jeffheaton/app_deep_learning

T81-558: PyTorch - Applications of Deep Neural Networks @Washington University in St. Louis

T81 558:Applications of Deep Neural Networks

Washington University in St. Louis

Instructor: Jeff Heaton

• Section 1. Spring 2024, Tuesday, 2:30 PM, Location: Cupples I / 215

Course Description

Deep learning is a group of exciting new technologies for neural networks. Through a combination of advanced training techniques and neural network architectural components, it is now possible to create neural networks that can handle tabular data, images, text, and audio as both input and output. Deep learning allows a neural network to learn hierarchies of information in a way that is like the function of the human brain. This course will introduce the student to classic neural network structures, Convolution Neural Networks (CNN), Long Short-Term Memory (LSTM), Gated Recurrent Neural Networks (GRU), General Adversarial Networks (GAN) and reinforcement learning. Application of these architectures to computer vision, time series, security, natural language processing (NLP), and data generation will be covered. High Performance Computing (HPC) aspects will demonstrate how deep learning can be leveraged both on graphical processing units (GPUs), as well as grids. Focus is primarily upon the application of deep learning to problems, with some introduction to mathematical foundations. Students will use the Python programming language to implement deep learning using PyTorch. It is not necessary to know Python prior to this course; however, familiarity of at least one programming language is assumed. This course will be delivered in a hybrid format that includes both classroom and online instruction.

Objectives

1. Explain how neural networks (deep and otherwise) compare to other machine learning models.
2. Determine when a deep neural network would be a good choice for a particular problem.
3. Demonstrate your understanding of the material through a final project uploaded to GitHub.

Syllabus

This syllabus presents the expected class schedule, due dates, and reading assignments. Download current syllabus

Module	Content
Module 1 Meet on 08/27/2024	Module 1: Python Preliminaries 1.1: Course Overview 1.2: Introduction to Python 1.3: Python Lists, Dictionaries, Sets & JSON 1.4: File Handling 1.5: Functions, Lambdas, and Map/ReducePython Preliminaries We will meet on campus this week! (first meeting)
Module 2 Week of 09/03/2024	Module 2: Python for Machine Learning 2.1: Introduction to Pandas for Deep Learning 2.2: Encoding Categorical Values in Pandas 2.3: Grouping, Sorting, and Shuffling 2.4: Using Apply and Map in Pandas 2.5: Feature Engineering in Padas Module 1 Program due: 09/04/2024 Icebreaker due: 09/04/2024
Module 3 Week of 09/10/2024	Module 3: PyTorch for Neural Networks 3.1: Deep Learning and Neural Network Introduction 3.2: Introduction to PyTorch 3.3: Encoding a Feature Vector for PyTorch Deep Learning 3.4: Early Stopping and Network Persistence 3.5: Sequences vs Classes in PyTorch Module 2: Program due: 09/11/2024
Module 4 Week of 09/17/2024	Module 4: Training for Tabular Data 4.1: Using K-Fold Cross-validation with PyTorch 4.2: Training Schedules for PyTorch 4.3: Dropout Regularization 4.4: Batch Normalization 4.5: RAPIDS for Tabular Data Module 3 Program due: 09/18/2024
Module 5 Week of 09/24/2024	Module 5: CNN and Computer Vision 5.1 Image Processing in Python 5.2 Using Convolutional Neural Networks 5.3 Using Pretrained Neural Networks 5.4 Looking at Generators and Image Augmentation 5.5 Recognizing Multiple Images with YOLO Module 4 Program due: 09/25/2024
Module 6 Meet on 10/01/2024	Module 6: ChatGPT and Large Language Models 6.1: Introduction to Transformers 6.2: Accessing the ChatGPT API 6.3: LLM Memory 6.4: Introduction to Embeddings 6.5: Prompt Engineering Module 5 Program due: 10/02/2024 We will meet on campus this week! (second meeting)
Module 7 Week of 10/15/2024	Module 7: Image Generative Models 7.1: Introduction to Generative AI 7.2: Generating Faces with StyleGAN3 7.3: GANS to Enhance Old Photographs Deoldify 7.4: Text to Images with StableDiffusion 7.5: Finetuning with Dreambooth Module 6 Program due: 10/16/2024
Module 8 Meet on 10/22/2024	Module 8: Kaggle 8.1 Introduction to Kaggle 8.2 Building Ensembles with Scikit-Learn and PyTorch 8.3 How Should you Architect Your PyTorch Neural Network: Hyperparameters 8.4 Bayesian Hyperparameter Optimization for PyTorch 8.5 Current Semester's Kaggle Module 7 Program due: 10/23/2024 We will meet on campus this week! (third meeting)
Module 9 Week of 10/29/2024	Module 9: Facial Recognition 9.1 Detecting Faces in an Image 9.2 Detecting Facial Features 9.3 Image Augmentation 9.4 Application: Emotion Detection 9.5 Application: Blink Efficiency Module 8 Assignment due: 10/30/2024
Module 10 Week of 11/05/2024	Module 10: Time Series in PyTorch Time Series Data Encoding for Deep Learning, PyTorch Seasonality and Trend LSTM-Based Time Series with PyTorch CNN-Based Time Series with PyTorch Predicting with Meta Prophet Module 9 Program due: 11/06/2024
Module 11 Week of 11/12/2024	Module 11: Natural Language Processing 11.1 Introduction to Natural Language Processing 11.2 Hugging Face Introduction 11.3 Hugging Face Tokenizers 11.4 Hugging Face Data Sets 11.5 Training a Model in Hugging Face Module 10 Program due: 11/13/2024
Module 12 Week of 11/19/2024	Module 12: Reinforcement Learning Introduction to Gymnasium Introduction to Q-Learning Stable Baselines Q-Learning Atari Games with Stable Baselines Neural Networks Future of Reinforcement Learning
Module 13 Week of 11/26/2024	Module 13: Deployment and Monitoring 13.1: Using Denoising AutoEncoders 13.2: Anomaly Detection 13.3: Model Drift and Retraining 13.4: Tensor Processing Units (TPUs) 13.5: Future Directions in Artificial Intelligence Kaggle Assignment due: 11/26/2024 (approx 4-6PM, due to Kaggle GMT timezone)
Week 14 Meet on 12/3/2024	Week 14: Kaggle Presentations Top Kaggle teams will present We will meet on campus this week! (fourth meeting) Final project due: 12/4/2024

Datasets

• Datasets can be downloaded here