Investigate and discuss domains in which bilinear methods are applied

[vdumoulin]

Besides VQA, do we know of more domains in which bilinear methods have been effectively applied?

[fstrub95]

Mieux vaut tard que jamais :) I tried to keep it short and extensif. I only keep papers that have an extensive number of citations(200 in average)

However, I am not sure where I can integrate this paragrah:

Bilinear models were first introduced by the vision community by to better untangle latent perceptual factors~\cite{tenenbaum1997separating}. As described by the paper's title, the authors wanted to separate the image style from its content, arguing that classic linear models were not rich enough to extract such complex interaction. There, they demonstrate the effectiveness of their approach by applying it for spoken vowel identification or zero-shot font classification. Among other the firsts bilinear applications, one remarkable case was to perform facial animation~\cite{chuang2002facial}. While the style describes the key facial features, the content encodes the visual emotions. This allows, for example, a sequence originally recorded with a happy expression to be modified so that the speaker appears to be speaking with an angry or neutral expression Bilinear models are also present in recommendation system by extracting user and item information in various settings~\cite{chu2009personalized,yang2011like}. More generally, recommendation system heavily rely on matrix factorization methods~\cite{koren2009matrix}, which is a bilinear model where one of the latent is fixed~\cite{tenenbaum1997separating}. More recently, bilinear models have inspired new neural architecture in visual recognition~\cite{lin2015bilinear}, video action recogintion~\cite{feichtenhofer2016convolutional} or visual question-answering~\cite{fukui2016multimodal}.

@inproceedings{tenenbaum1997separating, title={Separating style and content}, author={Tenenbaum, Joshua B and Freeman, William T}, booktitle={Advances in neural information processing systems}, year={1997} }

@inproceedings{chuang2002facial, author={E. S. Chuang and F. Deshpande and C. Bregler}, title={Facial expression space learning}, booktitle={10th Pacific Conference on Computer Graphics and Applications, 2002. Proceedings.}, year={2002} }

@inproceedings{yang2011like, title={Like like alike: joint friendship and interest propagation in social networks}, author={Yang, Shuang-Hong and Long, Bo and Smola, Alex and Sadagopan, Narayanan and Zheng, Zhaohui and Zha, Hongyuan}, booktitle={Proceedings of the 20th international conference on World wide web}, year={2011} }

@inproceedings{chu2009personalized, title={Personalized recommendation on dynamic content using predictive bilinear models}, author={Chu, Wei and Park, Seung-Taek}, booktitle={Proceedings of the 18th international conference on World wide web}, year={2009} }

@inproceedings{lin2015bilinear, title={Bilinear cnn models for fine-grained visual recognition}, author={Lin, Tsung-Yu and RoyChowdhury, Aruni and Maji, Subhransu}, booktitle={Proceedings of the IEEE International Conference on Computer Vision}, year={2015} }

@inproceedings{feichtenhofer2016convolutional, title={Convolutional two-stream network fusion for video action recognition}, author={Feichtenhofer, Christoph and Pinz, Axel and Zisserman, AP}, booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition}, year={2016} }

@inproceedings{fukui2016multimodal, title={Multimodal compact bilinear pooling for visual question answering and visual grounding}, author={Fukui, Akira and Park, Dong Huk and Yang, Daylen and Rohrbach, Anna and Darrell, Trevor and Rohrbach, Marcus}, booktitle={Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing}, year={2016} }

@article{koren2009matrix, title={Matrix factorization techniques for recommender systems}, author={Koren, Yehuda and Bell, Robert and Volinsky, Chris}, journal={Computer}, volume={42}, number={8}, year={2009}, publisher={IEEE} }

[vdumoulin]

@fstrub95 I would suggest putting it as a bibliographic note in the appendix, like the CTC article does (see this relevant portion of the source code).

We can later on add a reference to the bibliographic note in the main text.

[fstrub95]

Done -> https://github.com/vdumoulin/distill-film/pull/129/files