Faculty

Tuesday, September 25, 2018 1:00 pm - 1:00 pm EDT (GMT -04:00)

PhD Defence: Learning Sparse Orthogonal Wavelet Filters

Daniel Recoskie, PhD candidate
David R. Cheriton School of Computer Science

The wavelet transform is a well-studied and understood analysis technique used in signal processing. In wavelet analysis, signals are represented by a sum of self-similar wavelet and scaling functions. Typically, the wavelet transform makes use of a fixed set of wavelet functions that are analytically derived. We propose a method for learning wavelet functions directly from data. We impose an orthogonality constraint on the functions so that the learned wavelets can be used to perform both analysis and synthesis. We accomplish this by using gradient descent and leveraging existing automatic differentiation frameworks. Our learned wavelets are able to capture the structure of the data by exploiting sparsity. We show that the learned wavelets have similar structure to traditional wavelets.

Vineet John, Master’s candidate
David R. Cheriton School of Computer Science

This thesis tackles the problem of disentangling the latent style and content variables in a language modelling context. This involves splitting the latent representations of documents by learning which features of a document are discriminative of its style and content, and encoding these features separately using neural network models.

Royal Sequiera, Master’s candidate
David R. Cheriton School of Computer Science

With the advent of deep learning methods, researchers are abandoning decades-old work in Natural Language Processing (NLP). The research community has been increasingly moving away from otherwise dominant feature engineering approaches; rather, it is gravitating towards more complicated neural architectures. Highly competitive tools like Parts-of-Speech taggers that exhibit human-like accuracy are traded for complex networks, with the hope that the neural network will learn the features needed. In fact, there have been efforts to do NLP "from scratch" with neural networks that altogether eschew featuring engineering based tools (Collobert et al., 2011).

Professor Jeff Orchard and third-year undergraduate computer science student Louis Castricato received a best paper award at the 24th International Conference on Neural Informational Processing (ICONIP 2017) for their paper titled “Combating adversarial inputs using a predictive-estimator network.”

Read the full story

When an election is held we often employ a peculiar kind of logic. As we mull over the candidates we may have a top choice, but if we think our preferred candidate isn’t going to win we might vote for our second choice. Or maybe we cast a ballot for our second choice because we want to make sure that a frontrunner who doesn’t represent our view loses.

We live in a world increasingly dependent on the Internet for information retrieval, social interaction and general leisure. A growing number of Internet users with cognitive or visual impairments need assistive technology to make information accessible to them, but visually complex web pages can be difficult to navigate for assistive technology.

When you look at a scenic mountain photo typically everything in the distance is in sharp focus. But this scene might be even more captivating if something striking were in the foreground, perhaps a field of wild flowers in peak bloom. The problem is if the flowers are close to the lens relative to the mountains it’s impossible for all elements in the photo to be in perfect focus — if the flowers are sharp, the distant mountains will be blurry and vice versa.

Professor Robin Cohen has received a Lifetime Achievement Award from the Canadian Artificial Intelligence Association. She is the first female recipient of the Association’s highest honour, an award that is conferred to individuals who have distinguished themselves through outstanding research excellence in artificial intelligence during the course of their academic career.