Updated November 25, 2020
This page serves the purpose to help graduate students understand each graduate course offered during the 2020-2021 academic year. Please use this page as a guideline to help decide what courses to take.
Winter 2021 Graduate Course Information
CSE 202 - Algorithm Design and Analysis with Prof. Daniele Miccancio
Description: This course is intended primarily for non-theory computer science graduate students who will need to formalize computational problems, and design and implement or use existing efficient algorithms for these problems. The emphasis is on methods that can be used for a wide variety of cutting edge applications, as in genomics, data mining, or VLSI design. This course presents general techniques and paradigms for designing and analyzing algorithms for a variety of computational problems. Techniques include reductions between problems, graph search, greedy algorithms, divide and conquer, back-tracking, dynamic programming, and hill-climbing. For each technique, we will see a variety of applications of the technique, starting with relatively straight-forward applications and then going on to look at applications or settings that stretch the uses of the technique. These examples will be presented in class, in independent reading in the text, and discovered by students in assignments. A course project will be on formalizing computational problems and inventing algorithms for them.
Required Knowledge: We assume some undergraduate exposure to discrete mathematics, and to algorithms and their analysis, and the ability to read, recognize, and write a valid proof. For example, CSE 20, CSE 21, and CSE 101 cover the prerequisite material. We will cover many of the same topics from an undergraduate algorithms courses. However, after a quick review of the basics, we will move on to related advanced material for each topic. If your background in these areas is weak, you may need to do some extra work to catch up. Please talk to me about this. Students without prerequisites will be admitted at their own risk.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Please see above.
Link to Past Course: https://sites.google.com/a/eng.ucsd.edu/cse101/cse202
CSE 203B - Convex Optimization Algorithms with Prof. CK Cheng
Description: We study the formulations and algorithms solving convex optimization problems. The topics include convex sets, functions, optimality conditions, duality concepts, gradient descent, conjugate gradient, interior-point methods, and applications. The objective of the course is to provide students the background and techniques for scientific computing and system optimization.
Required Knowledge: Linear algebra
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Read the appendix of the textbook
Link to Past Course: http://cseweb.ucsd.edu/classes/wi20/cse203B-a/
CSE 216 - Interaction Design Research with Prof. Scott Klemmer
Description: This course is a broad graduate-level introduction to interaction design research. The course begins with seminal work on interactive systems, and moves through current and future research areas in interaction techniques and the design, prototyping, and evaluation of user interfaces. Topics include social computing, crowdsourcing, software tools, design and evaluation methods, ubiquitous and context-aware computing, tangible interfaces, and mobile interfaces.
COGS 230 / CSE 216 is a 4-unit course. Students will need four skills: critically reading research papers, undertaking a small research project, giving a presentation, and writing a paper. Students registered for the class will receive a letter grade; the "credit/no credit" option is not available. Students in this course are encouraged to enroll in the Design at Large (CSE 219) seminar for 1 unit.
Required Knowledge: Please see website below for more information.
Enforced Prerequisite: Open to PhDs. MS students need an Intro to HCI course (CSE 170 or equivalent).
Recommended Preparation for Those Without Required Knowledge: Please see website below for more information.
Link to Past Course: https://d.ucsd.edu/IxD/research/2020/
CSE 221 - Operating Systems with Prof. Geoffrey Voelker
Description: The purpose of this course is to learn about a wide variety of operating system design and implementation principles and techniques, examining their usage in the context of both important historical systems as well as modern systems. In addition, students learn how to read research papers in a critical manner, how to articulate an understanding of and insights into material described in research papers, and how to synthesize research themes and topics across multiple systems. The topics covered include operating system structures, synchronization, virtual memory, distribution, interaction of architecture and systems, scheduling, communication, file systems, multicore scalability, and virtual machine monitors.
Required Knowledge: Undergraduate operating systems.
Enforced Prerequisite: Yes. CSE 120 or equivalent.
Recommended Preparation for Those Without Required Knowledge: Take CSE 120.
Link to Past Course: https://cseweb.ucsd.edu/classes/wi14/cse202-a/
CSE 231 - Advanced Compiler Design with Prof. Joseph Politz
Description: The course will focus on optimizations and runtime systems for programming languages, and will be project-focused. Students will work individually or in groups to construct and measure interesting, modern approaches to compiler construction and program optimization. Topics may vary depending on the interests of the class and the trajectory of the projects. Likely topics are just-in-time (JIT) compilation, memory management, and compiling for the Web (e.g. WASM, JavaScript), along with standard control-flow and value-based ahead-of-time optimizations.
Required Knowledge: Preparation similar to CSE 131 (one offering: https://ucsd-cse131-f19.github.io/). This includes experience programming in a structurally recursive style as in Ocaml, Haskell, or similar; ability to implement syntax transformations/code generation up to a simple calling convention; ability to write C code that works with pointer representations; an understanding of process and memory layout.
Enforced Prerequisite: Yes. CSE 131 (or equivalent) or consent of instructor.
Recommended Preparation for Those Without Required Knowledge: The materials for CSE 131 are available online; you can complete the assignments at your own pace (https://ucsd-cse131-f19.github.io/).
Link to Past Course: https://ucsd-pl.github.io/cse231/wi18/ (though this offering may have some significant updates, the spirit of the material remains the same)
CSE 232B - Database System Implementation with Prof. Alin Deutsch
Description: The course explores the internal structure of database systems via a project that takes students from specification of the grammar and semantics of the query language to implementing an evaluation engine for it, then building in optimizations. We use as vehicle the W3C Standard XML query language XQuery.
Required Knowledge: Students should have taken an introductory course on principles of database systems (CSE 132A or CSE 232, or equivalent project experience with SQL databases). They should also be comfortable programming in Java.
Enforced Prerequisite: Yes. Java programming, introductory database course, or equivalent experience with SQL databases.
Recommended Preparation for Those Without Required Knowledge: Online Java programming course and intro db course or db project.
Link to Past Course: http://ubicomp.ucsd.edu/cse218
CSE 240C - Advanced Microarchitecture with Prof. Samira Mirbagher
Description: 240c is primarily a paper-driven course and has one project. This course has been presented before. To better accommodate for a remote class, changes are applied to improve learning quality online. You will be expected to read the papers before each class, turn in your paper summaries, and come prepared to participate in the discussion.
Computer architects have been striving to improve performance and efficiency since the first stored program computer was designed half a century ago. Superscalar and multithreading execution are key technique towards this aim but come with the cost of security and power. The course will explore microarchitectural approaches to improve performance, efficiency and security.
Topics covered will include: Historical Perspective, Branch Prediction, ILP and Speculation, Multithreading, Thread-level speculation, Cache and Memory, Prefetching, Energy and Power, Fault Tolerance, Application-Specific Processors, Spatial/grid processing, GPU Architecture, Hardware security
Five elements will factor into your grade in this class: (1) Paper slides and presentations (2) Future work summaries (3) Participation (4) Project (5) Final exam. The schedule of papers will be posted on Piazza.
1. Paper summaries (due via private piazza messages, 10 minutes before class (no late summaries accepted) -- a paper summary is no more than 1 page long (no more than 2 pages total for all papers for a particular day), and contains the following information: Summary of the paper [in the neighborhood of 2-3 sentences, everything else about 1 sentence each]; What is the most important result?; One strength of the paper; One weakness of the paper; One thing you didn't understand; If you wrote this paper, what would your next paper be about? (Don't overthink this -- if a question is inappropriate for the paper, skip it.)
2. Presentations. Each student will make a presentation of a particular topic to the class. Think of it as a mini research exam, where you become the expert on a topic, and lead discussion of a couple of papers. It is important that you not only establish yourself as an expert in the area, but also foster discussion among the class.
3. Participation. Each class will be a DISCUSSION of a particular set of papers. I expect each student to participate in, but not dominate, the conversations about the papers.
4. Projects -- more discussion of this on Piazza.
5. We will have a final exam. It will cover the papers you read during the quarter.
By the end of this course, you will gain a general knowledge of advanced microarchitecture designs and evaluation methodologies, as well as an in-depth understanding of one problem you pick for your project. You will also gain skills in critically reviewing papers and doing many gain experience on developing novel ideas that improve current studies.
Required Knowledge: General Computer Architecture and C/C++ skills for the project.
Enforced Prerequisite: Yes. CSE 240A.
Recommended Preparation for Those Without Required Knowledge: Get a book on C++ and practice. Students must have taken 240a.
Link to Past Course: https://cseweb.ucsd.edu/classes/sp15/cse240C-a/papers.html
CSE 240D - Application-Specific Processors with Prof. Hadi Esmaeilzadeh
Description: Deep learning is set to revolutionize medicine, robotics, commerce, transportation and numerous other aspects of our lives. However, these impacts are contingent upon providing high-performance compute capabilities alongside restrained power consumption. Significant effort has been made both by modifying hardware and software in the direction of enhancing the speed of neural networks and their consumed energy. The compute intensity of neural networks and the inability of general purpose processors to meet this high demand accentuate the need for applications specific hardware accelerators that are custom designed for deep neural network computations. In this course, you will gain insight on the design process of these accelerators, as well as deep neural network architectures and characteristics by discussing the prevalent literature in the area. This is a project-based course, so you will also acquire hands-on knowledge on how to actually construct an accelerator through the project. The project will step by step guide you through outlining your architecture and developing a functional and timing simulator for it.
The course will be a combination of lectures, student presentation, project and separate brainstorming sessions in which we will collectively explore and develop new ideas on each topic. The students will do a project in which they design an accelerator for deep neural networks and develop a simulator for their design.
Required Knowledge: There are no pre-requisites for the course but students will be expected to be comfortable with a) basic computer organization, and b) programming in a language such as Python or C/C++. The course is open to both PhD and MS students. Undergraduate students require permission from the instructor.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: See "Required Knowledge."
Link to Past Course: https://hadiclass.github.io/cse240d-fa19/index.html
CSE 241A - VLSI Integrated Computing Circuitry with Prof. Andrew Kahng
Description: Design methodologies and tradeoffs in the physical ("netlist to layout") implementation of VLSI systems. Algorithm and optimization foundations for solution of interconnection, clustering, partitioning, placement, power-ground distribution, clocking, routing, performance estimation, timing closure and other fundamental VLSI physical design problem classes. Computer-aided design and performance analysis algorithms and optimization frameworks are studied in the context of lectures/homeworks, design exercises and mini-projects.
Textbook: Weste and Harris, CMOS VLSI Design: A Circuits and Systems Perspective, 4th edition.
Required Knowledge: Layout (ECE 165 or ECE 260A) and logic design (CSE 140) are strongly recommended as background.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Study of a VLSI Design book such as Weste/Harris (used in ECE 260A), and browsing of a VLSI physical design algorithms book such as Sherwani.
Link to Past Course: Not possible (course website is password-protected due to IP access restrictions agreed to by UCSD and various companies (EDA, foundry).
CSE 251A - ML: Learning Algorithms with Prof. Kamalika Chaudhuri
Description: (Formerly CSE 250B. Students cannot receive credit for both CSE 250B and CSE 251A) The goal of this class is to provide a broad introduction to machine-learning at the graduate level. The topics covered in this class include some topics in supervised learning, such as k-nearest neighbor classifiers, linear and logistic regression, decision trees, boosting and neural networks, and topics in unsupervised learning, such as k-means, singular value decompositions and hierarchical clustering.
The topics covered in this class will be different from those covered in CSE 250-A. In addition to the actual algorithms, we will be focussing on the principles behind the algorithms in this class.
Required Knowledge: Knowledge of linear algebra, calculus, probability and programming
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Online probability and linear algebra class.
Link to Past Course: https://cseweb.ucsd.edu/classes/wi20/cse250B-a/index.html
CSE 251B: Neural Networks for Pattern Recognition (AKA “Deep Learning”) with Prof. Gary Cottrell
Description: (Formerly CSE 253. Students cannot receive credit for both CSE 253 and CSE 251B) Neural networks have come back into fashion since 2012 - especially for computer vision - when a deep network won the ImageNet Large Scale Visual Recognition Challenge (ILSVRC). They are now also used for language translation (google translate), language generation (GPT-3), image generation, game playing, and speech recognition. This course covers perceptrons, logistic and softmax regression, multilayer networks and back propagation, deep learning, convolutional networks, recurrent nets, transformers, reinforcement learning, and some recent papers. The course is in a lecture format, but most of your learning will occur from doing programming assignments. The course will involve four programming assignments roughly every two weeks, a final project (done in teams), and two in-class midterms. A typical final project will replicate some recent work, perhaps on a different dataset. This course is like a full-time job – don’t take any other courses that are time-consuming at the same time.
Required Knowledge: Students should be fluent in linear algebra, vector calculus, probability and programming. They should know what gradient descent is.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: If you are missing the math background, you should acquire it via online courses or by taking classes at UCSD. This math background is essential. Look up gradient descent.
Link to Past Course: Here is the syllabus from CSE 253 (the prior course number): https://cseweb.ucsd.edu/~gary/253/CSE_253_syllabus_w20.pdf
CSE 252B - Computer Vision II with Prof. Ben Ochoa
Description: This course covers topics in imaging geometry using techniques from computer vision, photogrammetry, and projective geometry. These topics include methods for projecting a 3D scene to a 2D image, reconstructing a 3D scene from 2D images, and camera parameter estimation. Upon completion of this course, students will have an understanding of single and multiple view geometry, and the algorithms used in practice.
Required Knowledge: Linear algebra, calculus, basic probability and statistics. MATLAB, Python, or other programming experience.
Enforced Prerequisite: None. CSE 252A is not a prerequisite to CSE 252B.
Recommended Preparation for Those Without Required Knowledge: Undergraduate courses and textbooks.
Link to Past Course: https://cseweb.ucsd.edu/classes/wi20/cse252B-a/
CSE 276E - Robotic System Design & Implementation with Prof. Steven Swanson
Description: End-to-end system design of embedded electronic systems including PCB design and fabrication, software control system development, and system integration. The course is an intensive capstone project course that provides students a chance to integrate knowledge from a wide range of domains into a single project.
Required Knowledge: Basic electronics and intermediate C/C++ programming.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Read through an undergraduate circuits textbook. If you don’t know how to program, you should not take the class.
Link to Past Course: https://sites.google.
CSE 280A - Algorithms/Computational Biol with Prof. Vineet Bafna
Description: The course presents selected topics in Computational Molecular Biology, with an emphasis on designing algorithms for biological data analysis. We focus specifically on problems from Population genetics and genomics.
Required Knowledge: The student should be comfortable with basic combinatorics and algorithms. Otherwise, the course should be self-contained. No knowledge of biology is assumed, but the student must have an interest in learning some new topics.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: One or more of the following classes would be great: CSE 101, CSE 202, CSE 181, CSE 282.
Link to Past Course: http://proteomics.ucsd.edu/vbafna/teaching-2/cse280a-algorithms-for-genetics/
CSE 282 - Bioinformatics II: Sequence and Structure Analysis - Methods and Applications with Prof. Pavel Pevzner
Description: This class closely follows the textbook Phillip Compeau and Pavel Pevzner. Bioinformatics Algorithms: An Active Learning Approach. 3rd edition. Active Learning Publishers 2018.
Required Knowledge: The course assumes some prior background in biology, some algorithmic culture (CSE 101 course on algorithms), and some programming skills.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: If you don’t have the required knowledge listed above, you should take the online UCSD course on Algorithms on Coursera.
Link to Past Course: N/A
CSE 291 (C00) - Security, Privacy, and US Law with Prof. Stefan Savage
Description: This course will explore the intersection of the technical and the legal around issues of computer security and privacy, as they manifest in the contemporary US legal system. The goal of the course is multifold: First, to provide a better understanding of how key portions of the US legal system operate in the context of electronic communications, storage and services. Second, to provide a pragmatic foundation for understanding some of the common legal liabilities associated with empirical security research (particularly laws such as the DMCA, ECPA and CFAA, as well as some understanding of contracts and how they apply to topics such as "reverse engineering"). Third, we will explore how changes in technology and law co-evolve and how this this process is highlighted in current legal and policy "fault lines". This will very much be a readings and discussion class, so be prepared to enage if you sign up.
Required Knowledge: None, but it we are going to assume you understand enough about the technical aspects of security and privacy (e.g., such as having taking an undergraduate class in security) that we, at most, need to do cursory reviews of any technical material.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: Review lectures/readings from CSE127
Link to Past Course: N/A
CSE 291 (D00) - Dependable Systems with Prof. Yuanyuan Zhou
Description: Fundamentals and modern techniques for building dependable systems including error detection, fault tolerance, and failure recovery.
Required Knowledge: Basic and advanced computer systems and computer architecture concepts and principles. CSE221
Enforced Prerequisite: Yes. CSE 120 or equivalent.
Recommended Preparation for Those Without Required Knowledge: CSE221 reading list and CSE120 lecture video
Link to Past Course: N/A
CSE 291 (E00) - Advanced Studies in Classical Operating Systems with Prof. Joseph Pasquale
Description: Current operating systems research is sometimes characterized as “polishing a round ball,” meaning that as most of the great ideas in operating systems were developed a long time ago, all we are doing today is refining them. This is somewhat unfair, but there is a grain of truth, and looking back at the development of those great ideas is a worthy and illuminating pursuit. Toward this end, we will study some of the most influential operating systems, where most of the modern-day abstractions, mechanisms, and policies were first developed. These systems are especially interesting because of their focus on justifying the design and implementation decisions they made -- Why a particular overall structuring of the kernel? Why a certain notion of process, or synchronization mechanism, or scheduling policy? How do these decisions interact? -- which are mostly either lost or taken for granted today. It is instructive to see what predictions they made, whether they came true or not, and how their thinking process might inform us as to how we might go about making such predictions today, based on changing user interests, application demands, and technological trends. Some of the systems we will look at include the early “classical” ones: MIT’s Multics, Dijkstra’s THE, Hansen’s RC 4000, Bell Labs’ UNIX, CMU’s Hydra, etc. We will also read some “reflection” papers, i.e., retrospective studies on lessons learned regarding overall operating system design. The pace will be informal and relaxed, about one paper per week, so that we have the time to really understand and get into the details. Grading will be based mostly on class participation and a short final paper.
Required Knowledge: Operating systems
Enforced Prerequisite: Yes. CSE 221 or equivalent.
Recommended Preparation for Those Without Required Knowledge: Reading list for CSE221 (or any graduate-level operating systems course)
Link to Past Course: http://cseweb.ucsd.edu/classes/wi20/cse291-i/
CSE 291 (F00) - Domain Adaptation in Computer Vision with Prof. Manmohan Chandraker
Description: Computer vision has made rapid progress in the era of deep learning. This is largely attributed to the availability of large-scale labeled data, coupled with GPU computation. Yet, computer vision models trained on a domain, say day images, often do not generalize to new domains, say images acquired at night. It is expensive to label data for all possible scenarios, but unlabeled data is easier to obtain. In this course, we will study concepts in unsupervised domain adaptation, with applications to various computer vision problems, such as image classification, semantic segmentation, object detection, face recognition and 3D reconstruction.
Required Knowledge: This is an advanced class, covering recent research progress. Graduate students with a strong interest in computer vision may enroll. Prior background in computer vision and machine learning is required, preferably through research experience or as covered by CSE 252 and similar offerings. Students are encouraged to contact the instructor if unsure about meeting any criteria for enrollment.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: CSE 252A, CSE 252C, or Stanford CS 231n
Link to Past Course: http://cseweb.ucsd.edu/~mkchandraker/classes/CSE291/Winter2020/
CSE 291 (H00) - Introduction to Computing Education Research with Prof. Gerald Soosai Raj
Description: Computer Science as a major has high societal demand. In addition, computer programming is a skill increasingly important for all students, not just computer science majors. However, computer science remains a challenging field for students to learn. This course examines what we know about key questions in computer science education: Why is learning to program so challenging? What pedagogical choices are known to help students? What barriers do diverse groups of students (e.g., non-native English speakers) face while learning computing? How do those interested in Computing Education Research (CER) study and answer pressing research questions?
This course surveys the key findings and research directions of CER and applications of those findings for secondary and post-secondary teaching contexts. We study the development of the field, current modes of inquiry, the role of technology in computing, student representation, research-based pedagogical approaches, efforts toward increasing diversity of students in computing, and important open research questions. CER is a relatively new field and there is much to be done; an important part of the course engages students in the design phases of a computing education research study and asks students to complete a significant project (e.g., a review of an area in computing education research, designing an intervention to increase diversity in computing, prototyping of a software system to aid student learning).
Required Knowledge: None.
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: None.
Link to Past Course: N/A
CSE 291 (I00) - Machine Learning Meets Geometry with Prof. Hao Su
Description: This is a graduate-level course to cover core concepts and algorithms of geometry that are being used in computer graphics, computer vision and machine learning. I will cover key concepts of differential geometry, the usage of geometry in computer graphics, vision, and machine learning. Students are required to do a course project in pairs.
Required Knowledge: Fluent in linear algebra, taken classes of numerical methods or machine learning, fluent in at least one deep learning library, e.g., pytorch
Enforced Prerequisite: None.
Recommended Preparation for Those Without Required Knowledge: None.
Link to Past Course: https://geoml.github.io/
CSE 291 (J00) - Program Synthesis with Prof. Nadia Polikarpova
Description: This course is a comprehensive introduction to program synthesis: an emerging area that sits at the intersection of programming systems, formal methods, and artificial intelligence. The goal of program synthesis is to generate programs automatically from high-level, possibly incomplete descriptions. The course will cover a wide variety of recent synthesis techniques that differ in the kind of program description they start with (from input-output examples to formal correctness specifications), the search strategy they employ (enumerative, stochastic, or symbolic search), and the kind of information they use to guide the search (counter-example guided synthesis, type-driven synthesis, synthesis with machine learning). The course will involve a project, as well as reading and reviewing research papers.
Required Knowledge: Programming languages (CSE 230 or equivalent)
Enforced Prerequisite: None, but CSE 230 is recommended.
Recommended Preparation for Those Without Required Knowledge: N/A
Link to Past Course: https://github.com/nadia-polikarpova/cse291-program-synthesis/wiki