Emanuele Rossi

icon indicating a website link www.emanuelerossi.co.uk icon indicating an email [email protected]

icon company Twitter icon location London, UK icon location Researcher at Twitter and PhD student at Imperial College. Working on graph deep learning. Previously at Fabula AI (acquired by Twitter).

Results 3 comments of Emanuele Rossi

Potential problem with embeddings computing?

Hi @antoniofilipovic, You are correct. In practice, it does not make much of a difference since we're mostly using only 1 graph layer with TGN, but this is indeed incorrect...

Why is the node_feature matrix + 1 larger than the largest index of nodes?

Hi @benelot, the reason for the `+1` is that the node indexes start from 0, so if the max index is `max_idx`, it means that in total we have `max_idx...

Why is the node_feature matrix + 1 larger than the largest index of nodes?

The node feature are actually added to the memory every time the embedding is computed, see [here](https://github.com/twitter-research/tgn/blob/286815a4e15de6a6d4e0b8e660189e9c3c9f1dbf/modules/embedding_module.py#L98)