There are thousand of hours of video interviews with Holocaust survivors that were filmed in the 90s on VHS. Many were proficient English speakers who communicated their ideas effectively, but still with pretty thick European accents (plus poor grammar, incomplete sentences, etc.)
Can you recommend any software that can take this into account and create accurate English transcripts?
The goal would be to apply text analysis to this massive corpus to discover any unique insights (e.g. cross-referencing stories and seeing what surfaces).
Thank you!
Thanks.
I personally think a startup-focused version would be useful.
One main obstacle I foresee is anonymity: a Staff ML Engineer at a 30-Engineer startup would be easily identifiable!
Your thoughts? Thanks.
Note: I did try the StackShare API and there is no filtering feature. So if you purchase the 1,000-company plan, you have no control over what they send you. It'll be a randomly generated list of 1,000 companies that use the technology you requested, a hodgepodge of companies all around the world, big and small, new and old.
I look forward to hearing your thoughts. Thanks!