AI & Data Foundations

Hallucination in AI refers to models producing confident but factually incorrect outputs, posing risks in critical fields like healthcare and humanitarian work. Managing hallucination is essential for trust and safe AI adoption.

Large Language Models enable natural language interaction, lowering barriers to digital participation and supporting diverse sectors like education, health, and humanitarian response with adaptable AI applications.

Machine Learning is a key AI subfield driving social innovation by analyzing data to predict outcomes, improve interventions, and support sustainable development with responsible technology use.

Model evaluation for large language models ensures accuracy, fairness, and safety, helping organizations deploy AI responsibly across diverse sectors like education, healthcare, and humanitarian aid.

Multilingual models enable AI systems to understand and generate text across many languages, supporting inclusion, communication, and services in diverse sectors like education, healthcare, and humanitarian aid.

Multimodal models combine text, images, audio, and video to improve AI understanding and decision-making across diverse sectors like humanitarian aid, health, education, and agriculture.

Named Entity Recognition (NER) identifies and classifies key information in text, helping organizations analyze unstructured data for better decision-making across sectors like health, humanitarian work, and governance.

Natural Language Processing enables machines to understand and generate human language, breaking down linguistic barriers and supporting inclusion across sectors like education, health, and humanitarian aid.

Natural Language Understanding enables machines to comprehend human language meaning, intent, and context, improving communication and decision-making across sectors like healthcare, education, agriculture, and humanitarian work.

Optical Character Recognition (OCR) converts printed and handwritten text into machine-readable formats, enabling digitization of physical documents for improved accessibility, analysis, and integration in AI systems across various sectors.