Despite the empirical success of knowledge distillation, there still lacks a theoretical foundation that can naturally lead to computationally inexpensive implementations. To address this concern, we forge an alternative connection between information theory and knowledge distillation using a recently proposed entropy-like functional. In doing so, we introduce two distinct complementary losses which aim to maximise the correlation and mutual information between the student and teacher representations. Our method achieves competitive performance to state-of-the-art on the knowledge distillation and cross-model transfer tasks, while incurring significantly less training overheads than closely related and similarly performing approaches. We further demonstrate the effectiveness of our method on a binary distillation task, whereby we shed light to a new state-of-theart for binary quantisation. The code, evaluation protocols, and trained models will be publicly available.
Feature-based image matching has been significantly improved through the use of deep learning and new large datasets. However, there has been little work addressing the computational cost, model size, and matching accuracy tradeoffs for the state of the art models. In this paper, we consider these practical aspects and improve the state-of-theart HardNet model through the use of depthwise separable layers and an efficient tensor decomposition. We propose the Convolution-Depthwise-Pointwise (CDP) layer, which partitions the weights into a low and full rank decomposition to exploit the naturally emergent structure in the convolutional weights. We can achieve an 8× reduction in the number of parameters on the HardNet model, 13× reduction in the computational complexity, while sacrificing less than 1% on the overall accuracy across the HPatches benchmarks. To further demonstrate the generalisation of this approach, we apply it to other state-of-the-art descriptor models, where we are able to a significant performance improvement.
We present an efficient alternative to the convolutional layer through utilising spatial basis filters (SBF). SBF layers exploit the spatial redundancy in the convolutional filters across the depth to achieve overall model compression, while maintaining the top-end accuracy of their dense counter-parts. Training SBF-Nets is modelled as a simple pruning problem, but instead of zeroing out the pruned channels, they are replaced with inexpensive transformations from the set of non-pruned features. To enable an adoption of these SBF layers, we provide a flexible training pipeline and an efficient implementation in CUDA with low latency. To further demonstrate the effective capacity of these models, we apply semi-supervised knowledge distillation that leads to significant performance improvements over the baseline networks. Our experiments show that SBF-Nets are effective and achieve comparable or improved performance to state-of-the-art across CIFAR10, CIFAR100, Tiny-ImageNet, and ILSCRC-2012.
In this paper we revisit the efficacy of knowledge distillation as a function matching and metric learning problem. In doing so we verify three important design decisions, namely the normalisation, soft maximum function, and projection layers as key ingredients. We theoretically show that the projector implicitly encodes information on past examples, enabling relational gradients for the student. We then show that the normalisation of representations is tightly coupled with the training dynamics of this projector, which can have a large impact on the students performance. Finally, we show that a simple soft maximum function can be used to address any significant capacity gap problems. Experimental results on various benchmark datasets demonstrate that using these insights can lead to superior or comparable performance to state-of-the-art knowledge distillation techniques, despite being much more computationally efficient. In particular, we obtain these results across image classification (CIFAR100 and ImageNet), object detection (COCO2017), and on more difficult distillation objectives, such as training data efficient transformers, whereby we attain a 77.2% top-1 accuracy with DeiT-Ti on ImageNet.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.