C01/C51 Modeling with Machine Learning: from Algorithms to Applications

Focuses on modeling with machine learning methods with an eye towards applications in engineering and sciences. Students will be introduced to modern machine learning methods, from supervised to unsupervised models, with an emphasis on newer neural approaches. Develops understanding of how and why the methods work from the point of view of modeling, and when they are applicable. Using concrete examples, covers formulation of machine learning tasks, adapting and extending methods to given problems, and how the methods can and should be evaluated. Students taking graduate version complete additional assignments. Enrollment limited.

Emphasizes the design and operation of sustainable systems. Illustrates how to leverage heterogeneous data from urban services, cities, and the environment, and apply machine learning methods to evaluate and/or improve sustainability solutions. Provides case studies from various domains, such as transportation and urban mobility, energy and water resources, environmental monitoring, infrastructure sensing and control, climate adaptation, and disaster resilience. Projects focus on using machine learning to identify new insights or decisions that can help engineer sustainability in societal-scale systems. Students taking graduate version complete additional assignments.

Encourages open-ended exploration of the increasingly topical intersection between artificial intelligence and the physical sciences. Uses energy and information, and their respective optimality conditions, to define supervised and unsupervised learning algorithms as well as ordinary and partial differential equations. Subsequently, physical systems with complex constitutive relationships are drawn from elasticity, biophysics, fluid mechanics, hydrodynamics, acoustics, and electromagnetics to illustrate how machine learning-inspired optimization can approximate solutions to forward and inverse problems in these domains. Students taking graduate version complete additional assignments.

Provides an introduction to the use of machine learning to solve problems arising in the science and engineering of biology, chemistry, and materials. Equips students to design and implement machine learning approaches to challenges such as analysis of omics (genomics, transcriptomics, proteomics, etc.), microscopy, spectroscopy, or crystallography data and design of new molecules and materials such as drugs, catalysts, polymer, alloys, ceramics, and proteins. Students taking graduate version complete additional assignments.

Focuses on in-depth modeling of engineering tasks as machine learning problems. Emphasizes framing, method design, and interpretation of results. In comparison to broader co-requisite 6.C01/6.C51, this project-oriented subject consists of deep dives into select technical areas or engineering tasks involving both supervised and exploratory uses of machine learning. Deep dives into technical areas, such as robustness, interpretability, privacy or causal discovery; engineering tasks such as recommender systems, performance optimization, or automated design.

Introduces machine learning as a tool to understand natural biological systems, with an evolving emphasis on problems in molecular and cellular biology that are being actively advanced using machine learning. Students design, implement, and interpret machine learning approaches to aid in predicting protein structure, probing protein structure/function relationships, and imaging biological systems at scales ranging from the atomic to cellular.

Explores applications of machine learning techniques to solve problems in modern finance. Provides an introduction to financial models that balance risk and reward, along with machine learning tools to uncover and analyze financial regularities. Applications include valuation, credit analysis, proprietary trading and hedge-fund strategies, portfolio management, market structure, risk management and stress testing, natural language processing, and personal finance.

Focuses on applying various machine learning techniques to a broad range of topics which are of core value in modern nuclear science and engineering. Relevant topics include machine learning on fusion and plasma diagnosis, reactor physics and nuclear fission, nuclear materials properties, quantum engineering and nuclear materials, and nuclear security. Special components center on the additional machine learning architectures that are most relevant to a certain field, the implementation, and picking up the right problems to solve using a machine learning approach. Final project dedicated to the field-specific applications. Students taking graduate version complete additional assignments.

Building upon 6.C51, students will design and develop applications targeted to solve key problems in supply chain using state-of-art machine learning methods and algorithms. Special emphasis will be given to the selection of the appropriate ML methods for the supply chain problems addressed, as well as requirements for code productionalization and code efficiency, both key in real industry applications.