SCoRe: Submodular Combinatorial Representation Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Majee, Anay, Kothawade, Suraj, Killamsetty, Krishnateja, Iyer, Rishabh
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917685625880576
author Majee, Anay
Kothawade, Suraj
Killamsetty, Krishnateja
Iyer, Rishabh
author_facet Majee, Anay
Kothawade, Suraj
Killamsetty, Krishnateja
Iyer, Rishabh
contents In this paper we introduce the SCoRe (Submodular Combinatorial Representation Learning) framework, a novel approach in representation learning that addresses inter-class bias and intra-class variance. SCoRe provides a new combinatorial viewpoint to representation learning, by introducing a family of loss functions based on set-based submodular information measures. We develop two novel combinatorial formulations for loss functions, using the Total Information and Total Correlation, that naturally minimize intra-class variance and inter-class bias. Several commonly used metric/contrastive learning loss functions like supervised contrastive loss, orthogonal projection loss, and N-pairs loss, are all instances of SCoRe, thereby underlining the versatility and applicability of SCoRe in a broad spectrum of learning scenarios. Novel objectives in SCoRe naturally model class-imbalance with up to 7.6\% improvement in classification on CIFAR-10-LT, CIFAR-100-LT, MedMNIST, 2.1% on ImageNet-LT, and 19.4% in object detection on IDD and LVIS (v1.0), demonstrating its effectiveness over existing approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2310_00165
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle SCoRe: Submodular Combinatorial Representation Learning
Majee, Anay
Kothawade, Suraj
Killamsetty, Krishnateja
Iyer, Rishabh
Machine Learning
Computer Vision and Pattern Recognition
In this paper we introduce the SCoRe (Submodular Combinatorial Representation Learning) framework, a novel approach in representation learning that addresses inter-class bias and intra-class variance. SCoRe provides a new combinatorial viewpoint to representation learning, by introducing a family of loss functions based on set-based submodular information measures. We develop two novel combinatorial formulations for loss functions, using the Total Information and Total Correlation, that naturally minimize intra-class variance and inter-class bias. Several commonly used metric/contrastive learning loss functions like supervised contrastive loss, orthogonal projection loss, and N-pairs loss, are all instances of SCoRe, thereby underlining the versatility and applicability of SCoRe in a broad spectrum of learning scenarios. Novel objectives in SCoRe naturally model class-imbalance with up to 7.6\% improvement in classification on CIFAR-10-LT, CIFAR-100-LT, MedMNIST, 2.1% on ImageNet-LT, and 19.4% in object detection on IDD and LVIS (v1.0), demonstrating its effectiveness over existing approaches.
title SCoRe: Submodular Combinatorial Representation Learning
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2310.00165