📄 Research Papers
Landmark AI/ML Papers
Interactive breakdowns of the papers that shaped modern AI — with diagrams, equations, and key insights.
1 paper · more comingAttention Is All You Need
The paper that introduced the Transformer architecture — replacing RNNs with pure self-attention. Foundation of every modern LLM including GPT, Claude, Gemini, and LLaMA. Covers scaled dot-product attention, multi-head attention, positional encoding, and SOTA translation results.
BERT: Pre-training of Deep Bidirectional Transformers
How masked language modeling and next-sentence prediction created the first truly bidirectional pre-trained model — changing NLP forever.
Language Models are Few-Shot Learners (GPT-3)
175B parameter model demonstrating that scale unlocks emergent few-shot learning — prompting without any gradient updates.
Denoising Diffusion Probabilistic Models (DDPM)
The paper that launched the diffusion model revolution — the backbone of Stable Diffusion, DALL·E 2, and Imagen.