bookmark_border

Machine Learning is Hard with Zayd Enam

Software Engineering Daily,

Originally posted on Software Engineering Daily

Machine learning frameworks like Torch and TensorFlow have made the job of a machine learning engineer much easier. But machine learning is still hard. Debugging a machine learning model is a slow, messy process.

A bug in a machine learning model does not always mean a complete failure. Your model could continue to deliver usable results even in the presence of a mistaken implementation. Perhaps you made a mistake when cleaning your data, leading to an incorrectly trained model.

It is a general rule in computer science that partial failures are harder to fix than complete failures. In this episode, Zayd Enam describes the different dimensions on which a machine learning model can develop an error. Zayd is a machine learning researcher at the Stanford AI Lab, so I also asked him about AI risk, job displacement, and academia versus industry.

Show Notes

Why ML is hard

Sponsors


GoCD is an on-premise, open source, continuous delivery tool. Get better visibility into and control of your teams’ deployments with GoCD. Say goodbye to deployment panic and hello to consistent, predictable deliveries. Visit gocd.io for a free download. 


Apica System helps companies with their end-user experience, focusing on availability and performance. Test, monitor, and optimize your applications with Apica System. Apica is hosting an upcoming webinar about API basics for big data analytics. You can also find past webinars, such as how to optimize websites for fast load time.


Couchbase is a document database with the flexibility of NoSQL and the power of SQL. With Couchbase Server, you can build a fast, powerful NoSQL database that scales. Running Couchbase in containers on Kubernetes, Mesos, or OpenShift is easy, and at developer.couchbase.com you can find tutorials on how to build out your Couchbase deployment.


debugging

About the Podcast