0:00 / 0:00
research
Learning From Less, with Origami | Researcher Spotlight: Emanuele Sansone
MIT CSAIL Alliances22 June 2026Watch on YouTube
Description
Most modern AI systems rely on massive datasets and enormous computing power. But what if AI could learn more like humans do? MIT CSAIL Postdoctoral researcher Emanuele Sansone explains how his work focuses on building AI systems that can learn from limited data. Using origami as a unique testing ground, his research explores how structured models can acquire knowledge in domains where large datasets simply don't exist. This work could help shape a new generation of AI that is more efficient, flexible, and capable of learning with far less information than today's typical models. #MIT #MITCSAIL #ArtificialIntelligence #MachineLearning #AIResearch #OrigamiAI #FutureOfAI #ScienceAndTechnology
What you'll learn
- Modern AI systems can learn more efficiently through structured models rather than massive datasets
- Origami serves as a practical testing ground for exploring machine learning with limited data
- Structured AI can acquire knowledge in domains where large datasets are unavailable
Frequently asked questions
Why does Emanuele Sansone use origami for his AI research?
Origami is a unique testing domain where large datasets don't exist, making it ideal for exploring how AI systems can learn with limited data.
How does Sansone's approach differ from typical modern AI systems?
Rather than relying on massive datasets and enormous computing power, his research focuses on structured models that can learn more like humans do.
What are potential benefits of AI systems that can learn from limited data?
Such systems could be more efficient, flexible, and suitable for domains without large datasets, leading to a new generation of more practical AI.