Unconstrained Robust Online Convex Optimization
By: Jiujia Zhang, Ashok Cutkosky
Potential Business Impact:
Learns from bad information without getting lost.
This paper addresses online learning with ``corrupted'' feedback. Our learner is provided with potentially corrupted gradients $\tilde g_t$ instead of the ``true'' gradients $g_t$. We make no assumptions about how the corruptions arise: they could be the result of outliers, mislabeled data, or even malicious interference. We focus on the difficult ``unconstrained'' setting in which our algorithm must maintain low regret with respect to any comparison point $u \in \mathbb{R}^d$. The unconstrained setting is significantly more challenging as existing algorithms suffer extremely high regret even with very tiny amounts of corruption (which is not true in the case of a bounded domain). Our algorithms guarantee regret $ \|u\|G (\sqrt{T} + k) $ when $G \ge \max_t \|g_t\|$ is known, where $k$ is a measure of the total amount of corruption. When $G$ is unknown we incur an extra additive penalty of $(\|u\|^2+G^2) k$.
Similar Papers
Exploiting Curvature in Online Convex Optimization with Delayed Feedback
Machine Learning (CS)
Improves computer learning with delayed information.
Robust Bayesian Optimisation with Unbounded Corruptions
Machine Learning (Stat)
Protects smart systems from bad data.
Beyond $\tilde{O}(\sqrt{T})$ Constraint Violation for Online Convex Optimization with Adversarial Constraints
Machine Learning (CS)
Makes computers learn better with tricky rules.