Peter Norvig, Google Director of Research, presented “State-of-the-Art AI: Building Tomorrow’s Intelligent Systems” at the EmTech Digital conference. He compares traditional software programming to machine learning development and highlights the new challenges of debugging and verifying ML-based systems.
Traditional software development vs. Machine Learning-based development:
“The problem here is the methodology for scaling this [machine learning verification] up to a whole industry is still in progress. We have been doing this for a while; we have some clues for how to make it work, but we don’t have the decades of experience that we have in developing and verifying regular software.”
“For some problems, we just don’t know what the truth is. So, how do you train a machine-learning algorithm on data for which there are no set results?”