Supervised ImageNet
Supervised ImageNet research focuses on improving image classification models by leveraging the massive ImageNet dataset. Current efforts concentrate on enhancing data curation strategies, developing more efficient training methods (including exploring alternative architectures like binary neural networks and leveraging self-supervised learning), and addressing challenges like dataset bias and the need for explainable AI. These advancements are crucial for improving the accuracy, efficiency, and trustworthiness of computer vision systems across various applications, from medical imaging to agricultural technology.
Papers
A Whac-A-Mole Dilemma: Shortcuts Come in Multiples Where Mitigating One Amplifies Others
Zhiheng Li, Ivan Evtimov, Albert Gordo, Caner Hazirbas, Tal Hassner, Cristian Canton Ferrer, Chenliang Xu, Mark Ibrahim
Optimizing Learning Rate Schedules for Iterative Pruning of Deep Neural Networks
Shiyu Liu, Rohan Ghosh, John Tan Chong Min, Mehul Motani
Spurious Features Everywhere -- Large-Scale Detection of Harmful Spurious Features in ImageNet
Yannic Neuhaus, Maximilian Augustin, Valentyn Boreiko, Matthias Hein
Self-Supervised Learning based on Heat Equation
Yinpeng Chen, Xiyang Dai, Dongdong Chen, Mengchen Liu, Lu Yuan, Zicheng Liu, Youzuo Lin
Join the High Accuracy Club on ImageNet with A Binary Neural Network Ticket
Nianhui Guo, Joseph Bethge, Christoph Meinel, Haojin Yang
Can we Adopt Self-supervised Pretraining for Chest X-Rays?
Arsh Verma, Makarand Tapaswi