Top-$k$ Classification and Cardinality-Aware Prediction

2024-03-28 · Anqi Mao, Mehryar Mohri, Yutao Zhong

We present a detailed study of top-$k$ classification, the task of predicting the $k$ most probable classes for an input, extending beyond single-class prediction. We demonstrate that several prevalent surrogate loss functions in multi-class classification, such as comp-sum and constrained losses, are supported by $H$-consistency bounds with respect to the top-$k$ loss. These bounds guarantee consistency in relation to the hypothesis set $H$, providing stronger guarantees than Bayes-consistency due to their non-asymptotic and hypothesis-set specific nature. To address the trade-off between accuracy and cardinality $k$, we further introduce cardinality-aware loss functions through instance-dependent cost-sensitive learning. For these functions, we derive cost-sensitive comp-sum and constrained surrogate losses, establishing their $H$-consistency bounds and Bayes-consistency. Minimizing these losses leads to new cardinality-aware algorithms for top-$k$ classification. We report the results of extensive experiments on CIFAR-100, ImageNet, CIFAR-10, and SVHN datasets demonstrating the effectiveness and benefit of these algorithms.

📄 PDF Abstract BibTeX arXiv:2403.19625

Code (0)

등록된 구현이 없습니다.

Tasks

Classification Multi-class Classification Prediction

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Cardinality-Aware Set Prediction and Top-$k$ Classification

2024-07-09 · Corinna Cortes, Anqi Mao, Christopher Mohri, Mehryar Mohri 외

We present a detailed study of cardinality-aware top-$k$ classification, a novel approach that aims to learn an accurate top-$k$ set predictor while maintaining a low cardinality. We introduce a new target loss function …

Classification Prediction

Predict and Constrain: Modeling Cardinality in Deep Structured Prediction

2018-02-13 · ICML 2018 7 · Nataly Brukhim, Amir Globerson

Many machine learning problems require the prediction of multi-dimensional labels. Such structured prediction models can benefit from modeling dependencies between labels. Recently, several deep learning approaches to st…

BIG-bench Machine Learning Deep Learning General Classification Multi-Label Classification+3

When Can Test-Time Adaptation Help Zero-Shot CT Vision-Language Models?

2026-07-17 · Ailar Mahdizadeh, Puria Azadi Moghadam, Xiangteng He, Leonid Sigal arxiv

3D CT vision-language models (VLMs) classify abnormalities from text prompts in a zero-shot manner, enabling cross-institution deployment where labels are scarce and clinical tasks shift faster than supervised models can…

Test-time Adaptation

Using Random Effects to Account for High-Cardinality Categorical Features and Repeated Measures in Deep Neural Networks

2021-12-01 · NeurIPS 2021 12 · Giora Simchoni, Saharon Rosset

High-cardinality categorical features are a major challenge for machine learning methods in general and for deep learning in particular. Existing solutions such as one-hot encoding and entity embeddings can be hard to sc…

Entity Embeddings

DeepSetNet: Predicting Sets with Deep Neural Networks

2016-11-28 · ICCV 2017 10 · S. Hamid Rezatofighi, Vijay Kumar B G, Anton Milan, Ehsan Abbasnejad 외

This paper addresses the task of set prediction using deep learning. This is important because the output of many computer vision tasks, including image tagging and object detection, are naturally expressed as sets of en…

image-classification Image Classification Object Counting object-detection+2