61.502 Deep Learning for Enterprise

Course description

This course serves as a pivotal bridge between foundational machine learning concepts and more advanced specialty topics in AI. It introduces students to the core principles and applications of deep learning through hands-on experience with popular frameworks, the most popular currently being PyTorch. Students will learn to implement, train, and fine-tune various deep learning architectures from scratch. These architectures include deep linear neural networks, convolutional neural networks (CNNs), recurrent neural networks (RNNs), and simple transformers, applying them to real-world tasks like image recognition, natural language processing, and sequence modelling.

The curriculum emphasises practical skills in handling data, selecting appropriate layers for models, and optimising neural network architectures, and using GPUs for efficient training. Additionally, the course introduces generative approaches including VAEs, GANs, and diffusion models, as well as cutting-edge topics like graph neural networks, self-supervised learning, semi-supervised learning, multimodal AI, and AI security.

Through project-based learning, students will integrate and apply their knowledge from earlier courses in machine learning, preparing them for subsequent specialised courses on topics such as Natural Language Processing, Computer Vision, and Reinforcement Learning.

Instructor

Pritee Agrawal

Number of credits: 12