Timezone: »

Hidden Technical Debt in Machine Learning Systems
D. Sculley · Gary Holt · Daniel Golovin · Eugene Davydov · Todd Phillips · Dietmar Ebner · Vinay Chaudhary · Michael Young · Jean-François Crespo · Dan Dennison

Mon Dec 07 04:00 PM -- 08:59 PM (PST) @ 210 C #24

Machine learning offers a fantastically powerful toolkit for building useful complexprediction systems quickly. This paper argues it is dangerous to think ofthese quick wins as coming for free. Using the software engineering frameworkof technical debt, we find it is common to incur massive ongoing maintenancecosts in real-world ML systems. We explore several ML-specific risk factors toaccount for in system design. These include boundary erosion, entanglement,hidden feedback loops, undeclared consumers, data dependencies, configurationissues, changes in the external world, and a variety of system-level anti-patterns.

Author Information

D. Sculley (Google Research)
Gary Holt
Daniel Golovin (Google, Inc.)
Eugene Davydov (Google, Inc.)
Todd Phillips (Google, Inc.)
Dietmar Ebner
Vinay Chaudhary (Google, Inc.)
Michael Young (Google, Inc.)
Jean-François Crespo (Google, Inc.)
Dan Dennison (Google, Inc.)

More from the Same Authors