I'm reading this patent https://www.google.com/patents/EP1679694B1. I hoped to read a way to get ASR confidence score, but all I got from the patent is that the ASR module (102) somehow produces the confidence scores based on number of hypothesized words. Does anyone find anything more than this in the patent?
Any other patents you'd like to recommend reading on ASR confidence score? (I have read a bunch of papers, and was hoping that I will find something different in patents)
Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Any other patents you'd like to recommend reading on ASR confidence score?
Patents are not a good place for finding a truth just because they often disclose a random ideas, not something useful.
Confidence estimation is a solved problem and the optimal algorithm to implement a good estimation is known. It's done with a framework of statistical hypothesis testing. Confidence estimation is well described in a papers like:
Hi, I'm also looking for some information about the conffidence measure.
the link to the paper is broken, could you please let me know the title or the author of this paper?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm reading this patent https://www.google.com/patents/EP1679694B1. I hoped to read a way to get ASR confidence score, but all I got from the patent is that the ASR module (102) somehow produces the confidence scores based on number of hypothesized words. Does anyone find anything more than this in the patent?
Any other patents you'd like to recommend reading on ASR confidence score? (I have read a bunch of papers, and was hoping that I will find something different in patents)
Thanks.
Patents are not a good place for finding a truth just because they often disclose a random ideas, not something useful.
Confidence estimation is a solved problem and the optimal algorithm to implement a good estimation is known. It's done with a framework of statistical hypothesis testing. Confidence estimation is well described in a papers like:
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.93.6890
Thanks for the advice.
Hi, I'm also looking for some information about the conffidence measure.
the link to the paper is broken, could you please let me know the title or the author of this paper?
doi 10.1.1.93.6890