This course provides an introduction to Bayesian Statistics and its applications to data analysis in various fields.

The class will focus on large language models (LLMs) and language-and-vision models, as well as on generative methods for artificial intelligence (AI). Topics include deep neural networks, transformers, large language models, generative adversarial networks, diffusion models, and applications of such architectures and methods to image analysis, image synthesis, and text-to-image translation.

This course will cover popular methods in machine learning and data mining, with an emphasis on developing a working understanding of how to apply these methods in practice. The course will focus on basic foundational concepts underpinning and motivating modern machine learning and data mining approaches. We will also discuss recent research developments.

Introduction to the theory, algorithms, and applications of automated learning. How much information is needed to learn a task, how much computation is involved, and how it can be accomplished. Special emphasis will be given to unifying the different approaches to the subject coming from statistics, function approximation, optimization, pattern recognition, and neural networks.

Statistical Inference is a branch of mathematical engineering that studies ways of extracting reliable information from limited data for learning, prediction, and decision making in the presence of uncertainty. This is an introductory course on statistical inference. The main goals are: develop statistical thinking and intuitive feel for the subject; introduce the most fundamental ideas, concepts, and methods of statistical inference; and explain how and why they work, and when they don't. Topics covered include summarizing data, fundamentals of survey sampling, statistical functionals, jackknife, bootstrap, methods of moments and maximum likelihood, hypothesis testing, p-values, the Wald, Student's t-, permutation, and likelihood ratio tests, multiple testing, scatterplots, simple linear regression, ordinary least squares, interval estimation, prediction, graphical residual analysis.

The main goal of the course is to provide an introduction to the central concepts and core methods of statistical learning, an interdisciplinary field at the intersection of applied mathematics, statistical inference, and machine learning. The course focuses on the mathematics and statistics of methods developed for learning from data. Students will learn what methods for statistical learning exist, how and why they work (not just what tasks they solve and in what built-in functions they are implemented), and when they are expected to perform poorly. The course is oriented for upper level undergraduate students in IDS, ACM, and CS and graduate students from other disciplines who have sufficient background in linear algebra, probability, and statistics. The course is a natural continuation of CS 157 and it can be viewed as a statistical analog of CS 155. Topics covered include elements of statistical decision theory, regression and classification problems, nearest-neighbor methods, curse of dimensionality, linear regression, model selection, cross-validation, subset selection, shrinkage methods, ridge regression, LASSO, logistic regression, linear and quadratic discriminant analysis, support-vector machines, tree-based methods, bagging, and random forests.

This course focuses on current topics in machine learning research. This is a paper reading course, and students are expected to understand material directly from research articles. Students are also expected to present in class, and to do a final project.

Lecture, laboratory, and project course aimed at understanding visual information processing, in both machines and the mammalian visual system. The course will emphasize an interdisciplinary approach aimed at understanding vision at several levels: computational theory, algorithms, psychophysics, and hardware (i.e., neuroanatomy and neurophysiology of the mammalian visual system). The course will focus on early vision processes, in particular motion analysis, binocular stereo, brightness, color and texture analysis, visual attention and boundary detection. Students will be required to hand in approximately three homework assignments as well as complete one project integrating aspects of mathematical analysis, modeling, physiology, psychophysics, and engineering. Given in alternate years; offered 2023-24.

This course aims at a quantitative understanding of how the nervous system computes. The goal is to link phenomena across scales from membrane proteins to cells, circuits, brain systems, and behavior. We will learn how to formulate these connections in terms of mathematical models, how to test these models experimentally, and how to interpret experimental data quantitatively. The concepts will be developed with motivation from some of the fascinating phenomena of animal behavior, such as: aerobatic control of insect flight, precise localization of sounds, sensing of single photons, reliable navigation and homing, rapid decision-making during escape, one-shot learning, and large-capacity recognition memory. Not offered 2023-2024.

This course is a both a theoretical and practical introduction to functional programming.

This course introduces students to the modern practice of interactive tactic-based theorem proving using the Coq theorem prover. Topics will be drawn from logic, programming languages and the theory of computation. Topics will include: proof by induction, lists, higher-order functions, polymorphism, dependently-typed functional programming, constructive logic, the Curry-Howard correspondence, modeling imperative programs, and other topics if time permits. Students will be graded partially on attendance and will be expected to participate in proving theorems in class.

CS 131 is a course on programming languages and their implementation. It teaches students how to program in a number of simplified languages representing the major programming paradigms in use today (imperative, object-oriented, and functional). It will also teach students how to build and modify the implementations of these languages. Emphasis will not be on syntax or parsing but on the essential differences in these languages and their implementations. Both dynamically-typed and statically-typed languages will be implemented. Relevant theory will be covered as needed. Implementations will mostly be interpreters, but some features of compilers will be covered if time permits. Enrollment limited to 30 students.

This course covers the construction of compilers: programs which convert program source code to machine code which is directly executable on modern hardware. The course takes a bottom-up approach: a series of compilers will be built, all of which generate assembly code for x86 processors, with each compiler adding features. The final compiler will compile a full-fledged high-level programming language to assembly language. Topics covered include register allocation, conditionals, loops and dataflow analysis, garbage collection, lexical scoping, and type checking. This course is programming intensive. All compilers will be written in the OCaml programming language.

Shannon's mathematical theory of communication, 1948-present. Entropy, relative entropy, and mutual information for discrete and continuous random variables. Shannon's source and channel coding theorems. Mathematical models for information sources and communication channels, including memoryless, Markov, ergodic, and Gaussian. Calculation of capacity and rate-distortion functions. Universal source codes. Side information in source coding and communications. Network information theory, including multiuser data compression, multiple access channels, broadcast channels, and multiterminal networks. Discussion of philosophical and practical implications of the theory. This course, when combined with EE 112, CS 127, CS 161, and CS 167, should prepare the student for research in information theory, coding theory, wireless communications, and/or data compression.

This course develops from first principles the theory and practical implementation of the most important techniques for combating errors in digital transmission and storage systems. Topics include highly symmetric linear codes, such as Hamming, Reed-Muller, and Polar codes; algebraic block codes, such as Reed-Solomon and BCH codes, including a self-contained introduction to the theory of finite fields; and low-density parity-check codes. Students will become acquainted with encoding and decoding algorithms, design principles and performance evaluation of codes.

This course introduces the basics of power systems analysis: phasor representation, 3-phase transmission system, transmission line models, transformer models, per-unit analysis, network matrix, power flow equations, power flow algorithms, optimal powerflow (OPF) problems, unbalanced power flow analysis and optimization,swing dynamics and stability.

There is a large gap between the public and private sectors' effective use of technology. This gap presents an opportunity for the development of innovative solutions to problems faced by society. Students will develop technology-based projects that address this gap. Course material will offer an introduction to the design, development, and analysis of digital technology with examples derived from services typically found in the public sector. Not offered 2023-24.

Programming distributed systems. Mechanics for cooperation among concurrent agents. Programming sensor networks and cloud computing applications. Applications of machine learning and statistics by using parallel computers to aggregate and analyze data streams from sensors. Not offered 2023-24.

Social networks, the web, and the internet are essential parts of our lives, and we depend on them every day. CS 143 and CS 144 study how they work and the 'big' ideas behind our networked lives. In this course, the questions explored include: Why is an hourglass architecture crucial for the design of the Internet? Why doesn't the Internet collapse under congestion? How are cloud services so scalable? How do algorithms for wireless and wired networks differ? For all these questions and more, the course will provide a mixture of both mathematical analysis and hands-on labs. The course expects students to be comfortable with graph theory, probability, and basic programming.

Social networks, the web, and the internet are essential parts of our lives, and we depend on them every day. CS 143 and CS 144 study how they work and the 'big' ideas behind our networked lives. In this course, the questions explored include: What do networks actually look like (and why do they all look the same)?; How do search engines work?; Why do epidemics and memes spread the way they do?; How does web advertising work? For all these questions and more, the course will provide a mixture of both mathematical analysis and hands-on labs. The course expects students to be comfortable with graph theory, probability, and basic programming.

Students are expected to execute a substantial project in networking, write up a report describing their work, and make a presentation.

This is a research-oriented course meant for undergraduates and beginning graduate students who want to learn about current research topics in networks such as the Internet, power networks, social networks, etc. The topics covered in the course will vary, but will be pulled from current research in the design, analysis, control, and optimization of networks.

Basics of information theory: entropy, mutual information, source and channel coding theorems. Basics of coding theory: error-correcting codes for information transmission and storage, block codes, algebraic codes, sparse graph codes. Basics of digital communications: sampling, quantization, digital modulation, matched filters, equalization.

Next generation networks will have tens of billions of nodes forming cyber-physical systems and the Internet of Things. A number of fundamental scientific and technological challenges must be overcome to deliver on this vision. This course will focus on (1) How to boost efficiency and reliability in large networks; the role of network coding, distributed storage, and distributed caching; (2) How to manage wireless access on a massive scale; modern random access and topology formation techniques; and (3) New vistas in big data networks, including distributed computing over networks and crowdsourcing. A selected subset of these problems, their mathematical underpinnings, state-of-the-art solutions, and challenges ahead will be covered. Not offered 2023-24.

The course develops the core concepts of robotics. The first quarter focuses on classical robotic manipulation, including topics in rigid body kinematics and dynamics. It develops planar and 3D kinematic formulations and algorithms for forward and inverse computations, Jacobians, and manipulability. The second quarter transitions to planning, navigation, and perception. Topics include configuration space, sample-based planners, A* and D* algorithms, to achieve collision-free motions. Course work transitions from homework and programming assignments to more open-ended team-based projects.

This course builds up, and brings to practice, the elements of robotic systems at the intersection of hardware, kinematics and control, computer vision, and autonomous behaviors. It presents selected topics from these domains, focusing on their integration into a full sense-think-act robot. The lectures will drive team-based projects, progressing from building custom robotic arms (5 to 7 degrees of freedom) to writing all necessary software (utilizing the Robotics Operating system, ROS). Teams are required to implement and customize general concepts for their selected tasks. Working systems will autonomously operate and demonstrate their capabilities during final presentations.

The class will focus on large language models (LLMs) and language-and-vision models, as well as on generative methods for artificial intelligence (AI). Topics include deep neural networks, transformers, large language models, generative adversarial networks, diffusion models, and applications of such architectures and methods to image analysis, image synthesis, and text-to-image translation.

Lecture, laboratory, and project course aimed at understanding visual information processing, in both machines and the mammalian visual system. The course will emphasize an interdisciplinary approach aimed at understanding vision at several levels: computational theory, algorithms, psychophysics, and hardware (i.e., neuroanatomy and neurophysiology of the mammalian visual system). The course will focus on early vision processes, in particular motion analysis, binocular stereo, brightness, color and texture analysis, visual attention and boundary detection. Students will be required to hand in approximately three homework assignments as well as complete one project integrating aspects of mathematical analysis, modeling, physiology, psychophysics, and engineering. Given in alternate years; offered 2023-24.

This course investigates computation by molecular systems, emphasizing models of computation based on the underlying physics, chemistry, and organization of biological cells. We will explore programmability, complexity, simulation of, and reasoning about abstract models of chemical reaction networks, molecular folding, molecular self-assembly, and molecular motors, with an emphasis on universal architectures for computation, control, and construction within molecular systems. If time permits, we will also discuss biological example systems such as signal transduction, genetic regulatory networks, and the cytoskeleton; physical limits of computation, reversibility, reliability, and the role of noise, DNA-based computers and DNA nanotechnology. Part a develops fundamental results; part b is a reading and research course: classic and current papers will be discussed, and students will do projects on current research topics.

Part a: This course will introduce students to the conceptual frameworks and tools of computer science as applied to molecular engineering, as well as to the practical realities of synthesizing and testing their designs in the laboratory. In part a, students will design and construct DNA circuits and self-assembled DNA nanostructures, as well as quantitatively analyze the designs and the experimental data. Students will learn laboratory techniques including fluorescence spectroscopy and atomic force microscopy and will use software tools and program in Mathematica. Part b is an open-ended design and build project

The theory of quantum information and quantum computation. Overview of classical information theory, compression of quantum information, transmission of quantum information through noisy channels, quantum error-correcting codes, quantum cryptography and teleportation. Overview of classical complexity theory, quantum complexity, efficient quantum algorithms, fault-tolerant quantum computation, physical implementations of quantum computation.

Design, construction, and characterization of engineered biological systems. Students propose and execute research projects in biomolecular engineering and synthetic biology, emphasizing projects that apply rational or library-based design strategies to the control of system behavior.