Pavlos Andreadis

Pavlos Andreadis

Auto-generated profile, based on publicly available information

Pavlos Andreadis

School of Informatics, University of Edinburgh, Edinburgh, United Kingdom

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 March 2026 in Journal of Artificial Intelligence Research
Verified
Journal Article
Closed access
PDF file
PKPatrick Kage
Patrick Kage
JRJay Rothenberger
Jay Rothenberger
PAPavlos Andreadis
Pavlos Andreadis
DDDimitrios Diochnos
Dimitrios Diochnos
Deep neural models have achieved state-of-the-art performance on a wide range of problems in computer science, especially in computer vision. However, deep neural networks often require large datasets of labeled samples to generalize effectively. An important area of active research is semi-supervised learning, which attempts to instead utilize large quantities of (easily acquired) unlabeled samples. One family of methods in this space is pseudo-labeling, a class of algorithms that use model outputs to assign labels to unlabeled samples which are then used as labeled samples during training. Such assigned labels, called pseudo-labels, are most commonly associated with the field of semi-supervised learning. In this work, we explore a broader interpretation of pseudo-labels within both self-supervised and unsupervised methods. After a thorough treatment of pseudo-labeling in these areas, we draw the connection between them and identify commonalities between fields, as well as new directions where advancements in one area would likely benefit others, such as curriculum learning and self-supervised regularization.

13 August 2024
Verified
Conference Paper
Open access
MVMarc Violides
Marc Violides
TTTanatip Timtong
Tanatip Timtong
KBKrystof Bezdek
Krystof Bezdek
PAPavlos Andreadis
Pavlos Andreadis
The COVID-19 pandemic has profoundly impacted mental health worldwide, potentially altering the linguistic expression of depression in online communities. This study investigates these changes by analyzing Reddit posts from depression-related and other mental health subreddits before and after the pandemic declaration. We employ a novel methodology combining advanced Natural Language Processing (NLP) techniques with statistical analysis. Fine-tuning the roBERTa model on the DepressionEmo dataset [24] achieves state-of-the-art performance in identifying depression-related emotions, with 80% precision, 82% recall, and 81% F1-score. Chi-square tests reveal significant post-pandemic increases in the expression of loneliness, sadness, suicide intent, and worthlessness (p < 0.001) in depression subreddits compared to other mental health forums. Anger and brain dysfunction exhibit divergent trends between the two groups. These findings underscore the pandemic's differential impact on depressive language, highlight the need for targeted online mental health support, and demonstrate the potential of interdisciplinary methodologies in analyzing mental health discourse at scale.