I lead the Machine Learning Research Group at the University of Guelph. I am interested in statistical machine learning and biologically-inspired computer vision, with an emphasis on deep learning and time series analysis.

News

Brief Biography

I received my PhD in Computer Science from the University of Toronto in 2009, where I was advised by Geoffrey Hinton and Sam Roweis. I spent two years as a postdoc at the Courant Institute of Mathematical Sciences, New York University working with Chris Bregler, Rob Fergus, and Yann LeCun. In 2012, I joined the School of Engineering at the University of Guelph as an Assistant Professor. In 2017, I was promoted to Associate Professor and became a member of the Vector Institute for Artificial Intelligence. In 2018, I was honoured as one of Canada's Top 40 under 40. In 2019, I was named a Canada CIFAR AI Chair. I spent 2018-2019 as a Visiting Faculty member at Google Brain, Montreal.

Research Highlights

A complete list of my publications is available on Google Scholar.

Terrance DeVries, Miguel Angel Bautista, Nitish Srivastava, Graham Taylor, and Joshua M. Susskind. Unconstrained scene generation with locally conditioned radiance fields. arXiv preprint arXiv:2104.00670, 2021. [ bib | http ]

Brendan Duke, Abdella Ahmed, Christian Wolf, Parham Aarabi, and Graham Taylor. SSTVOS: Sparse spatiotemporal transformers for video object segmentation. In Proc. of the 34th IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), 2021. [ bib | http ]

Rohit Saha, Brendan Dukeand Florian Shkurti, Graham Taylor, and Parham Aarabi. LOHO: Latent optimization of hairstyles via orthogonalization. In Proc. of the 34th IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), 2021. [ bib | http ]

Rylee Thompson, Elahe Ghalebi, Terrance DeVries, and Graham Taylor. Building LEGO using deep generative models of graphs. In Neural Information Processing Systems (NeurIPS) Workshop on Machine Learning for Engineering Modeling, Simulation, and Design, 2020. [ bib | http ]

Michal Lisicki, Arash Afkanpour, and Graham Taylor. Evaluating curriculum learning strategies in neural combinatorial optimization. In Neural Information Processing Systems (NeurIPS) Workshop on Learning Meets Combinatorial Algorithms, 2020. [ bib | http ]

Nolan Dey, Eric Taylor, Bryan Tripp, Alexander Wong, and Graham Taylor. Identifying and interpreting tuning dimensions in deep networks. In Neural Information Processing Systems (NeurIPS) Workshop on Shared Visual Representations in Human and Machine Intelligence, 2020. [ bib | http ]

Boris Knyazev, Michal Drozdzal, Graham Taylor, and Adriana Romero. Predicting pretrained weights of large-scale CNNs. In Neural Information Processing Systems (NeurIPS) Workshop on Beyond Backpropagation: Novel Ideas for Training Neural Architectures, 2020. [ bib ]

Terrance DeVries, Michal Drozdzal, and Graham Taylor. Instance selection for GANs. In Neural Information Processing Systems (NeurIPS), 2020. [ bib | http ]

Boris Knyazev, Harm de Vries, Cătălina Cangea, Graham Taylor, Aaron Courville, and Eugene Belilovsky. Graph density-aware losses for novel compositions in scene graph generation. In British Machine Vision Conference (BMVC), 2020. [ bib | http ]

Boris Knyazev, Harm de Vries, Cătălina Cangea, Graham Taylor, Aaron Courville, and Eugene Belilovsky. Generative graph perturbations for scene graph prediction. In International Conference on Machine Learning (ICML) Workshop on Object-Oriented Learning (OOL), 2020. [ bib | http ]

Eric Taylor, Shashank Shekhar, and Graham Taylor. Response time analysis for explainability of visual processing in CNNs. In IEEE CVPR Workshop on Minds vs. Machines: How Far Are We From the Common Sense of a Toddler?, 2020. [ bib | .html ]

Eu Wern Teh, Terrance DeVries, and Graham Taylor. ProxyNCA++: Revisiting and revitalizing proxy neighborhood component analysis. In European Conference on Computer Vision (ECCV), 2020. [ bib | http ]

Shivam Kalra, Mohammad Adnan, Graham Taylor, and Hamid Tizhoosh. Learning permutation invariant representations using memory networks. In European Conference on Computer Vision (ECCV), 2020. [ bib | http ]

Alaaeldin El-Nouby, Shuangfei Zhai, Graham Taylor, and Joshua Susskind. Skip-clip: Self-supervised spatiotemporal representation learning by future clip order ranking. In International Conference on Computer Vision (ICCV) Workshop on Holistic Video Understanding, 2019. [ bib | http ]

Terrance DeVries, Adriana Romero, Luis Pineda, Graham Taylor, and Michal Drozdzal. On the evaluation of conditional GANs. arXiv preprint arXiv:1907.08175, 2019. [ bib | http ]

Boris Knyazev, Graham Taylor, and Mohamed Amer. Understanding attention and generalization in graph neural networks. In Neural Information Processing Systems (NeurIPS), 2019. Early version appeared at the International Conference on Learning Representations (ICLR) Workshop on Representation Learning on Graphs and Manifolds. [ bib | http ]

Alaaeldin El-Nouby, Shikhar Sharma, Hannes Schulz, Devon Hjelm, Layla El Asri, Samira Ebrahimi Kahou, Yoshua Bengio, and Graham Taylor. Tell, draw, and repeat: Generating and modifying images based on continual linguistic instruction. In International Conference on Computer Vision (ICCV), 2019. Early version appeared at the Neural Information Processing Systems (NeurIPS) Workshop on Visually Grounded Interaction and Language (ViGIL). [ bib | http ]

Boris Knyazev, Xiao Lin, Mohamed Amer, and Graham Taylor. Image classification with hierarchical multigraph networks. In British Machine Vision Conference (BMVC), 2019. [ bib | http ]

Angus Galloway, Anna Golubeva, Thomas Tanay, Medhat Moussa, and Graham Taylor. Batch normalization is a cause of adversarial vulnerability. In International Conference on Machine Learning (ICML) Workshop on Identifying and Understanding Deep Learning Phenomena, 2019. [ bib | http ]

Angus Galloway, Anna Golubeva, and Graham Taylor. Adversarial examples as an input-fault tolerance problem. In Neural Information Processing Systems (NeurIPS) Workshop on Security in Machine Learning, 2018. [ bib | http ]

Boris Knyazev, Xiao Lin, Mohamed R. Amer, and Graham Taylor. Spectral multigraph networks for discovering and fusing relationships in molecules. In Neural Information Processing Systems (NeurIPS) Workshop on Machine Learning for Molecules and Materials, 2018. [ bib | http ]

Nihal Murali, Jonathan Schneider, Joel Levine, and Graham Taylor. Classification and re-identification of fruit fly individuals across days with convolutional neural networks. In IEEE Winter Conference on Applications of Computer Vision (WACV), pages 570--578, 2019. [ bib | http ]

Jonathan Schneider, Nihal Murali, Graham Taylor, and Joel Levine. Can Drosophila melanogaster tell who's who? PLOS One, 13(10), 2018. [ bib ]

Griffin Lacey, Graham Taylor, and Shawki Areibi. Stochastic layer-wise precision in deep neural networks. In Uncertainty in Artificial Intelligence (UAI), 2018. [ bib | http ]

Terrance DeVries and Graham Taylor. Leveraging uncertainty estimates for predicting segmentation quality. arXiv preprint arXiv:1807.00502, 2018. [ bib | http ]

Angus Galloway, Thomas Tanay, and Graham Taylor. Adversarial training versus weight decay. arXiv preprint arXiv:1802.04457, 2018. [ bib | http ]

Fabien Baradel, Christian Wolf, Julien Mille, and Graham Taylor. Glimpse clouds: Human activity recognition from unstructured feature points. In Proc. of the 31st IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), 2018. [ bib | http ]

Angus Galloway, Graham Taylor, and Medhat Moussa. Predicting adversarial examples with high confidence. arXiv preprint arXiv:1802.04457, 2018. [ bib | http ]

Terrance DeVries and Graham Taylor. Learning confidence for out-of-distribution detection in neural networks. arXiv preprint arXiv:1802.04865, 2018. [ bib | http ]

Daniel Jiwoong Im, He Ma, Graham Taylor, and Kristen Branson. Quantitatively evaluating GANs with divergences proposed for training. In International Conference on Learning Representations (ICLR), 2018. [ bib | http ]

Angus Galloway, Graham Taylor, and Medhat Moussa. Attacking binarized neural networks. In International Conference on Learning Representations (ICLR), 2018. [ bib | http ]

Group Members

Former Group Members