With otso, your expertise and understanding is a crucial part of the machine learning process.
Machine Learning is an iterative process, whereby improvements can be achieved through continued training, affirmation of correct labels, and correction of incorrect labels. We've designed otso to bring this process to the forefront of our platform, so that even if our AI makes a mistake, users have the ability to correct the error, and otso has the opportunity to learn from its blunder.
There's no tedious process for AI correction — we've built our correction system to be transparent and easy-to use. Simply find the document in our "Explore" view, click a drop-down menu to correct the sentiment, emotion, or a classification result, and click a check-mark to mark your correction as finished. That's it!
Any one-size-fits all approach is bound to have flaws. That's why we have built otso to be flexible to teaching, correction and customisation. Importantly, this means that we can customise the platform to adapt and react to baseline measurements in your data, so that you can more accurately tell when something is happening outside of the norm.