From Perceptron to Deep Learning at Industrial Scale
summary
Overview of neural networks from basic perceptron concepts to industrial-scale deep learning, explaining that modern production AI uses gradient-backpropagation-based models. Covers practical challenges encountered when scaling these networks in large corporations and illustrates them with real-world application examples.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:30Backpropagation basics
- 4:00Reggae detection experiment
- 8:08Sound perception and Fourier
- 11:40Simple perceptron
- 15:15Implementing autograd
- 22:02Backpropagation and gradients
- 27:46Weight update
- 30:15Training on music data
- 32:05Cross-entropy loss
- 33:55Switching to PyTorch
- 35:20Training on larger dataset
- 36:06Moving to GPU with CUDA
- 37:10Advanced neural architectures
- 40:43Deep learning scaling challenges
- 42:07AI ethics and future
- 43:35Future of transformers and backprop
- 45:01Summary and Q&A
related talks
16mDeep Learning in Production Seen by a Data Engineer
58mNeural Networks: A Primer
54mIntroduction to Understanding Deep Learning
45mIntroduction To Deep Learning
01h 02mNeural Networks: A Primer
42mDeep Learning with the GPUs in Production
27mDemystifying Deep Neural Networks
33mOvercoming the Barriers to Deep Learning in Production