I read some papers for table to text task, It’s understandable but i got some problems while re-implement the idea that combine the embedding of slot and field in some papers:

https://arxiv.org/pdf/1711.09724v1

https://aclanthology.org/2020.acl-main.355.pdf

https://arxiv.org/pdf/1809.01797v2

I have the image here

enter image description here

Let get the example in the image, and to be simpler, I just get the field name. So I have a sequence of slot values:

"George Mikell .... 4 April..." let say 10 tokens

and a sequence of corresponding field name (actually i’m not sure this is the meaning of the authors):

name, name, ..., day of birth, day of birth... Okay this is confused part. As the field name has several tokens. If I express it it would be (10, l) which l is max token for a filed, let say 3. -> (10, 3)

Okay in the forward phase, let say batch size = 1.

So slot values: (1,10)
field values: (1,10,3)

then after embedding slot embedding -> (1, 10, 64)
field embeddings -> (1, 10, 3, 32)

I guess the purpose is concat the field embedding to each slot embeding together. but I stuck at this step.

It would be nice if you take a look at paper too. thank you :<