5 papers
Reliable Active Learning from Unreliable Labels via Neural Collapse Geometry
Atharv Goel, Sharat Agarwal, Saket Anand +1
Active Learning (AL) promises to reduce annotation cost by prioritizing informative samples, yet its reliability is undermined when labels are noisy or when the data distribution s…
Exploiting Contextual Uncertainty of Visual Data for Efficient Training of Deep Models
Sharat Agarwal
Objects, in the real world, rarely occur in isolation and exhibit typical arrangements governed by their independent utility, and their expected interaction with humans and other o…
Reducing Annotation Effort by Identifying and Labeling Contextually Diverse Classes for Semantic Segmentation Under Domain Shift
Sharat Agarwal, Saket Anand, Chetan Arora
In Active Domain Adaptation (ADA), one uses Active Learning (AL) to select a subset of images from the target domain, which are then annotated and used for supervised domain adapta…
Does Data Repair Lead to Fair Models? Curating Contextually Fair Data To Reduce Model Bias
Sharat Agarwal, Sumanyu Muku, Saket Anand +1
Contextual information is a valuable cue for Deep Neural Networks (DNNs) to learn better representations and improve accuracy. However, co-occurrence bias in the training dataset m…
Contextual Diversity for Active Learning
Sharat Agarwal, Himanshu Arora, Saket Anand +1
Requirement of large annotated datasets restrict the use of deep convolutional neural networks (CNNs) for many practical applications. The problem can be mitigated by using active…