The original question was asked in 2015. So I believe it is appropriate to include surveys with mathematical flavor on more recent/advanced topics in neural networks and deep learning. 

 1. [**A Mathematical Introduction to Generative Adversarial Nets (GAN) - 2020**][1]

> The survey paper rigorously formulate the problem that a ***Generative Adversarial Network*** (GAN) tries to solve as a min-max optimization problem (Theorem 2.3).   

 2. [**Neural Network Theory - 2022**][2]

> A quite bit of analysis is involved in studying the ***Expressive Power of Neural Networks***, especially when it comes to approximation theorems. The lecture notes above survey [classical universal approximation results][3] for shallow neural nets as well as more advanced topics including approximation results under the manifold hypothesis, benefits of depth to expressivity, and results on the *VC dimension* and the convexity of the set of functions implemented by feedforward neural networks. 

> (Another mathematical question pertaining to this topic which can be of interest to a combinatorialist is to count/bound the number of *activation regions* of a *ReLU network* as a measure of its complexity/expressiveness; see this [paper][4] for instance.)

 3. [**Graph Representation Learning - 2020**][5]

> After a review of older methods of node embedding (e.g. using random walks on graphs), the book presents a careful treatment of ***Graph Neural Networks*** (GNNs). The *message passing framework* in GNNs is rigorously defined; and it is discussed how different *update* and *aggregation* operations in message passing yield different types of GNNs (Graph Convolutional Networks, Graph Attention Networks, GraphSAGE etc.). Furthermore, there is a chapter discussing the mathematical motivation such as the graph Fourier transform and the  Weisfeiler-Lehman algorithm. 

 4. [**Geometric Deep Learning: Going beyond Euclidean data - 2017**][6]

> The article provides an overview of ***Geometric Deep Learning***, an "umbrella term" for deep learning on "non-Euclidean structured domains". This goes beyond graph neural networks and involves many problems in which a *geometric prior* (e.g. symmetry or scale separation) is present such as when the data lives on a submanifold, or *Convolutional Neural Networks* (CNNs) in which a grid-like structure is exploited. A more comprehensive and recent treatment can be found in [this book][7] which brands geometric deep learning as an "attempt to apply the Erlangen Programme mindset to the domain of deep learning". 


 5. [**A mathematician's introduction to transformers and large language models - 2022**][8] <br/>
[**Formal Algorithms for Transformers - 2022**][9]

> No theorems are involved but these surveys on ***Transformers*** were rigorous enough for my taste. The first one explains the *attention mechanism*, and the second one precisely presents the algorithms for various components of a transformer (*positional embedding*, attention, *layer normalization* etc.). Both references are self-contained.  

 6. [**Large Language Models - 2023**][10]

> The paper gives an introduction to ***Large Language Models (LLMs)*** "for mathematicians, physicists, and other scientists and readers who are mathematically knowledgeable but not necessarily expert in machine learning or artificial intelligence."

 


  [1]: https://arxiv.org/abs/2009.00169v1
  [2]: http://pc-petersen.eu/Neural_Network_Theory.pdf
  [3]: https://link.springer.com/article/10.1007/BF02551274
  [4]: https://arxiv.org/abs/1711.02114
  [5]: https://www.cs.mcgill.ca/~wlh/grl_book/
  [6]: https://ieeexplore.ieee.org/document/7974879
  [7]: https://arxiv.org/abs/2104.13478
  [8]: https://x-dev.pages.jsc.fz-juelich.de/2022/07/13/transformers-matmul.html
  [9]: https://arxiv.org/abs/2207.09238
  [10]: https://arxiv.org/abs/2307.05782v2