Intelligence Lab
Advanced Machine Learning Architectures • Predictive Analytics • Computer Vision Systems
The Intelligence Lab is where we put our trained models in front of anyone who wants to try them. Each one is a model we built and evaluated ourselves, served from our own inference API, and exposed here with the raw inputs it expects — no sign-up, no sample data hidden behind a form.
What you can run here
- Sonar Mine Detector — a classifier that reads 60 sonar frequency returns and predicts whether the object is rock or a mine.
- Diabetes Risk Predictor — takes eight patient metrics — glucose, blood pressure, BMI, age and others — and estimates risk of onset.
- Fake News Detector — a natural-language model that judges whether a headline reads as authentic or fabricated.
- Wine Quality Analyst — a regression model that scores quality from chemical properties.
- Deep Learning Vision — MobileNetV2 exported to ONNX, classifying types of clothing from an image.
How it works
Each model is trained offline, evaluated against a held-out set, then exported and served behind a FastAPI endpoint. When you press run, your inputs are sent to that endpoint and the prediction comes back as JSON, which the page renders.
The vision model is the exception: it runs entirely in your browser through ONNX Runtime Web, so the image you pick never leaves your device.
What it is not
These are research models, published so students can see what a trained classifier actually does rather than read about one. They are not calibrated for professional use, and the medical model in particular is a teaching example — it is not a diagnostic tool and must not be used as one.
Neural Workspace
Neural.Awaiting_Input