Non-text recognition

[Sojin Lee]

Hello, 

I have a question about ML recognition for the non-text based numbers.

For example, we process 3-5 pages documents and we have some of the labels that should pick up only numbers. (e.g. 100,000 for a label, and $200,000 for another label).

1) Does AutoML recognize numbers?

2) Does AutoML distinguish different numbers?

Thank you

Sojin

[Jinjun Q]

Hello Sojin,

Labels are treated as strings. You can freely use numeric values as labels as they are treated as strings anyway. 

If I didn't understand your questions correctly, please provide some sample data to describe your questions.

[Sojin Lee]

Hello JinJun,

You mentioned "You can freely use numeric values as labels as they are treated as strings anyway." -> What do you mean by "treated as strings"?

Just to confirm, would the AutoML recognize a specific number? For example, if we are looking for $120 in the document, can we tag the $120 in the AI training ? Then would ML will find sentences with $120 from other documents?

Thanks !

Sojin

[Sojin Lee]

Just to ensure and add a point to the previous query, I am not asking whether label name can be a number. I am asking whether number can be used as a training phrase.

[shameemf]

Hello Sojin,

Yes, anything can be trained.

Cloud Natural Language uses machine learning to reveal the structure and meaning of text. It has a "Content classification" feature.

Following documentation might be helpful for further reference and clarification.

[1] https://cloud.google.com/natural-language/docs/basics#specifying_text_content

[2] https://cloud.google.com/natural-language