The newsletter of kiselbsthilfegruppe.de is now called rundbrief.ai
Curated free courses and channels from Stanford, MIT, DeepMind, fast.ai and more – from beginner to expert, all free on YouTube. The companion to our AI tool directory.
Learning resources
20 coursesYaser Abu-Mostafa explains the mathematical theory of why machine learning works at all — including VC dimension, the bias-variance tradeoff and regularization.
Watch on YouTubeAndrew Ng's classic Stanford course on the algorithms and math behind supervised, unsupervised and reinforcement learning.
Watch on YouTubeJosh Starmer breaks down decision trees, random forests, gradient boosting, PCA and statistics fundamentals with simple drawings and clear language.
Watch on YouTubeFrom simple multilayer perceptrons to attention mechanisms, GANs, variational autoencoders and self-supervised learning.
Watch on YouTubeA fast run through neural networks, CNNs, RNNs, generative models and large language models.
Watch on YouTubeAndrej Karpathy builds neural networks, backpropagation and finally a GPT-style language model with you, completely from scratch in Python.
Watch on YouTubeTuring Award winner Yann LeCun shares his view on energy-based models, self-supervised learning, world models and the future of AI.
Watch on YouTubeJeremy Howard (fast.ai) has you building real models from the very first lesson — image classifiers, NLP pipelines and tabular models.
Watch on YouTubeAndrew Ng covers CNNs, recurrent architectures, optimization methods like Adam and BatchNorm, and GANs.
Watch on YouTubeChelsea Finn teaches multi-task learning, transfer learning, meta-learning algorithms like MAML and few-shot learning.
Watch on YouTubeA hands-on intro to the Transformers library: tokenization, fine-tuning pretrained models, text classification and named entity recognition.
Watch on YouTubeA deep dive into word vectors, transformer architectures, attention, pretraining strategies and models like BERT and GPT.
Watch on YouTubeResearchers from OpenAI, Google Brain, DeepMind and Anthropic present current work on transformer architectures, scaling laws and multimodal AI.
Watch on YouTubeImage classification, object detection, semantic segmentation and generative models — including visualizing what CNNs actually learn.
Watch on YouTubeHado van Hasselt (DeepMind) teaches multi-armed bandits, Markov decision processes, policy-gradient methods and deep RL at UCL.
Watch on YouTubeEmma Brunskill covers Markov decision processes, policy evaluation, Q-learning, deep RL, exploration strategies and imitation learning.
Watch on YouTubeGrant Sanderson uses his signature animations to vividly explain gradient descent, backpropagation and how neural networks actually learn.
Watch on YouTubeKároly Zsolnai-Fehér summarizes recent AI research papers in short, visual clips just a few minutes long.
Watch on YouTubeAndrew Ng shows how to take ML models out of the Jupyter notebook and into production — with ML pipelines, data management and model monitoring.
Watch on YouTubeMIT teaches the paradigm of improving AI by improving the data rather than tuning the model — including data quality and labeling strategies.
Watch on YouTubeNo matching courses found.
Rundbrief
Good tools. Straight to your inbox.
New tools and the AI developments that matter. Free, in our German-language newsletter.
