AI with or without human supervision‪?‬ AI-podden

AI-podden is happy to host Pelin Ozbozkurt, Data Science Solution Director at Oracle. Pelin shares with us some of the interesting projects they have been working on at Oracle. The first one is on fraud detection which is an anomaly detection problem that is challenging because of the fact that fraud doesn’t happen that often which makes the data small. Pelin describes how they use other fields to tackle this problem. She mentions in particular the help of human supervision which is vital in building good predictive models. We also make the connection between this methodology and one of our previous guests at AI-podden, Gary Marcus, who wrote in his book “Rebooting AI” the importance of having a new fresh take on AI and not only rely solely on Deep Learning. What’s interesting is that Pelin mentions the development of unsupervised learning as one of the most interesting fields to explore within AI research. More interesting topics are discussed in this episode! Read more: https://ai-podden.se/2020/11/ai-med-eller-utan-mansklig-overvakning/ Download Oracle Analytics for Dummies for free here>> https://go.oracle.com/LP=103205?elqCampaignId=261794