DS

Understand Deep Learning covering neural networks, activation functions, and optimization algorithms. Gain experience with TensorFlow and PyTorch, mastering key techniques such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Generative Adversarial Networks (GANs). Explore transfer learning, reinforcement learning, and natural language processing (NLP), along with industry applications and ethical considerations.

Explores new models of database design: graph, vector, and ledger. These have become required infrastructure in service of social media (graph databases), Large Language Models (vector databases), and cryptocurrency (ledger databases). Will learn their basic operations with an eye toward other purposes as well as the key advantages and drawbacks of these data models. Center on student projects built using one of these databases.

Comprehensive exploration of the multifaceted aspects of data creation, emphasizing the symbiotic relationship between design and data. Students will gain insight into the intentional and unintentional mechanisms that contribute to data creation, including human input, technological processes, environmental factors, and systemic influences.

Introduces complex interplay between technology, regulation, and data science and exposes regulatory realities confronting the field. Read and parse regulatory texts. Navigate the international technology regulatory landscape, identify key actors, and appreciate how rules governing different kinds of data, platforms, copyright and intellectual property, and digital services and markets shape data science and AI/ML development practices.

This course provides selected special topics in data science.

Reading and research under the direction of a faculty member. Students must obtain approval from a faculty advisor to approve and direct the independent study. Final approval by the Director of Undergraduate Programs is also required.

Covers foundations and applications of NLP with a focus on the most popular form of unstructured data - text. Convert source texts into structure-preserving analytical form and then apply information theory, NLP tools, and vector-based methods to explore language models, topic models, sentiment analyses, and GenAI techniques. Focus is on unsupervised methods to explore cognitive patterns in texts, with real-world examples and demonstrations.

Train your own LLM for a custom task. Learn about the LLM lifecycle from architecture, to pre-training, to supervised finetuning, to deployment, to model editing/updating, including discussing LLM limitations. End up with your own trained LLM, a HuggingFace model card you can show off in technical interviews, and a plan for how to stay up to date with this fast-moving field.

Provides healthcare domain knowledge, healthcare data understanding, and data science methodologies to solve problems. Understand data types, models, and sources, including electronic health record data; health outcomes, quality, risk, and safety data; and unstructured data, such as clinical text data; biomedical sensor data; and biomedical image data. Querying with SQL, data visualization with Tableau, and analysis and prediction with Python.

Reinforcement Learning is a dynamic area in machine learning that allows an agent to learn by interacting with its environment. This enables learning when the ground truth is unavailable or outdated (think predicting Netflix usage before and during a pandemic). This course will introduce topics including k-armed bandits, Markov Decision Processes, value functions, Policy Gradients, Q-Learning, and deep Q-Learning.

This course looks into the past, present, and future of technologies that impact labor, with an eye to empowering students with knowledge about the social, economic, and political dimensions of the tools they use both inside and outside of work. The course covers labor history, whistleblowers, and hidden histories of common technologies that reorient common assumptions about what technologies can do, and what they have done in the past.

Provides a foundation in discrete mathematics, data structures, algorithmic design and implementation, computational complexity, parallel computing, and data integrity and consistency. Case studies and exercises will be drawn from real-world examples (e.g., bioinformatics, public health, marketing, and security).

Principles of interactivity in application and dashboard development using R, Python, and JavaScript programming languages. Design visually appealing and user-friendly interfaces, develop interactive applications for data visualization, and build dynamic dashboards for effective data communication with end-users. Covers theoretical concepts and hands-on implementation to provide a comprehensive understanding of the full design process at the graduate level.

Provides an in-depth exploration of probabilistic and statistical methods used to understand, quantify, and manage uncertainty. Learn foundational concepts in probability and statistics, simulation techniques, and modern approaches to parameter estimation, decision theory, and hypothesis testing. Topics include parametric and nonparametric methods, Bayesian and frequentist paradigms, and applications of uncertainty in real-world problems. 

Equips students with some of the most used deep learning architectures. Explore feed-forward networks, convolutional neural networks, UNETs, encoders-decoders, generative adversarial networks and transformers. Analyze tools of explainable AI. Focused on climate applications, apply these techniques to real-world data, solving problems in prediction, pattern recognition, and data-driven insights.

This course will equip students with some of the most commonly used deep learning architectures. We will explore feed-forward networks, convolutional neural networks, UNETs, encoders-decoders, generative adversarial networks and transformers. We will also analyze tools of explainable AI. Focused on environmental applications, students will apply these techniques to real-world data, solving problems in prediction, pattern recognition, and data-driven insights. Solid background in probability, statistics, and in coding (preferably Python) is recommended for enrollment in this course.

Provides domain knowledge of humanitarian action, the data and data systems used by humanitarian organizations, the lifesaving implications of predictive analytics in disaster contexts and the complex data ethics considerations unique to the social impact sector. Communicating with clients/stakeholders/decision makers about resource allocation strategy and critically evaluating tools, data sources and methods are key features of the course.

Genomics Foundations introduces core concepts in modern genomics and human genetics underlying computational biology and public health genomics. The course integrates key biological principles with quantitative reasoning and hands-on use of real genomics data and databases.

An introduction to essential programming concepts, structures, and techniques. Students will gain confidence in not only reading code, but learning what it means to write good quality code. Additionally, essential and complementary topics are taught, such as testing and debugging, exception handling, and an introduction to visualization. This course is project based, consisting of a semester project and final project presentations.

Trends in hardware and software for Big Data Systems and applications. Cover principles driving data infrastructures, which enabled the training of AI models on datasets (speech, sounds, images, video, languages) and may extend to structured data (text, images, time series). AI and machine learning practitioners build and deploy data science projects on Amazon Web Services unifying data science, data engineering, and application development.

Code an end-to-end data science project with core software engineering and automation to quickly integrate into a corporate environment. Use version control to focus on solutions, leverage automation at your command line and in the cloud, deliver solid code by incorporating testing, lower extension and maintenance time with OOP and Design Patterns, ensuring your code's path to production to deliver a complete package to the enterprise.

An intensive overview of cloud infrastructure and their role in data science. Topics will include storage as a service, ephemeral computing resources, auto-scaling, and event-driven workloads. Special attention will be paid to cloud-native design patterns, which are built assuming the unique functionality of cloud computing resources.

This course focuses on making students more effective at identifying and designing AI use cases to create novel AI-powered products and services. Students will work with a variety of AI technologies across several projects. They will gain a deep understanding of design considerations for incorporating AI into products in ways that create value for users and businesses.

Focuses on the application of data science to critical problems and opportunities in business. You will learn business concepts in strategy, markets and competition, and will apply data science to analytical projects in operations, marketing, human resources and finance. Additional topics include experimentation, business cases, team leadership and executive communication. Students will use Python or R, and Dataiku DSS.

This course provides selected special topics in data science to graduate and undergraduate students.