yeah. with FL it should be possible to make sense out of all data that is distributed across devices without ever having to move it to a central location (i.e. collect it). In the case of speech data, users participating in a federated setting would likely come from different backgrounds, which could be reflected in their accent or use of language.
Comments
I’m guessing this will also help with thick accents?
yeah. with FL it should be possible to make sense out of all data that is distributed across devices without ever having to move it to a central location (i.e. collect it). In the case of speech data, users participating in a federated setting would likely come from different backgrounds, which could be reflected in their accent or use of language.