this is absolutely FANTASTIC I watched Albert Gu's stanford lecture on state space models/ Mamba, and it was a great high level overview. But I really appreciate you taking it slower, and going farther into detail on the basic/ fundamental concepts. A lot of us aren't mathematicians or ML engineers, so it's much appreciated to be helped along with those concepts.
I rarely comment on videos, but this one was worth it. Thank you so much for such a clear explanation. You explained all the nuances that I previously did not understand in a very clear way. God bless you.
Your teaching approach is very good. You started from fundamental concepts and went deeper. This helped in gaining intuitions, understanding and avoid confusions in later part. Brilliant!
I just read about mamba and wanted to find a detailed explanation video. All you covered in this video is everything I need, thank you so much, keep on cooking
You are just too amazing! You can understand these stuff in great detail. Then you take the time and explain to us in educative videos. A true gem channel!
Thanks for your clear explanation of MAMBA, coming from a control theory background, very much appreciate its usage in LLMs. One small error that I noted was that the A matrix must be N x N to translate the previous N-dimensional hidden states h(t-1) to h(t). I believe the A matrix is also time-varying to produce selective output tokens.
Thanks a ton! Excellent explanation and great analogies to introduce the more advanced material. This is an absolute masterclass on how to teach advanced material.
Even I understood much of this. I have no education. Thank you! Mamba looks really cool. Especially like the long context and further refinement. It looks like a model that could be made to learn as it goes. Plasticity potential
As others have mentioned, you have a keen ability to explain difficult topics succinctly and completely. Keep up the awesome work! I could of used this when I took a class on time-series modeling! Hah!
Amazing explanation. I love this video because it covers sufficient depth and explains each concept with proper examples. I've subscribed instantly, and look forward to more such videos on recent papers.
Thank you for this great and smooth explanation. I think the model you are showing at 36:14 is valid if matrix A ( and B also to send each input directly to the corresponding ssm) is diagonal. Now in this way each hidden state at different canonical direction ( or different element of the vector) is independent of each other. So if A is not diagonal then assuming an eigen decomposition exist, then we may say there exist an equivalent ssm which can be represented independent ( if we change the basis to eigen basis) .
You're making very useful content, thank you!!! Maybe you could consider using larger text, so that one could read easily from a phone. Also a plus would be if the presentation were white on black (or bright color on black), it is less tiring to look at a dark screen for long periods of time.
Coding one is not very interesting, because the most interesting part is the selective scan algorithm, which is a CUDA Kernel. The architecture is not so different from any other language model. Of course it would be super cool to code the CUDA kernel from scratch ;-)
Hi, I was wondering if you could explain 36:40 a bit more where you talk about multi head attention. From what I understand each head in multi-head attention each head looks at the whole input vector. Our key value and query matrices are all of size Dx(head_size) where D being dimension of embedding, so when we find key say we do key = X @ key_matrix where X is an CxD dimensional matrix, C is context len. This means each head looks at the whole dimension of the embedding D and represents it a head_size vector meaning that arrows going into each head should point at every single input dim.
Very good lecture ! Thank you very much for putting this for free on youtube :) I have question though, if my understanding of the HiPPO framework is correct, the A matrix is built to uniformly approximate the input signal (name HiPPO LegS in the paper). "Our novel scaled Legendre measure (LegS) assigns uniform weight to all history [0, t]". But however at 41:49 you explain that it is decaying exponentially similarly to HiPPO LagT. Do they opt for HiPPO LagT when moving to s4 and Mamba or am I missing something ?
Dear Umar, referring to 53:50, recurrent SSM is indeed similar as prefix-sum (i.e., y=x_0+x_1+....x_N), but I the difference is that h_t=Ah_{t-1}+Bx_t, where h_{t_1} depends on h_{t-2}. I know how Blelloch parallel prefix scan works for calculating the sum of constants, but I do not know how parallel scan works for h_t=Ah_{t-1}+Bx_t. Could you please elaborate on it ? Thank you. @Umar
Hi Professor! Very good explanation as always. However, I have huge difficulties to understand the dimensions of objects. Why the hell A matrix would be of (D,N) dimensions since it is used to project a vector h_t-1 of N dimensions into N dimensions? By the way, why is it written "Represents structured N x N matrix" ?????!!!!
I have one question in terms of the example which you provided, 'the number of buddies'. I think the function should be like this : b(t)=5squ(3)^λt . please comment to me if I am wrong.
Anyone here to tell me what is the dimension of h_t? I thought it was a vector (D,1), but according to the slides, it seems it is a matrix of (D,N)!!?? Thanks in advance!