Graph Neural Networks (GNN) | Nodes, Edges, Adjacency Matrix, Message Passing, Aggregation explained

Рет қаралды 3,391

Күн бұрын

Пікірлер: 35

@Fire_AJ_ 4 ай бұрын

I rarely comment on videos, but the quality of explanation is so good that it is worth it to appreciate! Waiting for the next video!

@vizuara 4 ай бұрын

Thank you! Stay tuned for more videos!

@HnM0720 3 ай бұрын

Very clear explanation. Can you suggest some good resourses for learning Graph Networks? We need more lecturers like this.

@physicsanimated1623 5 ай бұрын

This is Vivek Karmarkar - really loved the visuals and the simplicity of the lecture!

@vizuara 5 ай бұрын

Thank you!

@anishj847 5 ай бұрын

Great lecture! Covered everything from theory to math concepts.

@vizuara 5 ай бұрын

Thank you! Stay tuned for more videos.

@indigo1417 3 ай бұрын

thank you. I found this video very helpful. In particular i liked how you went through the message aggregation steps. In other videos i have watched, message aggregation is described with abstract maths notation that has left me still vague about what is happening. Also, great to see an up to date series of tutorials coming out. I can find a lot of tutorials from 3-4 years ago, but not much since then.

@padmanabhannamboothiri5314 3 ай бұрын

Good example ( Sam) and simple syle. Anyone can follow it easly.

@khushbukhushbu1363 3 ай бұрын

Thank you, Very well explained, I have a small doubt, that at 22:06 in the node intialisation process, are the nodes intialised randomly or there is some method to do so?

@devanshuagarwal4513 4 ай бұрын

Very well explained and easy to understand. When are you guys releasing next video

@bay-bicerdover 5 ай бұрын

A sublime lecturer! 👏🏻

@vizuara 5 ай бұрын

Thank you! Stay tuned for more videos.

@kinglobby6684 5 ай бұрын

This is really good. It would be great to know how to make message passing graph embedding for molecular structure and doing pooling operation to create single vector. will there be coding session in upcoming sessions.

@vizuara 4 ай бұрын

Thank you! Sure, in the upcoming lectures, there will be implementation of GNNs. Stay tuned for more videos!

@rajehkumarmishra 4 ай бұрын

loved it but can you please share the slides also

@PraveenKumar-lv6cu 5 ай бұрын

Very Interesting, Please keep it up!

@vizuara 5 ай бұрын

Thank you!

@tusharmuratkar7290 5 ай бұрын

Great teaching ....Very Nice

@vizuara 4 ай бұрын

Thank you! Stay tuned for more videos!

@lekshmip3043 5 ай бұрын

Why arnt you considering the edge between a and d for aggregation?

@vizuara 5 ай бұрын

Yes, you are correct. The aggregated messages will be [3.5,2,1.5] , [2.3,2,2.6] , [2.5,1.5,2.5] , [3,1.5,2].

@naiduruthika7 4 ай бұрын

@@vizuara If edge between A and D is considered, then the aggregated message of A has to be [10/4, 8/4,10/4] and that of D has to be [2.3,2,2.6]. Please correct me if I'm wrong.