The CHALLENGE: Be yourself the classifier!
I invite you to BECOME a Fall/ADL Classifier… and to compare yourself (“face-off”) against some Artificial Intelligence (AI) models in a classification task. By playing and experimenting with The Classifier Mastery Game (CMG) – a.k.a “Confusion Matrix Game” – you will learn about AI-based classifiers, their performance and related measurement instruments, such as the Confusion Matrix (CM).
What is it like to be a CLASSIFIER?
AI-based (or ML-machine learning based) classifiers are mathematical/computational models that are “trained” in order to “learn” patterns that are embodied in data (often in large datasets), so that when they are presented with new data/samples, they can classify them based on their previous learning (“experience”). To train an AI/ML model, different techniques can be used. In a type of machine learning – known as Supervised Learning – the classes of the samples used during training are known – or as usually said, the samples are “labelled”. Thus, a first step in the creation of an AI-classifier, is actually to “train” it by “showing” it some (labelled) samples.
To ensure “fair-playing” you get a chance to train yourself, which is your first step to “become” a Fall/ADL classifier!
How to PLAY:
- TRAIN YOURSELF. Observe the following “labelled” samples of each class: FALLs and ADLs.
- TEST YOURSELF. When you first go-to the “PLAY the game” tab (or by pressing the buttom “Start new game”), eight samples of generative artworks will be displayed. Each one is either a FALL or an ADL (Activities of Daily Living: seating, walking, etc.). As a “classifier”, your job is to “guess” (classify!) each one as FALL or ADL.
- CHECK YOUR RESULTS. When you have classified all artworks, click the buttom “Show my Metrics” to see your performnace results. Then click/go-to the “Confusion Matrix & Stats” tab, to learn more…
- LEARN MORE. Read the notes & questions in the “Confusion Matrix & Stats” tab; and try/experiment again…
- Last but not least, ENJOY & HAVE FUN!
Note: The generative-data-art pieces used in this challenge/game, have been created using several algorithms with sensor data captured from a FALLs and Activities of Daily Living (ADL). Learn more about this project in the About page.
BE yourself the classifier!
Play one game to see your score & statistics.
Steps to BECOME a GREAT classifier:
- Get some (more) training in the CHALLENGE tab
- Classify each of the artworks above as a FALL or an ADL (Activity of Daily Living). DO YOUR BEST!
- Click on “Show My Metrics” to see your results… …and/or go to the “Confusion Matrix” tab to learn more.
- Come back to play & experiment again!
What is a Confusion Matrix?
A Confusion Matrix (CM) is a very useful measurement instrument used to understand and communicate the performance of a ML (Machine Learning) – based classifier. The CM below, displays the results of the challenge:
|ACTUAL VALUES (classes in the sample)|
|Positive class (Falls)
|Negative class (ADLs)
|YOUR predicted values||Predicted Falls
|True Positive (TP)
|False Positive (FP)
|False Negative (FN)
|True Negative (TN)
Questions and call-to-actions:
- Do you think an AI (Artificial Intelligence) could do better than you…?
- “Compare” your results against some AI-based models by clicking the “AI Benchmark” button [coming soon].
- What could you do to improve your performance…? How could an AI’s performance be improved…?
- HOW CAN FALLS BE PREVENTED? …and/or PREDICTED/DETECTED by “AI/ML-based classifiers“?
- Play again and try to achieve better results!