broad range of key research results in the area as well as their applications.

- know how to read and assess high quality research papers

- be able to highlight practical examples/applications, limitations of existing work, and outline potential improvements.
ContentThe course will be structured as a sequence of presentations of high-quality research papers, spanning both theory and practice. These papers will have typically appeared in top conferences spanning several areas such as POPL, PLDI, OOPSLA, OSDI, ASPLOS, SOSP, AAAI, ICML and others.
LiteratureThe publications to be presented will be announced on the seminar home page at least one week before the first session.
Prerequisites / NoticePapers will be distributed during the first lecture.
263-2926-00LDeep Learning for Big Code Information Restricted registration - show details
Number of participants limited to 24.
W2 credits2SM. Vechev
AbstractThe seminar covers some of the latest and most exciting developments (industrial and research) in the field of Deep Learning for Code, including new methods and latest systems, as well as open challenges and opportunities.
ObjectiveThe objective of the seminar is to:

- Introduce students to the field of Deep Learning for Big Code.

- Learn how machine learning models can be used to solve practical challenges in software engineering and programming beyond traditional methods.

- Highlight the latest research and work opportunities in industry and academia available on this topic.
ContentThe last 5 years have seen increased interest in applying advanced machine learning techniques such as deep learning to new kind of data: program code. As the size of open source code increases dramatically (over 980 billion lines of code written by humans), so comes the opportunity for new kind of deep probabilistic methods and commercial systems that leverage this data to revolutionize software creation and address hard problems not previously possible. Examples include: machines writing code, program de-obfuscation for security, code search, and many more.

Interestingly, this new type of data, unlike natural language and images, introduces technical challenges not typically encountered when working with standard datasets (e.g., images, videos, natural language), for instance, finding the right representation over which deep learning operates. This in turn has the potential to drive new kinds of machine learning models with broad applicability.

Because of this, there has been substantial interest over the last few years in both industry (e.g., companies such as Facebook starting, various start-ups in the space such as Link), academia (e.g., Link) and government agencies (e.g., DARPA) on using machine learning to automate various programming tasks.

In this seminar, we will cover some of the latest and most exciting developments in the field of Deep Learning for Code, including new methods and latest systems, as well as open challenges and opportunities.

The seminar is carried out as a set of presentations chosen from a list of available papers. The grade is determined as a function of the presentation, handling questions and answers, and participation.
Prerequisites / NoticeThe seminar is carried out as a set of presentations chosen from a list of available papers. The grade is determined as a function of the presentation, handling questions and answers, and participation.

The seminar is ideally suited for M.Sc. students in Computer Science.
263-2930-00LBlockchain Security Seminar Information Restricted registration - show details
Number of participants limited to 26.
W2 credits2SM. Vechev, D. Drachsler Cohen, P. Tsankov
AbstractThis seminar introduces students to the latest research trends in the field of blockchains.
ObjectiveThe objectives of this seminar are twofold: (1) learning about the blockchain platform, a prominent technology receiving a lot of attention in computer Science and economy and (2) learning to convey and present complex and technical concepts in simple terms, and in particular identifying the core idea underlying the technicalities.
ContentThis seminar introduces students to the latest research trends in the field of blockchains. The seminar covers the basics of blockchain technology, including motivation for decentralized currency, establishing trust between multiple parties using consensus algorithms, and smart contracts as a means to establish decentralized computation. It also covers security issues arising in blockchains and smart contracts as well as automated techniques for detecting vulnerabilities using programming language techniques.
  •  Page  1  of  1