@nierenstein
I assumed that captcha was being used to train image models.
I didn't realize you could feed it incorrect information.
I figured they were smarter than this, and assumed that some of the answers were true/false and those that were not they would take a confidence level of 9/10 people saying it was a "fire hydrant" would then be used to feed back into the model.
I assumed that captcha was being used to train image models.
I didn't realize you could feed it incorrect information.
I figured they were smarter than this, and assumed that some of the answers were true/false and those that were not they would take a confidence level of 9/10 people saying it was a "fire hydrant" would then be used to feed back into the model.
- replies
- 1
- announces
- 0
- likes
- 0