Browsing by Subject "AI"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Open Access
    Evaluating deep learning for enhanced breast cancer diagnosis: a comparative analysis of CNN architectures
    (2025) Frankle, Solyle; Sinkala, Musalula
    Artificial Intelligence (AI), particularly its machine learning (ML) subfield, has revolutionised various sectors, including healthcare. In breast cancer care, AI's ability to analyse vast datasets and extract complex patterns from medical images has the potential to transform diagnostics and treatment strategies. Breast cancer remains one of the most prevalent cancers affecting women globally, with early and accurate diagnosis being crucial for effective treatment. AI, through its advanced image analysis capabilities, significantly improves the accuracy and efficiency of breast cancer diagnosis, specifically in distinguishing between cancer subtypes. Here, we aim to explore the application of deep learning, particularly convolutional neural networks (CNNs), in breast cancer subtype classification using histology images. A custom CNN model, alongside well-established models like ResNet50 and EfficientNetB0, was developed and evaluated for its accuracy in predicting benign and malignant breast cancer subtypes. The results demonstrated that while the custom CNN achieved an accuracy of 65% for malignant and 67% for benign subtypes with ROC-AUC scores of 0.86 and 0.90, respectively, ResNet50 significantly outperformed both the custom model and EfficientNetB0. ResNet50 attained an accuracy of 77% in classifying malignant subtypes and 77% for benign subtypes, accompanied by ROC-AUC scores of 0.92 and 0.96, respectively. Additionally, ResNet50 exhibited higher precision (0.68 for malignant, 0.67 for benign), recall (0.65 for malignant, 0.67 for benign), and F1 scores (0.65 for malignant, 0.67 for benign) across most subtypes, underscoring its robust performance and reliability in clinical settings. In conclusion, AI, specifically through advanced CNN architectures, can greatly enhance breast cancer diagnosis by providing more accurate subtype classifications. Future work should focus on integrating these models into clinical workflows, enabling faster and more personalised treatment planning. Moreover, continued refinement of these models, including addressing the complexities of tumour heterogeneity and incorporating multimodal data, will be crucial for their widespread adoption in oncology.
  • No Thumbnail Available
    Item
    Open Access
    Examining personality assessment in asynchronous video interviews (AVI): convergence between human personality judgements and AI/ML scoring
    (2025) Cronje, Jacobus Fouche; de Kock, Francois
    The assessment of personality is an essential component of personnel selection due to its validity in predicting job performance. To assess personality, asynchronous video interviews (AVIs) scored using artificial intelligence (AI) algorithms are increasingly used, allowing candidates to record responses to interview prompts that are subsequently evaluated automatically by AI algorithms and/or human raters. As questions remain about the validity of AI-based AVI scoring approaches, this study examines the convergence between human-and AI-scored personality assessments. To measure personality, the study focuses on the HEXACO model, which measures Honesty-Humility, Emotionality, Extraversion, Agreeableness, Conscientiousness, and Openness to Experience. Verbal responses were transcribed from videotaped AVIs of 161 mock interview candidates who answered five AVI questions. Responses were scored by 15 trained human raters and a closed-dictionary text-analysis keyword-counting AI algorithm developed for this study, respectively. The correlation between trait-level scores produced by human judges and AI scoring was tested both across traits and within traits (trait-level) to assess scoring convergence. Moreover, in addition to comparing score levels produced by the two scoring methods (AI vs. human raters), score spread (i.e., variability), rank-order stability, and rating reliability were evaluated. The findings revealed a moderately positive and significant overall convergence (r = .29, p < .001) across traits between human and AI evaluations, which suggests that AI scoring may potentially be useful as a replacement of human evaluations when general screening is desired. Trait-level convergence varied between scoring methods, with the scoring consensus between human raters and AI being higher for some traits than for others, suggesting that these methods rely on different information and/or may interpret interview responses differently. The research highlights the potential of AI to complement human- based scoring of AVIs used in recruitment, selection, and assessment while also identifying the limitations of algorithm-based scoring in capturing complex human behaviour in interviews. The findings may further contribute to understanding the role of AI in personality assessment and implications for organisational practices.
  • No Thumbnail Available
    Item
    Open Access
    Use of ChatGPT for student co-creation of open textbooks
    (Digital Open Textbooks for Development, 2024-02) Cox, Glenda; Willmers, Michelle; Held, Michael; Brown, Robyn
    This is a presentation by members of the Digital Open Textbooks for Development (DOT4D) initiave, Asso. Prof Glenda Cox and Michelle Willmers, and collaborators Dr Michael Held and Robyn Brown, as part of the Centre for Innovation in Learning and Teaching's (CILT) Brown Bag seminar series in February 2024.