Learning with Imprinted Weights
By Hang Qi, Matthew A. Brown, David G. Lowe. 2017
12 citations recorded by the source corpus.
Human vision is able to immediately recognize novel visual categories after seeing just one or a few training examples. We describe how to add a similar capability to ConvNet classifiers by directly setting the final layer weights from novel training examples during low-shot learning. We call this process weight imprinting as it directly sets weights for a new category based on an appropriately scaled copy of the embedding layer activations for that training example. The imprinting process provides a valuable complement to training with stochastic gradient descent, as it provides immediate good classification performance and an initialization for any further fine-tuning in the future. We show how this imprinting process is related to proxy-based embeddings. However, it differs in that only a single imprinted weight vector is learned for each novel category, rather than relying on a nearest-neighbor distance to training instances as typically used with embedding methods. Our experiments show that using averaging of imprinted weights provides better generalization than using nearest-neighbor instance embeddings.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Matthieu Cord
- Nikos Komodakis
- Patrick Pérez
- Andrei Bursuc
- Spyros Gidaris
- Dacheng Tao
- Bernt Schiele
- Trevor Darrell
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Matthieu Cord
- Dacheng Tao
- Trevor Darrell
- Bernt Schiele
- Xiaogang Wang
- Ming-Hsuan Yang
- Andrew Zisserman
Citations in the corpus
- Dynamic Few-Shot Visual Learning Without Forgetting (2018, methodology): “[15] also propose to use the cosine similarity function in a similar way for the few-shot learning task.”
- Adaptive Masked Weight Imprinting for Few-Shot Segmentation (2019, methodology): “Most of the few-shot learning literature focused on image classification (Finn et al., 2017)(Koch et al., 2015)(Lin et al., 2017)(Vinyals et al., 2016)(Snell et al., 2017) (Qi et al., 2017)(Qiao et al., 2017).”
- Learning Representations by Predicting Bags of Visual Words (2020, methodology): “For comparison with this protocol we provide results of the supervised Cosine Classifier (CC) few-shot model [19, 55].”
- AMP: Adaptive Masked Proxies for Few-Shot Segmentation (2019, methodology): “The imprinting process was initially proposed in the context of classification [23].”