15 ms·
Much of the programming with ML has moved towards cleaning, extrapolating and generating the data. But this type of programing is - miracles- bugfree. We never
by honestoHeminway 9y ago
Much of the programming with ML has moved towards cleaning, extrapolating and generating the data.
But this type of programing is - miracles- bugfree. We never hear of data-conversion gone wrong, data corrupted or data-mining withou conclusive results here. Obviously such bugs lack the glamour of security bugs.
- skgoa 9y agoIt's also very difficult to catch these errors. Your trained model just doesn't work as well as it could, but how would you be able to tell?
- honestoHeminway 9y agoWell you can filter the data for extreme cases, and then write test for correct behaviour on those. If validating NN was solved, Darpa had some millions less and alot of plc programming and robotics would be done for. Basically its validating for every, decision treshold (formerly know as if-else/switch-case) the outcome, in combination with all other decision thresholds. n!. Maybe one would have a easier time for validating that the NN does not fullfill a certain outcome, by reversing it like a sock, filling the undesired outcome into it, and watch wether a valid input variable appears.