Timezone: »

Reproducibility in Optimization: Theoretical Framework and Limits
Kwangjun Ahn · Prateek Jain · Ziwei Ji · Satyen Kale · Praneeth Netrapalli · Gil I Shamir

Wed Nov 30 09:00 AM -- 11:00 AM (PST) @ Hall J #722

We initiate a formal study of reproducibility in optimization. We define a quantitative measure of reproducibility of optimization procedures in the face of noisy or error-prone operations such as inexact or stochastic gradient computations or inexact initialization. We then analyze several convex optimization settings of interest such as smooth, non-smooth, and strongly-convex objective functions and establish tight bounds on the limits of reproducibility in each setting. Our analysis reveals a fundamental trade-off between computation and reproducibility: more computation is necessary (and sufficient) for better reproducibility.

Author Information

Kwangjun Ahn (MIT)
Prateek Jain (Google Research)
Ziwei Ji (Google)
Satyen Kale (Google)
Praneeth Netrapalli (Google Research)
Gil I Shamir (Google)

More from the Same Authors