Course Description

Covers advanced statistical machine learning methods, including high-dimensional data, Bayesian approaches, and deep learning. Topics include sparse and graphical models, neural networks, advanced architectures, generative methods, reinforcement learning, and ethical use of machine learning. Students gain experience with modern machine learning frameworks and real-world model deployment.

---

Athena Title

Adv Stat ML for Data Sci

---

Non-Traditional Format

This course will be taught 95% or more online.

---

Prerequisite

STAT 6384E and STAT 6385E

---

Semester Course Offered

Offered fall

---

Grading System

A - F (Traditional)

---

Student learning Outcomes

• Students will analyze high-dimensional data and apply sparse modeling and nonlinear dimensionality reduction techniques.
• Students will implement probabilistic and Bayesian machine learning methods, including graphical models, and apply approximate inference.
• Students will design and optimize deep learning architectures using modern machine learning frameworks.
• Students will apply advanced deep learning models such as autoencoders, variational methods, and generative approaches.
• Students will develop reinforcement learning solutions using Markov Decision Processes, Q-learning, and deep reinforcement methods.
• Students will evaluate strategies for handling missing data, including imputation and model-based approaches.
• Students will assess machine learning models with respect to interpretability, fairness, and causal reasoning.

---

Topical Outline

• I. Recap of Supervised and Unsupervised Learning a. Review of classification and clustering paradigms b. Bias–variance tradeoff and model complexity c. Recap of regularization methods
• II. High-Dimensional and Sparse Models a. Curse of dimensionality and modeling challenges b. Sparse modeling approaches c. Feature selection techniques for high-dimensional settings d. Nonlinear dimensionality reduction (t-SNE, UMAP)
• III. Probabilistic and Bayesian Machine Learning a. Bayesian inference: priors, likelihoods, posteriors b. Probabilistic graphical models (Bayesian networks, HMMs) c. Approximate inference (MCMC, Gibbs sampling, Variational Inference) d. Gaussian Processes for regression and classification e. Mixture models and Bayesian clustering
• IV. Deep Learning Foundations a. Fundamentals of neural networks and backpropagation b. Optimization methods (SGD, Adam, learning rate schedules) c. Regularization strategies (dropout, early stopping, batch normalization) d. Core architectures (CNNs, RNNs/LSTMs, Transformers) e. Frameworks for implementation (PyTorch, TensorFlow)
• V. Advanced Deep Learning and Modern AI a. Autoencoders and variational autoencoders b. Generative adversarial networks c. Semi-supervised and self-supervised learning approaches d. Emerging frontier methods (e.g., diffusion models) e. Interpretability and fairness (SHAP, LIME, influence functions) f. Causal inference methods (DAGs, counterfactuals, propensity models) g. Deployment strategies such as Flask, Docker, and AutoML pipelines
• VI. Reinforcement Learning a. Introduction to reinforcement learning paradigms b. Markov Decision Processes c. Value functions and Q-learning d. Policy gradient methods e. Deep reinforcement learning (Deep Q-Networks) f. Applications in data science

---

Institutional Competencies Learning Outcomes

Analytical Thinking

The ability to reason, interpret, analyze, and solve problems from a wide array of authentic contexts.

---

Syllabus

Select Faculty / Instructor to download syllabus:   Basnet, Man

---

Public CV

Select a file to download:   Basnet, Man

---