Seminar

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

The wavelet transform has seen success when incorporated into neural network architectures, such as in wavelet scattering networks. More recently, it has been shown that the dual-tree complex wavelet transform can provide better representations than the standard transform.

Michael Cormier, PhD candidate

This thesis is focused on the development of computer vision techniques for parsing web pages using an image of the rendered page as evidence, and on understanding this under-explored class of images from the perspective of computer vision. This project is divided into two tracks — applied and theoretical — which complement each other. Our practical motivation is the application of improved web page parsing to assistive technology, such as screenreaders for visually impaired users or the ability to declutter the presentation of a web page for those with cognitive deficit. From a more theoretical standpoint, images of rendered web pages have interesting properties from a computer vision perspective; in particular, low-level assumptions can be made in this domain, but the most important cues are often subtle and can be highly non-local. The parsing system developed in this thesis is a principled Bayesian segmentation-classification pipeline, using innovative techniques to produce valuable results in this challenging domain. The thesis includes both implementation and evaluation solutions.

Priyank Jaini, PhD candidate
David R. Cheriton School of Computer Science

At their core, many unsupervised learning models provide a compact representation of homogeneous density mixtures, but their similarities and differences are not always clearly understood. In this work, we formally establish the relationships among latent tree graphical models (including special cases such as hidden Markov models and tensorial mixture models), hierarchical tensor formats and sum-product networks.

Nisarg Shah, Department of Computer Science
University of Toronto

Algorithms are increasingly making decisions that affect humans. The field of computational social choice deals with algorithms for eliciting individual preferences and making collective decisions. Everyday examples of such decisions include citizens electing their representatives, roommates dividing collectively purchased items, or residents voting over allocation of city's budget. Making reasonable collective decisions requires viewing the problem through the lenses of elicitation, fairness, efficiency, incentives, and ethics.

Speaker: Ivana Kajić, PhD candidate

The representation of semantic knowledge poses a central modelling decision in many models of cognitive phenomena. However, not all such representations reflect properties observed in human semantic networks. Here, we evaluate the psychological plausibility of two distributional semantic models widely used in natural language processing: word2vec and GloVe. We use these models to construct directed and undirected semantic networks and compare them to networks of human association norms using a set of graph-theoretic analyses. 

Speaker: Meng Tang, PhD candidate

Minimization of regularized losses is a principled approach to weak supervision well established in deep learning, in general. However, it is largely overlooked in semantic segmentation currently dominated by methods mimicking full supervision via "fake" fully-labeled training masks (proposals) generated from available partial input. To obtain such full masks the typical methods explicitly use standard regularization techniques for "shallow" segmentation, e.g., graph cuts or dense CRFs. In contrast, we integrate such standard regularizers and clustering criteria directly into the loss functions over partial input. This approach simplifies weakly-supervised training by avoiding extra MRF/CRF inference steps or layers explicitly generating full masks, while improving both the quality and efficiency of training. 

Speaker: Daniel Recoskie, PhD candidate

We propose a new method for learning filters for the 2D discrete wavelet transform. We extend our previous work on the 1D wavelet transform in order to process images. We show that the 2D wavelet transform can be represented as a modified convolutional neural network (CNN). Doing so allows us to learn wavelet filters from data by gradient descent. Our learned wavelets are similar to traditional wavelets which are typically derived using Fourier methods.