I absolutely love the concise explanations in your video series. So clear AND concise. I have not seen any more clear explanations on TH-cam than yours. The animations help so much. THANK YOU!
Wow. Thanks! I'm a graduate student in AI and I may not get 100% of the video content but it helped me understand where we can use GATs. I'd love to learn more about them now.
That's the most understandable introductions on GAT I've ever seen! Really love the way you explain very complicated things using simple examples. Thank you so much! Any chance you can do a video on transformer some time?
This is top notch stuff. I'm trying to leverage my biophysics background for a GNN-related project, and this channel is kinda my go-to for concepts. Wanted to take your course, but being a grad student in a third-world country, and when the conversion rate kicks in, it gets out of reach :/
Hi Zak, thank you for such a great explanation. I'm currently facing trouble keeping up with the terminology in GNNs. Do you have any blog recommendations that serve as a good overview of the field and might be easier to follow for a beginner in GNNs (unfortunately I cannot afford your course and I've tried looking at your blogs and I could only find one on GCN). Nevertheless, I hope that you continue to make such cool videos!
Reach out on Discord if affordability is an issue and we can work something out. Otherwise, you might enjoy Jure's Stanford CS224W course, which has all the lectures on TH-cam. It has excellent material, but might be tough for beginners.
Out of curiosity, what are your thoughts on combining RGCNs with GATs, as in, instead of regular degree based aggregations in RGCNs for each Relation, what do you think of calculating an attention based weight in the formula?
What's the significant difference between Relational GCN & graph attention network ? is it the concatenation and the dot product of a ? coz I saw both are also assigned with different weights
Hi hope you're doing well Is there any graph neural network architecture that receives multivariate dataset instead of graph-structured data as an input? I'll be very thankful if you answer me i really nead it Thanks in advanced
@@welcomeaioverlords thanks for your responding, but you've said GNNs use both multivarite and graph-structured dataset. I wanna know how it can deal with multivariate dataset?
I absolutely love the concise explanations in your video series. So clear AND concise. I have not seen any more clear explanations on TH-cam than yours. The animations help so much. THANK YOU!
Brief explanation for GAT. Thanks alot. This is currently what I need for thesis
Wow. Thanks!
I'm a graduate student in AI and I may not get 100% of the video content but it helped me understand where we can use GATs. I'd love to learn more about them now.
I don't think any other explanation can do better than this. Thanks a lot!
Good to see that you are doing your part to bring Roko's Basilisk into existence with your TH-cam channel.
That's the most understandable introductions on GAT I've ever seen! Really love the way you explain very complicated things using simple examples. Thank you so much! Any chance you can do a video on transformer some time?
Great! .. I really like your explanation .. we need more, please
I like you videos. Easy to understand with the essential things revealed
awesome content! thank you!
Incredible video. Super well explained and much better than how university taught it
=
Great video!
Just excellent.
Thank you!
This is top notch stuff. I'm trying to leverage my biophysics background for a GNN-related project, and this channel is kinda my go-to for concepts. Wanted to take your course, but being a grad student in a third-world country, and when the conversion rate kicks in, it gets out of reach :/
Contact me on Discord if you want to work something out for the course.
well explained...
Fantastic
Hi Zak, thank you for such a great explanation. I'm currently facing trouble keeping up with the terminology in GNNs. Do you have any blog recommendations that serve as a good overview of the field and might be easier to follow for a beginner in GNNs (unfortunately I cannot afford your course and I've tried looking at your blogs and I could only find one on GCN). Nevertheless, I hope that you continue to make such cool videos!
Reach out on Discord if affordability is an issue and we can work something out. Otherwise, you might enjoy Jure's Stanford CS224W course, which has all the lectures on TH-cam. It has excellent material, but might be tough for beginners.
Out of curiosity, what are your thoughts on combining RGCNs with GATs, as in, instead of regular degree based aggregations in RGCNs for each Relation, what do you think of calculating an attention based weight in the formula?
Anybody told you you look exactly like Chad Michael Murray from One Tree Hill? The resemblance is remarkable
What is the link to the first video for Graph NNs?
What's the significant difference between Relational GCN & graph attention network ? is it the concatenation and the dot product of a ? coz I saw both are also assigned with different weights
Error in Min. 2:53 --> alpha u,v expression --> denominator exponent should be a(hk, hv)
Great overview, thanks. What does the T in GAT stand for?
I assumed it's Graph ATtention networks.
please answer me. did GAT considered as a variant of GCNs ?
Wouldn't concatination of multiple head results lead to exponentially growing node representation with regards to the number of gat layers?
The vector gets longer, yes, but after fully connected layer, you end up with any dimension you want
Hi hope you're doing well
Is there any graph neural network architecture that receives multivariate dataset instead of graph-structured data as an input?
I'll be very thankful if you answer me i really nead it
Thanks in advanced
GNNs use data that is both graph structured and multivariate. If your data has no graph structure, you don't need a *graph* based method like GNN.
@@welcomeaioverlords thanks for your responding, but you've said GNNs use both multivarite and graph-structured dataset. I wanna know how it can deal with multivariate dataset?
@@nastaranmarzban1419 The "messages" during message passing are the nodes' feature vectors. The feature vectors are the variates of the nodes.
@@welcomeaioverlords I got it.thanks
what do u mean by the degree of a node?
how many neighbours/connections