4 ms·
It seems to me the limitations with assistants aren't voice recognition but NLP. If I have to craft my query for a machine then you've failed to make a 'natural
by cameronbrown 7y ago
It seems to me the limitations with assistants aren't voice recognition but NLP. If I have to craft my query for a machine then you've failed to make a 'natural language' input.
- robbrown451 7y agoI agree (mostly) but not sure how that applies to my comment. Yes my major frustrations with it is that it isn't able to form a helpful response, not that it gets the actual words wrong. And in the case above, it goes beyond that. Designers of the product (Google maps) have it present onscreen options you need to click, in response to a spoken request (e.g. "find gas stations on my route") that is likely to be by a person who is driving and should not have to look at a screen. I would argue that that is not just non-helpful, but irresponsible and unsafe. But I guess you are right, in that if they have the transcripts, they don't need to actually listen to recordings. Then again they might need to listen to recordings if they are trying to gauge user frustration.
- dodobirdlord 7y ago> It seems to me the limitations with assistants aren't voice recognition but NLP. You probably speak English with a common American, Canadian, or British accent. Quality for common-accent English is fairly high, but is much lower for people with thick accents or who are speaking languages with smaller available training datasets. These sorts of initiatives are mostly about bringing quality up for those demographics.
- cameronbrown 7y agoThat's very true but I wish they'd work on making a better product as well. The growth in that area seems stagnant from both Siri and Google