A5015584294- Computational and Text Analysis Methods
- Commonwealth, Australian Politics and Federalism
- Data Analysis with R
- Big Data Technologies and Applications
- Electoral Systems and Political Participation
- Big Data and Business Intelligence
- Scientific Computing and Data Management
- Migration and Labor Dynamics
- Biomedical Text Mining and Ontologies
- Hate Speech and Cyberbullying Detection
- Academic Publishing and Open Access
- Topic Modeling
- Urban, Neighborhood, and Segregation Studies
- Natural Language Processing Techniques
- Data-Driven Disease Surveillance
- Statistical Methods and Inference
- scientometrics and bibliometrics research
- Software Engineering Techniques and Practices
- COVID-19 Pandemic Impacts
- Data Quality and Management
- Advanced Causal Inference Techniques
- Social Policy and Reform Studies
- Advanced Text Analysis Techniques
- Hydrology and Drought Analysis
- Business Process Modeling and Analysis
University of Toronto
2020-2023
North Bristol NHS Trust
2020
Australian National University
2018-2019
University of Victoria
1993
Sophisticated language models such as OpenAI's GPT-3 can generate hateful text that targets marginalized groups. Given this capacity, we are interested in whether large be used to identify hate speech and classify sexist or racist. We use racist passages with zero-, one-, few-shot learning. find zero- one-shot learning, an average accuracy between 55 per cent 67 cent, depending on the category of type With model's high 85 cent. Large have a role play detection, further development they could...
We estimate the effect of age at arrival for immigrant outcomes with a new dataset arrivals linked to 1940 U.S. Census. Using within-family variation, we find that arriving an older age, or having more childhood exposure European environment, led negative wage gap relative native born. Infant had positive natives, in contrast teenage arrivals. Therefore, key determinant during Age Mass Migration was country residence critical periods development.
To examine the reproducibility of COVID-19 research, we create a dataset pre-prints posted to arXiv, bioRxiv, and medRxiv between 28 January 2020 30 June 2021 that are related COVID-19. We extract text from these parse them looking for keyword markers signaling availability data code underpinning pre-print. For in our sample, unable find either open or 75% those on 67% 79% medRxiv.
We investigated whether large language models (LLMs) can develop data validation tests. considered 96 conditions each for both GPT-3.5 and GPT-4, examining different prompt scenarios, learning modes, temperature settings, roles. The scenarios were: 1) Asking expectations, 2) expectations with a given context, 3) after requesting simulation, 4) provided sample. modes zero-shot, one-shot, few-shot learning. also tested four settings: 0, 0.4, 0.6, 1. And the two distinct roles helpful...
The purpose of this research is to know the perspectives technological pedagogical environment and teaching adaptation in times pandemic Ecuador regarding online education within framework health emergency Ecuadorian middle-level institutions; since, current represents a methodological challenge characterized by digital tools during teaching, processes that have abruptly modified space forms teaching. Therefore, retrospective level study assisted case methodology used as data collection...
Public knowledge of what is said in parliament a tenet democracy, and critical resource for political science research. In Australia, following the British tradition, written record known as Hansard. While Australian Hansard has always been publicly available, it difficult to use purpose large-scale macro- micro-level text analysis because only available PDFs or XMLs. Following lead Linked Parliamentary Data project which achieved this Canada, we provide new, comprehensive, high-quality,...