Update on work with the Military Policies and Standards Chatbot.
Navigating through the extensive and bilingual military policies of the Canadian Armed Forces (CAF) can be a complex task. To address the need for streamlined access to these documents, this paper explores the implementation of containerized artificial intelligence (AI) models to develop a bilingual knowledge management system. This system is designed to enhance information retrieval from military policies by leveraging AI, metadata tagging, text embeddings, and vector databases. Our approach involves creating a comprehensive data processing pipeline to store and retrieve military policies and documents. We utilize a question-answering pipeline that incorporates language detection, semantic search, and large language models (LLM) to process queries in both English and French. Preliminary evaluations demonstrate an accuracy rate of 87.78\% for English and 73.33\% for French queries. This paper builds upon our previous work on semantic search for military policies by integrating a generative AI component, thus extending beyond traditional retrieval-based methods to provide more precise and contextually relevant responses.

In the Canadian Armed Forces (CAF), navigating through extensive policies and standards can be a challenging task. To address the need for streamlined access to these vital documents, this paper explores the usage of artificial intelligence (AI) and natural language processing (NLP) to create a question-answering chatbot. This chatbot is specifically tailored to pinpoint and retrieve specific passages from policy documents in response to user queries. Our approach involved first developing a comprehensive and systematic data collection technique for parsing the multi-formatted policy and standard documents. Following this, we implemented an advanced NLP-based information retrieval system to provide the most relevant answers to users’ questions. Preliminary user evaluations showcased a promising accuracy rate of 88.46%. Even though this chatbot is designed to operate on military policy documents, it can be extended for similar use cases to automate information retrieval from long documents.

This paper outlines a robust approach to automate the detection of military badges on official government documents utilizing YOLOv5 computer vision model. In an era where the rapid classification and management of sensitive documents is paramount, developing a system capable of accurately identifying and classifying distinct badge types plays a crucial role in supporting data management and security protocols. To address the challenges posed by the lack of accessible, real-world government and military documents for research, we introduced a novel method to simulate training data. We employ a technique that automates the data labelling process, facilitating the generation of a comprehensive and versatile dataset while eliminating the risk of compromising sensitive information. Through careful model training and hyper-parameter tuning, the YOLOv5 model demonstrated exemplary performance, successfully detecting a wide spectrum of badge types across various documents.

Our study suggests that individuals using more linguistic shibboleths are more likely to have a stronger bond with this online community and develop a tendency to display characteristics of mental illness themselves. We believe that our study has several far-reaching consequences, including a new psycholinguistic lens through which to examine propaganda and an explanation of echo chambers, parasocial relationships, cults of personality, and other socio-psychological phenomena

Barrett, R.C.A., Poe, R., O’Camb, J.W., Woodruff, C., Harrison, S.M., Dolguikh, K., Chuong, C., Klassen, A.D., Zhang, R., Joseph, R.B. and Blair, M.R., 2022. Comparing virtual reality, desktop-based 3D, and 2D versions of a category learning experiment. Plos one, 17(10), p.e0275119.

Starting work into testing the viability of immersive virtual reality in conducting cognitive science research. Using the category learning paradigm, we are exploring how people change their behaviour over time as they learn to categorize different stimuli into their appropriate groups.

The ADP database and directory is designed for reflection, study and analysis of online literary events in Canada during the COVID-19 pandemic. This paper presents our approaches to data collection and structuring, stack development, data visualization, and front-end design, pursued through community-oriented processes of design, research and development.

In 2019, SpokenWeb SFU Project Manager Cole Mash (SFU) and SpokenWeb Systems Task Force member Tomasz Neugebauer (Concordia) began work on editing SWALLOW entries. SWALLOW is an open-source metadata ingestion system developed by the SpokenWeb team to describe and manage the project’s object of study: literary audio. Since the implementation of SWALLOW in 2018, SpokenWeb team members have ingested over 4,700 entries into the system..

Our aim is to facilitate investigation of potential variants by presenting a computational corpus approach towards measuring one of many ontological parameters: linguistic distinctiveness. Using the HKCSE and the BNC, we perform part-of-speech tagging, dependency parsing, and word order assessment to identify morpho-syntactic features.

The objective of this study is to develop a model that can take in the content of a microblog (such as a tweet on Twitter) that contains opinions (or sentiments), and determine the prevalent emotions behind the tweet as a function of uncertainty about the membership functions of fuzzy sets associated with linguistic terms. This information can then be used for sentiment analysis.

An exploration of my process developing a bot to traverse through the metadata ingestion system in order to facilitate editing large volumes of data, usage of NLP word prediction techniques to fill in gaps in areas with missing data and supervised learning to automate the data entry process. Besides this, I also include considerations of more experimental applications such as lip reading.

An exploration into illustrating linguistic trends in the age of misinformation with respect to targeted advertisements creating echo chambers, manufactured consent, subjective language, online disinhibition, and how news organizations create newsworthiness.

In this review, I look at the history of these pieces of code, from Socrates’ Daemon, to the Child Machine; and examine the applications these bots could currently have in Computing Science in general and as an answer to the Turing Test in Specific.

We use the real-time strategy game StarCraft 2 in our lab as it is a useful domain in which to study learning and attention. I was involved with a project that utilized Starcraft 2 Gameplay Chatlogs datasets and used NLP techniques such as frequency mapping and NER to attempt to classify positive or negative expression of emotion during gameplay

ExNovo is aimed at designing and testing a new computer interface that is grounded in what we know about human cognition. I played a minor role in correcting C# code for the Ex-Novo game, testing equipment before experiments and re-formatting data

Minor contribution to data management of field notes etc. for the Hul’q’umi’num’ Language & Culture Society (HLCS). HLCS is a not-for-profit organization whose mandate is to ensure the survival of traditional knowledge and values. Our work proceeds under the guidance of a board of Elders, teachers, and researchers, all speakers of the Hul’q’umi’num’ language.

Web Page

Minor contributions to a variety of projects involving the study of clefts, information structure and intonation