A*STAR Centre for Frontier AI Research (A*STAR CFAR)

By reasoning and accessing knowledge in an alternative language, large language models can produce more reliable answers in the user’s original language.

A new approach to training AI models can help improve fairness in decision outcomes for the most disadvantaged groups even when their identities are unknown.

A new comprehensive benchmark for audio-based artificial intelligence allows researchers to gauge and improve the way models follow speech, soundscapes and everyday cues.

From biomedical research to generative AI models, scientific progress depends on strong ethical standards that maintain integrity, transparency and accountability.

A*STAR frontiers in the national push for a quantum computing advantage

Researchers teach artificial intelligence systems how to process information by mimicking human introspection.

As machines take their first full-fledged steps into the world of human language, A*STAR is paving the way through innovative work in natural language processing.

A new framework gives pre-trained machine learning models the ability to perform optimally in real-world settings, even with limited training data.