Alice's Adventures in a Differentiable Wonderland -- Volume I, A Tour of the Land

TL;DR

This paper provides an accessible introduction to differentiable programming, covering core concepts, common neural network architectures, and practical implementation tips for understanding advanced models like LLMs.

Contribution

It offers a comprehensive, beginner-friendly overview of differentiable programming techniques and neural network designs, bridging theory and practical coding in PyTorch and JAX.

Findings

Explains automatic differentiation and its role in neural network training.

Describes key neural network components like convolutional, attentional, and recurrent blocks.

Provides practical guidance for understanding and implementing advanced models.

Abstract

Neural networks surround us, in the form of large language models, speech transcription systems, molecular discovery algorithms, robotics, and much more. Stripped of anything else, neural networks are compositions of differentiable primitives, and studying them means learning how to program and how to interact with these models, a particular example of what is called differentiable programming. This primer is an introduction to this fascinating field imagined for someone, like Alice, who has just ventured into this strange differentiable wonderland. I overview the basics of optimizing a function via automatic differentiation, and a selection of the most common designs for handling sequences, graphs, texts, and audios. The focus is on a intuitive, self-contained introduction to the most important design techniques, including convolutional, attentional, and recurrent blocks, hoping to bridge the gap between theory and code (PyTorch and JAX) and leaving the reader capable of understanding some of the most advanced models out there, such as large language models (LLMs) and multimodal architectures.