World Models · Predictive Learning · Computer Vision · Embodied AI
Shiuan-Ting (Jeremy) Lin
I am a Solution Architect at Nokia with a foundation in mathematics and statistics. My work spans trustworthy machine learning, spatial representation analysis, and AI systems for network operations—with a growing interest in self-supervised and predictive representation learning, world models, and embodied intelligence.
A statistical lens on learning systems
I move between statistical reasoning, representation learning, and real-world AI systems. Across each setting, I care about what information a model retains, how reliably it generalizes, and whether its decisions can be inspected.
- Structure
Learning representations that preserve spatial and temporal relationships.
- Interpretability
Turning opaque predictions into faithful, human-readable explanations.
- Systems
Testing learning methods against the constraints of operational environments.