TransBoost: Improving the Best ImageNet Performance using Deep Transduction

05/26/2022
by   Omer Belhasin, et al.
0

This paper deals with deep transductive learning, and proposes TransBoost as a procedure for fine-tuning any deep neural model to improve its performance on any (unlabeled) test set provided at training time. TransBoost is inspired by a large margin principle and is efficient and simple to use. The ImageNet classification performance is consistently and significantly improved with TransBoost on many architectures such as ResNets, MobileNetV3-L, EfficientNetB0, ViT-S, and ConvNext-T. Additionally we show that TransBoost is effective on a wide variety of image classification datasets.

READ FULL TEXT

page 5

page 6

page 7

page 12

page 13

research
07/25/2018

Do Better ImageNet Models Transfer Better... for Image Recommendation ?

Visual embeddings from Convolutional Neural Networks (CNN) trained on th...
research
05/23/2018

Do Better ImageNet Models Transfer Better?

Transfer learning has become a cornerstone of computer vision with the a...
research
05/27/2019

On Mixup Training: Improved Calibration and Predictive Uncertainty for Deep Neural Networks

Mixup zhang2017mixup is a recently proposed method for training deep neu...
research
02/26/2021

Unifying Remote Sensing Image Retrieval and Classification with Robust Fine-tuning

Advances in high resolution remote sensing image analysis are currently ...
research
04/27/2021

Efficient Pre-trained Features and Recurrent Pseudo-Labeling inUnsupervised Domain Adaptation

Domain adaptation (DA) mitigates the domain shift problem when transferr...
research
08/01/2022

giMLPs: Gate with Inhibition Mechanism in MLPs

This paper presents a new model architecture, gate with inhibition MLP (...
research
05/21/2018

A Simple Cache Model for Image Recognition

Training large-scale image recognition models is computationally expensi...

Please sign up or login with your details

Forgot password? Click here to reset