Alexander Amini Courses

MIT: Recurrent Neural Networks

2 ratings

Explore recurrent neural networks for sequence modeling, including long-term dependencies, LSTMs, and attention mechanisms.

• YouTube
• 37 minutes
• Self-Paced
• Free Video

Neurosymbolic AI

2 ratings

Explores how combining neural networks with symbolic reasoning can improve AI's out-of-distribution performance, compositional understanding, and data efficiency.

• YouTube
• 41 minutes
• Self-Paced
• Free Video

Deep Generative Modeling

1 rating

Explores how deep neural networks learn latent structure from unlabeled data to generate synthetic examples, including VAEs, GANs, and unpaired domain translation.

• YouTube
• 1 hour 2 minutes
• Self-Paced
• Free Video

MIT 6.S191 - Automatic Speech Recognition

Explore end-to-end speech recognition with mel-scale features, subword units, attention-based encoder-decoders, CTC decoding, language models, and bias-aware data strategies.

• YouTube
• 42 minutes
• Self-Paced
• Free Video

MIT: Reinforcement Learning

Explores how deep neural networks learn through interaction with dynamic environments, using Q-learning and policy gradients for tasks such as gameplay.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

Deep Generative Modeling

Explore how variational autoencoders and generative adversarial networks learn latent distributions to synthesize and transform data.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

Convolutional Neural Networks

Explains how convolutional neural networks learn spatial image features for computer vision tasks including classification, segmentation, and autonomous navigation.

• YouTube
• 49 minutes
• Self-Paced
• Free Video

Recurrent Neural Networks and Transformers

Learn how recurrent networks and attention mechanisms model sequential data, from backpropagation through time to transformer applications.

• YouTube
• 58 minutes
• Self-Paced
• Free Video

MIT Introduction to Deep Learning

An MIT lecture builds deep learning from perceptrons into neural networks, then trains them with loss functions, gradient descent, backpropagation, and regularization.

• YouTube
• 49 minutes
• Self-Paced
• Free Video

AI in Healthcare

An introduction to how deep learning can expand healthcare access while addressing bias, uncertainty, model limitations, and patient safety.

• YouTube
• 34 minutes
• Self-Paced
• Free Video

Towards AI for 3D Content Creation

Explores how deep learning can synthesize 3D worlds, objects, medical data, and simulations through differentiable rendering and neural models.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Taming Dataset Bias via Domain Adaptation

Learn how domain adaptation transfers deep-learning models across shifted datasets using unlabeled target data, adversarial alignment, pixel translation, self-supervision, and consistency training.

• YouTube
• 43 minutes
• Self-Paced
• Free Video

Deep CPCFG for Information Extraction

Explore Deep Conditional Probabilistic Context-Free Grammars for extracting structured information from business documents through end-to-end deep learning.

• YouTube
• 41 minutes
• Self-Paced
• Free Video

AI Bias and Fairness

Examines how bias enters AI pipelines and introduces data- and latent-space debiasing methods to reduce racial and gender disparities.

• YouTube
• 43 minutes
• Self-Paced
• Free Video

Evidential Deep Learning and Uncertainty

Learn how evidential deep learning helps neural networks quantify predictive uncertainty and signal when their outputs should not be trusted.

• YouTube
• 49 minutes
• Self-Paced
• Free Video