SketchEmbedNet: Learning Novel Concepts by Imitating Drawings

Alexander Wang, Mengye Ren, Richard Zemel
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:10870-10881, 2021.

Abstract

Sketch drawings capture the salient information of visual concepts. Previous work has shown that neural networks are capable of producing sketches of natural objects drawn from a small number of classes. While earlier approaches focus on generation quality or retrieval, we explore properties of image representations learned by training a model to produce sketches of images. We show that this generative, class-agnostic model produces informative embeddings of images from novel examples, classes, and even novel datasets in a few-shot setting. Additionally, we find that these learned representations exhibit interesting structure and compositionality.

Cite this Paper


BibTeX
@InProceedings{pmlr-v139-wang21s, title = {SketchEmbedNet: Learning Novel Concepts by Imitating Drawings}, author = {Wang, Alexander and Ren, Mengye and Zemel, Richard}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {10870--10881}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/wang21s/wang21s.pdf}, url = {https://proceedings.mlr.press/v139/wang21s.html}, abstract = {Sketch drawings capture the salient information of visual concepts. Previous work has shown that neural networks are capable of producing sketches of natural objects drawn from a small number of classes. While earlier approaches focus on generation quality or retrieval, we explore properties of image representations learned by training a model to produce sketches of images. We show that this generative, class-agnostic model produces informative embeddings of images from novel examples, classes, and even novel datasets in a few-shot setting. Additionally, we find that these learned representations exhibit interesting structure and compositionality.} }
Endnote
%0 Conference Paper %T SketchEmbedNet: Learning Novel Concepts by Imitating Drawings %A Alexander Wang %A Mengye Ren %A Richard Zemel %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-wang21s %I PMLR %P 10870--10881 %U https://proceedings.mlr.press/v139/wang21s.html %V 139 %X Sketch drawings capture the salient information of visual concepts. Previous work has shown that neural networks are capable of producing sketches of natural objects drawn from a small number of classes. While earlier approaches focus on generation quality or retrieval, we explore properties of image representations learned by training a model to produce sketches of images. We show that this generative, class-agnostic model produces informative embeddings of images from novel examples, classes, and even novel datasets in a few-shot setting. Additionally, we find that these learned representations exhibit interesting structure and compositionality.
APA
Wang, A., Ren, M. & Zemel, R.. (2021). SketchEmbedNet: Learning Novel Concepts by Imitating Drawings. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:10870-10881 Available from https://proceedings.mlr.press/v139/wang21s.html.

Related Material