LSTM is dead. Long Live Transformers!

I don't think you really want a 5yo's version, because that's "computers can be smart". But for an SDE1-level explanation I made this video which has gotten tons of compliments and keeps people watching year after year: ▶ this video It's in the context of NLP, which is where transformers started of course.

The attention mechanism started as a simple trick to not use recurrent neural networks. Read the intro in the original paper "Attention is all you need paper" ( link ) This video explains the drawbacks to RNNs and how transformers solve that: ▶ this video Andrej Karpathy explains attention here: link He explains how attention is seen as a communication network: link

Yannic Kilcher is great but this video worked better for me: "LSTM is dead. Long live transformers!" (Leo Dirac): ▶ this video

I've just recently been on a journey to understand transformers in and out, here are the resources I've found managed to drill it into my head: 1. Chapters 7, 9, 10 link This was really useful to really build up to the concepts of attention (although the actual attention section is still brief). 2. link This was great for visualization and understanding that attention wasn't exclusively for transformers and actually was for RNNs first. 3. link Getting to understand how transformers actually work visually. 4. ▶ this video This lecture by Leo Dirac was extremely helpful to finish off with, not only because it actually includes some pseudocode but it also revisits some key topics and covers why transformers are needed. One of the big confusion points for me was that the concept of ATTENTION and SELF-ATTENTION are not the same thing. Hope this helps.

I'm not sure I understand it well enough to say but watching a video on it [0] I think there were a few key points: * "Attention is all you need" introduced positional encoding which allows you to keep context of the word, allowing for more complex translation (and thus generative/chatgpt like tasks?) because words now have context relative to each other. Contrast this with "bag of words" models that only tells you whether the word is present or not. * I don't quite understand why but transformers (which "AiaYN" introduced) can be made fully parallel, compared with the RNN/LSTM networks which has to be serial per token. Fully parallel allows for GPU optimization, which means you can take advantage of Moore's law for training. I'm always a bit suspicious when people claim a breakthrough of this sort. There's no doubt that better algorithms give better results but how much is due to just faster computers, cheaper compute, memory, etc. [0] ▶ this video