Research & exploration
Three proposed directions for learning through experiments. These are starting points for collaboration, not completed projects or established research programs.
Proposed direction / 01
Explore local models, Thai-language tools, and the constraints of running useful AI on available hardware.
A question to test: How reliably can a local model summarize a small set of Thai-language documents?
A possible first step: Define a narrow task, choose appropriate sample documents, and compare factual accuracy, response time, and resource use.
Understand local AI ↗Proposed direction / 02
Investigate how a promising demonstration becomes a tool that someone can use for a specific, repeatable task.
A question to test: Does a prototype help someone complete a task more accurately or with less effort?
A possible first step: Map a simple workflow, build the smallest useful version, and compare it with the existing way of doing the task.
Learn how to evaluate it ↗Proposed direction / 03
Explore perception, robotics, and the challenge of connecting software decisions to the physical world.
A question to test: How does a perception system perform when lighting, position, or background changes?
A possible first step: Begin with recorded examples or a simulation, measure the errors, and identify the conditions that make the system less reliable.
Find out about getting involved ↗A shared approach
Each direction can begin with a clear task, a simple baseline, and a record of what happened. Useful findings include the limits of an approach and the reasons a first idea did not work.
Teams, resources, and participation arrangements for these directions are still to be determined.
Read about the initiative ↗