← Goal templates
Ten landmark AI papers
The papers modern AI is built on, read the researcher's way.
Read the papers modern AI is built on, from Adam and ResNet to Attention Is All You Need, BERT, GPT-3 and the scaling laws, with the three-pass method: know which deserve a deep read and what each one changed.
intermediateabout 5 weeks at 45 min a day10 free sources
What you'll cover
Phase 1
Training and architectures
- Adam
- Generative adversarial networks
- Residual networks
Phase 2
Reinforcement learning
- Deep Q-networks
- PPO
Phase 3
Transformers
- Attention Is All You Need
- BERT
- Vision Transformers
Phase 4
Scale
- Scaling laws
- GPT-3 and few-shot learning
An outline. Your own plan is made when you start, around your level and deadline.
Sources
Free, and fetched for you
- PaperAdam: A Method for Stochastic Optimizationarxiv.org ↗
- PaperGenerative Adversarial Networksarxiv.org ↗
- PaperDeep Residual Learning for Image Recognitionarxiv.org ↗
- PaperPlaying Atari with Deep Reinforcement Learningarxiv.org ↗
- PaperProximal Policy Optimization Algorithmsarxiv.org ↗
- PaperAttention Is All You Needarxiv.org ↗
- PaperBERT: Pre-training of Deep Bidirectional Transformers for Language Understandingarxiv.org ↗
- PaperAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scalearxiv.org ↗
- PaperScaling Laws for Neural Language Modelsarxiv.org ↗
- PaperLanguage Models are Few-Shot Learnersarxiv.org ↗
All sources belong to their creators and are free to watch or read; Confu links to them and turns them into study PDFs for you.
More templates
- Neural networks from scratchFrom a single neuron to backpropagation you could code yourself.
- Transformers and large language modelsHow ChatGPT-style models work, down to attention and a GPT you build yourself.
- Reinforcement learning, from basics to PPOAgents, rewards and policies, up to the algorithm behind RLHF.
- Linear algebra, the visual wayVectors, matrices and eigenvectors you can actually picture.