deep learning 12
- Direction in Noise
- 1000 Steps Back: The Math Behind DDPM
- Finding Signal in the Static
- The Paper That Started the Neural NLP Revolution: Bengio's Neural Probabilistic Language Model
- Reading Both Ways: BERT and the End of Left-to-Right
- The Right Half: Decoder-Only Transformers
- Where Do Facts Go to Live? MLPs, Superposition, and a Basketball Player Named Michael
- Attention Is a Third of What You Need: QKV, Dot Products, and the Other Two-Thirds
- GANs: From a Thought Experiment to Photorealistic Faces
- When Neural Networks Lie: Adversarial Examples and the Art of Fooling AI
- From Autoencoders to VAEs: Learning to Generate, Not Just Compress
- LSTMs: How We Taught Neural Networks to Remember