It's an exciting time to be thinking about artificial intelligence and machine learning, but it can be challenging to figure out a practical and reasonable approach to using these technologies. In this webinar, we'll discuss the history of machine learning, why now is the time to think about investing, and how to identify and execute on practical machine learning opportunities.
24. Rich media is now perceptible.
Example use cases:
• Quality control systems.
• Surgical robotics realtime error
detection.
• Real world navigation in constrained
environments.
• Legacy media made useful and
searchable.
28. Language is becoming computable.
Example use cases:
• Insurance company understanding
personas of applicants for marketing.
• Bank parsing customer service call
transcripts to better recommend
actions.
• Investment bank automatically parsing
the news effectively for commodities
traders.
32. We must understand what black boxes do!
Example use cases:
• Regulatory compliance and bias
testing
• Telecom churn reasoning
• Reverse engineering 3rd party models
Let’s make sure we are using the same robust vocabulary
Conflation of what’s actually happening and what’s possible
1952, Claude Shannon
1952, Claude Shannon
Linear regression (thanks to Wikipedia for the image). The line represents a generalized function we can learn from data. It’s a 200+ year old technique.
Perceptrons were inspired by how he thought the brain worked at the time, but it is not a human mind in itself.