Bias in Machine Learning Models for Depression Detection
9 Jun 2025
A systematic review of AI models for detecting depression on social media reveals key biases that limit accuracy, generalizability, and ethical application.
How Social Media Data Is Used to Detect Depression
9 Jun 2025
Systematic review of 47 studies on using social media for depression detection, revealing key biases, sampling flaws, and machine learning model insights.
AI, Depression & Social Media: What New Research Tells Us
9 Jun 2025
A systematic review of how AI and machine learning models detect depression using social media data from platforms like Twitter, Facebook, and Reddit.
Can AI Tell When You’re Depressed?
9 Jun 2025
Systematic review shows how bias and poor methodology limit ML models used to detect depression through social media posts.
Understanding the Role of Neuron Weights in AI Model Performance
30 May 2025
Explore how neuron weight and activation correlate with AI bias and generation quality—and what it tells us about model collapse and mitigation.
Measuring Text Decay in AI
30 May 2025
How GPT-2's text quality deteriorates over generations using deterministic, beam, and nucleus sampling. Includes real examples and perplexity analysis.
How Bias Amplifies Across AI Generations
29 May 2025
Investigates WMLE and qualitative frameworks to reveal how AI models amplify political bias in media-style text across generations.
GPT-2 Study Shows How Language Models Can Amplify Political Bias
29 May 2025
GPT-2 experiments reveal how LLMs can amplify bias during training, raising urgent ethical concerns about fairness, transparency, and model accountability.
How GPT-2 Gets More Politically Biased Over Time
29 May 2025
Study shows how GPT-2 amplifies political bias and loses text quality through self-training—and explores strategies to prevent AI model collapse.