PhD defence

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).

Ricardo Salmon, PhD candidate
David R. Cheriton School of Computer Science

Stochastic satisfiability (SSAT), Quantified Boolean Satisfiability (QBF) and decision theoretic planning in infinite horizon partially observable Markov decision processes (POMDPs) are all PSPACE-Complete problems. Since they are all complete for the same complexity class, I show how to convert them into one another in polynomial time and space.

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.

Speaker: Junnan Chen, Master’s candidate

Conversations depend on information from the context. To go beyond one-round conversation, a chatbot must resolve contextual information such as: 1) co-reference resolution, 2) ellipsis resolution, and 3) conjunctive relationship resolution.

Speaker: Amir-Hossein Karimi, Master’s candidate

The story of this work is dimensionality reduction. Dimensionality reduction is a method that takes as input a point-set P of n points in \(R^d\) where d is typically large and attempts to find a lower-dimensional representation of that dataset, in order to ease the burden of processing for down-stream algorithms. In today’s landscape of machine learning, researchers and practitioners work with datasets that either have a very large number of samples and/or include high-dimensional samples. Therefore, dimensionality reduction is applied as a pre-processing technique primarily to overcome the curse of dimensionality.

Speaker: Zhucheng Tu, Master's Candidate

Modelling the similarity of two sentences is an important problem in natural language processing and information retrieval, with applications in tasks such as paraphrase identification and answer selection in question answering. The Multi-Perspective Convolutional Neural Network (MP-CNN) is a model that improved previous state-of-the-art models in 2015 and has remained a popular model for sentence similarity tasks. However, until now, there has not been a rigorous study of how the model actually achieves competitive accuracy. 

Speaker: Zhengkun Shang, Master's Candidate

Emotions are an essential part of human social interactions. By integrating an automatic affect recognizer into an artificial system, the system can detect humans’ emotions and provide personal responses. We aim to build a prompting system that uses a virtual human with emotional interaction capabilities to help persons with a cognitive disability to complete daily activities independently.

Speaker: Ali Wytsma, Master's Candidate

The naive Bayes model is a simple model that has been used for many decades, often as a baseline, for both supervised and unsupervised learning. With a latent class variable it is one of the simplest latent variable models, and is often used for clustering. The estimation of its parameters by maximum likelihood (e.g., gradient ascent, expectation maximization) is subject to local optima since the objective is non-concave. However, the conditions under which global optimality can be guaranteed are currently unknown.

Speaker: Areej Alhothali, PhD Candidate

Natural language texts are often meant to express or impact individuals emotions. Sentiment analysis researchers are increasingly interested in investigating natural language processing techniques as well as emotions theories to classify sentiments expressed in natural language text. Most sentiment analysis research effort focuses on classifying highly opinionated documents from the writer's perspectives and uses either count-based word representations that ignore sentence structure or a small set of lexicon resources that do not cover the wide range of words used on the Internet.

Speaker: Mariah Shein, Master's Candidate

The development of the field of reinforcement learning was based on psychological studies of the instrumental conditioning of humans and other animals. Recently, reinforcement learning algorithms have been applied to neuroscience to help characterize neural activity and animal behaviour in instrumental conditioning tasks. A specific example is the hybrid learner developed to match human behaviour on a two-stage decision task. This hybrid learner is composed of a model-free and a model-based system.