Machine Learning: No, It Can’t Do That!

A SkillsCast for this session is not available.

Artificial Intelligence (AI) in general, and Machine Learning (ML) specifically are indisputably the hottest fields in the industry at the moment and have demonstrably advanced many areas of technology and science alike: web page classification, spelling correction, search ranking, graph building, large-molecule database screening for receptor-protein binding, predictive analytics, etc. etc. However, despite all the aforementioned advances, there are classes of problems (e.g. where “baselining” is required, such as network security) where ML is not a suitable technology to apply to. In fact, there are facets of ML that we simply don’t yet understand. In this session we describe the basics of AI and ML, discuss examples in network security and why ML is not a suitable solution, and finally discuss general shortcomings ML, namely reasoning and debugging.

KEYWORDS

Machine Learning, Deep Learning, Artificial Intelligence, General Intelligence, Machine Cognition, Data Science, Analytics, Security

YOU MAY ALSO LIKE:

• Practical Learning To Learn (SkillsCast recorded in May 2019)
• Making The Black Box Transparent: Lessons in Opacity (SkillsCast recorded in May 2019)