Master of Information and Data Science (MIDS) Program at UC Berkeley, Summaries of Machine Learning

Download Master of Information and Data Science (MIDS) Program at UC Berkeley and more Summaries Machine Learning in PDF only on Docsity! Information and Data Science: MIDS 1 Information and Data Science: MIDS The (https://datascience.berkeley.edu/form/)Master of Information and Data Science (MIDS) is an online, part-time professional degree program that prepares students to work effectively with heterogeneous, real- world data and to extract insights from the data using the latest tools and analytical methods. The program emphasizes the importance of asking good research or business questions as well as the ethical and legal requirements of data privacy and security. The Fifth Year Master of Information and Data Science (Fifth Year MIDS) is a customized pathway in the MIDS program that is designed for students who have just completed their undergraduate education at UC Berkeley. Students attend weekly live ("synchronous") sessions with classmates and instructors via an online platform as well as engaging with online ("asynchronous") videos and assignments on their own time. The curriculum includes research design and applications for data and analysis, statistics for data science, data engineering, applied machine learning, data visualization, and data ethics. MIDS features a project- based approach to learning and encourages the pragmatic application of a variety of different tools and methods to solve complex problems. Graduates of the program will be able to: • Imagine new and valuable uses for large datasets; • Retrieve, organize, combine, clean, and store data from multiple sources; • Apply appropriate data mining, statistical analysis, and machine learning techniques to detect patterns and make predictions; • Design visualizations and effectively communicate findings; and • Understand the ethical and legal requirements of data privacy and security. The I School also offers a master's in Information Management and Systems (MIMS) (http://guide.berkeley.edu/graduate/degree- programs/information-management-systems/), a master's in Information and Cybersecurity (MICS) (http://guide.berkeley.edu/graduate/ degree-programs/information-cybersecurity/), and a Ph.D (http:// guide.berkeley.edu/graduate/degree-programs/information-management- systems-phd/). (http://guide.berkeley.edu/graduate/degree-programs/ information-management-systems/#doctoraldegreerequirementstext) Admission to the University Minimum Requirements for Admission The following minimum requirements apply to all graduate programs and will be verified by the Graduate Division: 1. A bachelor’s degree or recognized equivalent from an accredited institution; 2. A grade point average of B or better (3.0); 3. If the applicant has completed a basic degree from a country or political entity (e.g., Quebec) where English is not the official language, adequate proficiency in English to do graduate work, as evidenced by a TOEFL score of at least 90 on the iBT test, 570 on the paper-and-pencil test, or an IELTS Band score of at least 7 on a 9-point scale (note that individual programs may set higher levels for any of these); and 4. Sufficient undergraduate training to do graduate work in the given field. Applicants Who Already Hold a Graduate Degree The Graduate Council views academic degrees not as vocational training certificates, but as evidence of broad training in research methods, independent study, and articulation of learning. Therefore, applicants who already have academic graduate degrees should be able to pursue new subject matter at an advanced level without the need to enroll in a related or similar graduate program. Programs may consider students for an additional academic master’s or professional master’s degree only if the additional degree is in a distinctly different field. Applicants admitted to a doctoral program that requires a master’s degree to be earned at Berkeley as a prerequisite (even though the applicant already has a master’s degree from another institution in the same or a closely allied field of study) will be permitted to undertake the second master’s degree, despite the overlap in field. The Graduate Division will admit students for a second doctoral degree only if they meet the following guidelines: 1. Applicants with doctoral degrees may be admitted for an additional doctoral degree only if that degree program is in a general area of knowledge distinctly different from the field in which they earned their original degree. For example, a physics PhD could be admitted to a doctoral degree program in music or history; however, a student with a doctoral degree in mathematics would not be permitted to add a PhD in statistics. 2. Applicants who hold the PhD degree may be admitted to a professional doctorate or professional master’s degree program if there is no duplication of training involved. Applicants may apply only to one single degree program or one concurrent degree program per admission cycle. Required Documents for Applications 1. Transcripts: Applicants may upload unofficial transcripts with your application for the departmental initial review. Unofficial transcripts must contain specific information including the name of the applicant, name of the school, all courses, grades, units, & degree conferral (if applicable). 2. Letters of recommendation: Applicants may request online letters of recommendation through the online application system. Hard copies of recommendation letters must be sent directly to the program, by the recommender, not the Graduate Admissions. 3. Evidence of English language proficiency: All applicants who have completed a basic degree from a country or political entity in which the official language is not English are required to submit official evidence of English language proficiency. This applies to institutions from Bangladesh, Burma, Nepal, India, Pakistan, Latin America, the Middle East, the People’s Republic of China, Taiwan, Japan, Korea, Southeast Asia, most European countries, and Quebec (Canada). However, applicants who, at the time of application, have already completed at least one year of full-time academic course work with grades of B or better at a US university may submit an 2 Information and Data Science: MIDS official transcript from the US university to fulfill this requirement. The following courses will not fulfill this requirement: • courses in English as a Second Language, • courses conducted in a language other than English, • courses that will be completed after the application is submitted, and • courses of a non-academic nature. Applicants who have previously applied to Berkeley must also submit new test scores that meet the current minimum requirement from one of the standardized tests. Official TOEFL score reports must be sent directly from Educational Test Services (ETS). The institution code for Berkeley is 4833 for Graduate Organizations. Official IELTS score reports must be sent electronically from the testing center to University of California, Berkeley, Graduate Division, Sproul Hall, Rm 318 MC 5900, Berkeley, CA 94720. TOEFL and IELTS score reports are only valid for two years prior to beginning the graduate program at UC Berkeley. Note: score reports can not expire before the month of June. Where to Apply Visit the Berkeley Graduate Division application page (http:// grad.berkeley.edu/admissions/apply/). Admission to the Program Applications are evaluated holistically on a combination of prior academic performance, work experience, essays, letters of recommendation, and goals that are a good fit for the program. The UC Berkeley School of Information seeks students with the academic abilities to meet the demands of a rigorous graduate program. To be eligible to apply to the Master of Information and Data Science program, applicants must meet the following requirements: • A bachelor’s degree or its recognized equivalent from an accredited institution. • Superior scholastic record, normally well above a 3.0 GPA. • A high level of quantitative ability as conveyed by significant work experience that demonstrates your quantitative abilities and/or academic coursework that demonstrates quantitative aptitude • A high level of analytical reasoning ability and a problem-solving mindset as demonstrated in academic and/or professional performance. • A working knowledge of fundamental concepts including: data structures, algorithms and analysis of algorithms, and linear algebra. • Proficiency in programming languages, such as Python or Java, demonstrated by prior work experience or advanced coursework. Applicants who lack this experience in their academic or work background but meet all other admission requirements will be required to take the Introduction to Data Science Programming course in their first term. • The ability to communicate effectively, as demonstrated by academic performance, professional experience, and/or strong essays that demonstrate effective communication skills. • Not Required: Official Graduate Record Examination (GRE) (http://www.princetonreview.com/mids/) General Test or Graduate Management Admission Test (GMAT) (http:// www.princetonreview.com/mids/) scores. As of Fall 2020, we have eliminated the GRE/GMAT requirement. We recommend you put your time and effort towards the required application materials. • Official Test of English as a Foreign Language (TOEFL) (http:// www.toefl.org/) scores for applicants whose academic work has been in a country other than the US, UK, Australia, or English-speaking Canada. Note: Admission to the Fifth Year Master of Information and Data Science program requires that the applicant complete their undergraduate education at UC Berkeley in the year prior to starting the program. Consequently, applicants are not required to submit TOEFL scores. However, applicants are required to submit three letters of recommendation and additional short answer essays. For more information and application instructions, prospective MIDS students should visit the datascience@berkeley Admissions Overview (http://datascience.berkeley.edu/admissions/admissions-overview/) and prospective 5th Year MIDS students should visit the I School Admissions page (https://www.ischool.berkeley.edu/programs/5th-year-mids/apply/). Unit Requirements The Master of Information and Data Science is designed to be completed in 20 months, but other options are available to complete the program. You will complete 27 units of course work over an average of five terms, taking a maximum of 9 units each term. Courses are divided into foundation courses (15 units), advanced courses (9 units), and a synthetic capstone (3 units). Students also complete an Immersion Program. The unit and coursework requirements for the Fifth Year Master of Information and Data Science are identical, although fifth-year students are also encouraged and supported to seek internships to complement and encourage professional development. Curriculum Foundation Courses DATASCI 200 Introduction to Data Science Programming 3 DATASCI 201 Research Design and Applications for Data and Analysis 3 DATASCI 201A Research Design and Applications for Data and Analysis for Early Career Data Scientists 4 DATASCI 203 Statistics for Data Science 3 DATASCI 205 Fundamentals of Data Engineering 3 DATASCI 207 Applied Machine Learning 3 Advanced Courses DATASCI 209 Data Visualization 3 DATASCI 231 Behind the Data: Humans and Values 3 DATASCI 233 Privacy Engineering 3 DATASCI 241 Experiments and Causal Inference 3 DATASCI 251 Deep Learning in the Cloud and at the Edge 3 DATASCI 255 Machine Learning Systems Engineering 3 DATASCI 261 Machine Learning at Scale 3 DATASCI 266 Natural Language Processing with Deep Learning 3 DATASCI 271 Statistical Methods for Discrete Response, Time Series, and Panel Data 3 DATASCI 281 Computer Vision 3 DATASCI 290 Special Topics 3 Information and Data Science: MIDS 5 DATASCI 203 Statistics for Data Science 3 Units Terms offered: Not yet offered This course provides students with a foundational understanding of classical statistics within the broader context of data science. Topics include exploratory analysis and descriptive statistics, probability theory and the foundations of statistical modeling, estimators, hypothesis testing, and classical linear regression. Causal inference and reproducibility issues are treated briefly. Students will learn to apply the most common statistical procedures correctly, checking assumptions and responding appropriately when they appear violated; to evaluate the design of a study and how the variables being measured relate to research questions; and to analyze real-world data using the open-source language R. Statistics for Data Science: Read More [+] Rules & Requirements Prerequisites: MIDS students only. Intermediate competency in calculus is required. A college-level linear algebra course is recommended Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Formerly known as: Data Science W203 Statistics for Data Science: Read Less [-] DATASCI 205 Fundamentals of Data Engineering 3 Units Terms offered: Not yet offered Storing, managing, and processing datasets are foundational processes in data science. This course introduces the fundamental knowledge and skills of data engineering that are required to be effective as a data scientist. This course focuses on the basics of data pipelines, data pipeline flows and associated business use cases, and how organizations derive value from data and data engineering. As these fundamentals of data engineering are introduced, learners will interact with data and data processes at various stages in the pipeline, understand key data engineering tools and platforms, and use and connect critical technologies through which one can construct storage and processing architectures that underpin data science applications. Fundamentals of Data Engineering: Read More [+] Rules & Requirements Prerequisites: MIDS students only. Intermediate competency in Python, C, or Java, and competency in Linux, GitHub, and relevant Python libraries. Knowledge of database management including SQL is recommended but not required Credit Restrictions: Students will receive no credit for DATASCI W205 after completing DATASCI 205. A deficient grade in DATASCI W205 may be removed by taking DATASCI 205. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Instructor: Crook Formerly known as: Data Science W205 Fundamentals of Data Engineering: Read Less [-] 6 Information and Data Science: MIDS DATASCI 207 Applied Machine Learning 3 Units Terms offered: Not yet offered Machine learning is a rapidly growing field at the intersection of computer science and statistics concerned with finding patterns in data. It is responsible for tremendous advances in technology, from personalized product recommendations to speech recognition in cell phones. This course provides a broad introduction to the key ideas in machine learning. The emphasis will be on intuition and practical examples rather than theoretical results, though some experience with probability, statistics, and linear algebra will be important. Course must be taken for a letter grade to fulfill degree requirements. Applied Machine Learning: Read More [+] Rules & Requirements Prerequisites: MIDS students only. DATASCI W201 and DATASCI W203. Intermediate competency in Python, C, or Java, and competency in Linux, GitHub, and relevant Python libraries; or permission of instructor. Linear algebra is recommended Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Instructor: Gillick Formerly known as: Data Science W207 Applied Machine Learning: Read Less [-] DATASCI 209 Data Visualization 3 Units Terms offered: Not yet offered Visualization enhances exploratory analysis as well as efficient communication of data results. This course focuses on the design of visual representations of data in order to discover patterns, answer questions, convey findings, drive decisions, and provide persuasive evidence. The goal is to give you the practical knowledge you need to create effective tools for both exploring and explaining your data. Exercises throughout the course provide a hands-on experience using relevant programming libraries and software tools to apply research and design concepts learned. Data Visualization: Read More [+] Objectives & Outcomes Student Learning Outcomes: Analyze data using exploratory visualization. Build commonly requested types of visualizations as well as more advanced visualizations using ground-up customization. Constructively critique existing visualizations, identifying issues of integrity as well as excellence. Create useful, performant visualizations from real-world data sources, including large and complex datasets. Design aesthetically pleasing static and interactive visualizations with perceptually appropriate forms and encodings. Improve your own work through usability testing and iteration, with attention to context. Select appropriate tools for building visualizations, and gain skills to evaluate new tools. Rules & Requirements Prerequisites: MIDS students only. DATASCI W203. Students must take DATASCI W205 concurrently or prior to DATASCI W209. If taken concurrently, students may not drop W205 and remain in W209. Recommended: experience with HTML, CSS, and JavaScript, or ability to learn new programming languages quickly. If Python is the only programming language you know, you will probably benefit from learning the basics of web development with JavaScript in advance Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Formerly known as: Data Science W209 Data Visualization: Read Less [-] Information and Data Science: MIDS 7 DATASCI 210 Capstone 3 Units Terms offered: Not yet offered The capstone course will cement skills learned throughout the MIDS program – both core data science skills and “soft skills” like problem- solving, communication, influencing, and management – preparing students for success in the field. The centerpiece is a semester-long group project in which teams of students propose and select project ideas, conduct and communicate their work, receive and provide feedback (in informal group discussions as well as formal class presentations), and deliver compelling presentations along with a Web- based final deliverable. Includes relevant readings, case discussions, and real-world examples and perspectives from panel discussions with leading data science experts and industry practitioners. Capstone: Read More [+] Rules & Requirements Prerequisites: MIDS students only. Must be taken in final term of the MIDS program Credit Restrictions: Students will receive no credit for DATASCI W210 after completing DATASCI 210. A deficient grade in DATASCI W210 may be removed by taking DATASCI 210. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Formerly known as: Data Science W210 Capstone: Read Less [-] DATASCI 231 Behind the Data: Humans and Values 3 Units Terms offered: Not yet offered Intro to the legal, policy, and ethical implications of data, including privacy, surveillance, security, classification, discrimination, decisional- autonomy, and duties to warn or act. Examines legal, policy, and ethical issues throughout the full data-science life cycle — collection, storage, processing, analysis, and use — with case studies from criminal justice, national security, health, marketing, politics, education, employment, athletics, and development. Includes legal and policy constraints and considerations for specific domains and data-types, collection methods, and institutions; technical, legal, and market approaches to mitigating and managing concerns; and the strengths and benefits of competing and complementary approaches. Behind the Data: Humans and Values: Read More [+] Rules & Requirements Prerequisites: MIDS and MPA students only Credit Restrictions: Students will receive no credit for DATASCI W231 after completing DATASCI 231. A deficient grade in DATASCI W231 may be removed by taking DATASCI 231. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Instructor: Morgan Formerly known as: Data Science W231 Behind the Data: Humans and Values: Read Less [-] 10 Information and Data Science: MIDS DATASCI 261 Machine Learning at Scale 3 Units Terms offered: Not yet offered This course teaches the underlying principles required to develop scalable machine learning pipelines for structured and unstructured data at the petabyte scale. Students will gain hands-on experience in Apache Hadoop and Apache Spark. Machine Learning at Scale: Read More [+] Rules & Requirements Prerequisites: MIDS students only. DATASCI W205 and DATASCI W207. Intermediate programming skills in an object-oriented language (e.g., Python) Credit Restrictions: Students will receive no credit for DATASCI W261 after completing DATASCI 261. A deficient grade in DATASCI W261 may be removed by taking DATASCI 261. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Formerly known as: Data Science W261 Machine Learning at Scale: Read Less [-] DATASCI 266 Natural Language Processing with Deep Learning 3 Units Terms offered: Not yet offered Understanding language is fundamental to human interaction. Our brains have evolved language-specific circuitry that helps us learn it very quickly; however, this also means that we have great difficulty explaining how exactly meaning arises from sounds and symbols. This course is a broad introduction to linguistic phenomena and our attempts to analyze them with machine learning. We will cover a wide range of concepts with a focus on practical applications such as information extraction, machine translation, sentiment analysis, and summarization. Natural Language Processing with Deep Learning: Read More [+] Rules & Requirements Prerequisites: MIDS students only. DATASCI W207 Credit Restrictions: Students will receive no credit for DATASCI W266 after completing DATASCI 266. A deficient grade in DATASCI W266 may be removed by taking DATASCI 266. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Instructor: Gillick Formerly known as: Data Science W266 Natural Language Processing with Deep Learning: Read Less [-] Information and Data Science: MIDS 11 DATASCI 271 Statistical Methods for Discrete Response, Time Series, and Panel Data 3 Units Terms offered: Not yet offered A continuation of Data Science W203 (Exploring and Analyzing Data), this course trains data science students to apply more advanced methods from regression analysis and time series models. Central topics include linear regression, causal inference, identification strategies, and a wide-range of time series models that are frequently used by industry professionals. Throughout the course, we emphasize choosing, applying, and implementing statistical techniques to capture key patterns and generate insight from data. Students who successfully complete this course will be able to distinguish between appropriate and inappropriate techniques given the problem under consideration, the data available, and the given timeframe. Statistical Methods for Discrete Response, Time Series, and Panel Data: Read More [+] Rules & Requirements Prerequisites: MIDS students only. DATASCI W203 taken in Fall 2016 or later and completed with a grade of B+ or above; strong familiarity with classical linear regression modeling; strong hands-on experience in R; working knowledge of calculus and linear algebra; familiarity with differential calculus, integral calculus and matrix notations Credit Restrictions: Students will receive no credit for DATASCI W271 after completing DATASCI 271. A deficient grade in DATASCI W271 may be removed by taking DATASCI 271. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Formerly known as: Data Science W271 Statistical Methods for Discrete Response, Time Series, and Panel Data: Read Less [-] DATASCI 281 Computer Vision 3 Units Terms offered: Summer 2022, Spring 2022 This course introduces the theoretical and practical aspects of computer vision, covering both classical and state of the art deep-learning based approaches. This course covers everything from the basics of the image formation process in digital cameras and biological systems, through a mathematical and practical treatment of basic image processing, space/ frequency representations, classical computer vision techniques for making 3-D measurements from images, and modern deep-learning based techniques for image classification and recognition. Computer Vision: Read More [+] Objectives & Outcomes Student Learning Outcomes: Be able to read and understand research papers in the computer-vision literature. Build computer vision systems to solve real-world problems. Properly formulate problems with the appropriate mathematical and computational tools. Understand the building blocks of classical computer vision techniques. Understand the building blocks of modern computer vision techniques (primarily artificial neural networks). Understand the process by which images are formed and represented. Rules & Requirements Prerequisites: MIDS students only. DATASCI W207 Applied Machine Learning: We assume you are familiar with machine learning techniques. Linear Algebra: You should also be comfortable with linear algebra, which we'll use for vector representations and when we discuss deep learning. Language: This course will use Python for all examples, exercises, and assignments Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Computer Vision: Read Less [-] 12 Information and Data Science: MIDS DATASCI 290 Special Topics 3 Units Terms offered: Fall 2021 Specific topics, may vary from section to section, year to year. Special Topics: Read More [+] Rules & Requirements Prerequisites: MIDS students only Repeat rules: Course may be repeated for credit when topic changes. Students may enroll in multiple sections of this course within the same semester. Hours & Format Fall and/or spring: 14 weeks - 3 hours of lecture per week Summer: 14 weeks - 3 hours of lecture per week Additional Details Subject/Course Level: Data Science/Graduate Grading: Letter grade. Special Topics: Read Less [-]