Torch: Season 1, Episode 3

Can machines learn on their own? – Supervised learning & misclassification

Do machines learn as independently as we humans? As a machine learning expert, I can tell you: not exactly. In "Torch", our new data science series, I elaborate on how "intelligent" machines really are and for which tasks they need support from us humans.


Download takeaway

In my “Torch” episode, you will learn:

  • The principles by which ML algorithms learn.
  • What "adversarial attacks" are and how they can change the output of your ML algorithm.
  • How to teach your ML algorithm to handle unclassified objects.

Want to see a current promising approach to machine learning? In my takeaway, I present a research project that brings machine learning closer to human learning and elaborate on how this is relevant for your own business applications.

Hello there, I'm Frank, and I've been working as a platform architect at Trivadis – Part of Accenture since 2020. Before that, I spent more than 14 years at the chemical company BASF in various roles and positions. As a scientifically oriented person with a strong practical orientation, I enjoy explaining the world of AI and ML to others. Off-duty, I'm a passionate live guitarist with a penchant for the musically rather heavier style and torture my fellow musicians in various bands with way too much distortion. PS: Did you enjoy my “Torch” episode? Then let’s get in touch – I would love to pass the torch and discuss the world of data science further!

What I share with you

The idea that machines learn as independently as humans tends to lead to false expectations. Are there approaches that bring machine learning closer to human learning? And how would that be useful? Find out in this takeaway ...

Download takeaway

Your contact