Timezone: »
Poster
Near-Tight Margin-Based Generalization Bounds for Support Vector Machines
Allan Grønlund · Lior Kamma · Kasper Green Larsen
Support Vector Machines (SVMs) are among the most fundamental tools for binary classification.
In its simplest formulation, an SVM produces a hyperplane separating two classes of data using the largest possible margin to the data. The focus on maximizing the margin has been well motivated through numerous generalization bounds.
In this paper, we revisit and improve the classic generalization bounds in terms of margins. Furthermore, we complement our new generalization bound by a nearly matching lower bound, thus almost settling the generalization performance of SVMs in terms of margins.
Author Information
Allan Grønlund (Aarhus University)
Lior Kamma (Aarhus University)
Kasper Green Larsen (Aarhus University, MADALGO)
More from the Same Authors
-
2023 Oral: AdaBoost is not an Optimal Weak to Strong Learner »
Mikael Møller Høgsgaard · Kasper Green Larsen · Martin Ritzert -
2023 Poster: The Fast Johnson-Lindenstrauss Transform Is Even Faster »
Ora Nova Fandina · Mikael Møller Høgsgaard · Kasper Green Larsen -
2023 Poster: AdaBoost is not an Optimal Weak to Strong Learner »
Mikael Møller Høgsgaard · Kasper Green Larsen · Martin Ritzert -
2021 Poster: CountSketches, Feature Hashing and the Median of Three »
Kasper Green Larsen · Rasmus Pagh · Jakub Tětek -
2021 Spotlight: CountSketches, Feature Hashing and the Median of Three »
Kasper Green Larsen · Rasmus Pagh · Jakub Tětek -
2019 Poster: Optimal Minimal Margin Maximization with Boosting »
Alexander Mathiasen · Kasper Green Larsen · Allan Grønlund -
2019 Oral: Optimal Minimal Margin Maximization with Boosting »
Alexander Mathiasen · Kasper Green Larsen · Allan Grønlund