← RCAI4S · Research Centre on AI for Science

Ask a question

You know which questions are worth asking. Tell us one.

The quickest way is to tell us three things:

Problem
What you have, and what you would like to predict or find from it.
Data
A link to your data: a public link, or a PolyU OneDrive folder shared with jessie-zh.li@polyu.edu.hk. As a rough guide for a first look: about 1,000 samples with known answers, under 1 GB. If yours is larger, share a representative part; we’ll plan the full work with you together.
Metric
How you would judge a good answer.
Example: the message that started Jiong Zhao’s case
PROBLEMour goal is to train a machine learning model, predict from /input to /target

DATAhere is the data: …/simulation/train_data1   ← in your message, put the link to your data here
in the /input folder, it is the data with some noise
in the /target folder, it is the data with another type of noise
in the /point_map folder, it is the ground truth
there are 1500 samples

METRICuse the following metrics:
PSNR: Peak signal-to-noise ratio
SSIM: Structural similarity index measure

Typed on 22 September 2026. See what came of it.

Can’t describe your problem in the format above? Just tell us about it in your own words, and we’ll find a time to talk it through together.

I am

How we’ll get back to you. We’d love to hear from everyone. If you’re a PolyU faculty member, we’ll get back to you within a week. Students and postdocs, we’d be glad to help too; please include your supervisor so we can work with your group together. If you’re from outside PolyU, we read every message and will reach out wherever we can help.