Our Newsletters and Articles
Latest Newsletters & Articles
Being aware of the permanent need to increase the professional skills, we take care of sharing topics of interest for the field in which we operate
-
From Accuracy to Reliability: How Conformal Prediction Is Changing Machine Learning Adoption
Conformal Prediction brings statistically calibrated confidence intervals to machine learning, overcoming the limitations of point predictions. This technique enables evaluation and communication of uncertainty at the individual level, increasing the reliability and transparency of ML models. Integrating CP can accelerate ML adoption in organizations, providing a competitive edge in risk management and automated processes.
-
Artificial Intelligence in Migration Management: Ethical Risks and Necessary Limits
The use of AI in migration processes exposes migrants to the risk of discrimination and automated decisions that are difficult to challenge. Sensitive data and lack of transparency exacerbate these risks. Regulations, independent oversight, and professional accountability are necessary to protect fundamental rights.
-
What Does the Quantization Threshold Miss? Two Essential Numbers for Optimizing AI Models (in Plain English)
Quantizing AI models helps with efficiency, but it doesn’t always show true performance. Two additional metrics—variation on real-world data and impact on critical cases—are essential for better business decisions. Don’t stop at the overall average when optimizing AI!