The models and ideas behind the projects — grouped by area, mapped to your track.
What it is: Layers of simple units that learn patterns from examples.
Where it is used: The base of almost every modern AI system.
What it is: Models trained on huge text corpora that understand and generate language.
Where it is used: Chatbots, writing help, code, tutoring.
What it is: Teaching computers to interpret images and video.
Where it is used: Recognizing objects, scenes, and actions.
What it is: An architecture that uses attention to relate every part of the input.
Where it is used: Powers LLMs, vision, and speech models.
What it is: Learning by trial and reward to make good decisions over time.
Where it is used: Game AI, robotics, strategy.
What it is: Vision models that flag findings in scans and slides.
Where it is used: Triage support in clinics and radiology.
What it is: Language models that read clinical notes and guidelines.
Where it is used: Summarizing records, patient guidance.
What it is: Models that translate between languages while keeping meaning.
Where it is used: Preserving and sharing heritage languages.
What it is: Turning written text into natural spoken audio.
Where it is used: Accessibility, storytelling, archives.
What it is: Models that surface relevant content for each person.
Where it is used: Connecting communities to what matters.
What it is: Detecting body joints to analyze movement.
Where it is used: Technique, injury prevention, coaching.
What it is: Models that forecast outcomes from historical data.
Where it is used: Scouting, tactics, fan engagement.
What it is: Following players and the ball across a match.
Where it is used: Performance analytics and highlights.