Submission instructions

The final submission must be uploaded at the following link:

A quick testing page and a final submission page are provided in the above link to support the self-evaluation on the submission website and the final submission.

Detailed instruction is as follows:

Fig. 1 – Sign-in page. Click on ‘CREATE ACCOUNT’ to sign-up an account for your team.

Fig. 2 – Sign-up page. Click on ‘SIGNIN’ if you have already registered an account for this data challenge.

Fig. 3 – Home page to perform self-evaluation of your model. Drag and drop your zip file to the top-right panel to upload. A sample data can be retrieved by clicking on “Sample Data” to see the format (i.e., a list of Pandas DataFrame). An example submission can be downloaded by clicking on “Example.zip”.

Submission format: Per each submission, please create a folder, whose name should differ from your previous submissions, put your ‘requirements.txt’ and ‘Model.py’ in it, then zip it without changing its folder name. In ‘requirements.txt’, please specify the Python modules you are using in this submission. The platform will automatically install these modules via “pip install”. Only the “evaluate” function in “Model.py” script will be used to evaluate your submission. The format of function is:

def evaluate(data):

#Your codes here

return detected

Here, detected should be a list of lists, where a sub-list contains the binary indicators (i.e., 0 for an in-control measurement point, 1 for an out-of-control measurement point). Number of binary indicators in each sub-list equals to the number of measurement points in the corresponding layer. And the number of sub-lists in detected equals to the number of layers. You are encouraged to use open source codes.

Fig. 4 – Results will be visualized on the home page.

Fig. 5 – Results from historical trials can be reviewed in “ACTION LOGS”. Errors will also be displayed here.

Fig. 6 – Final submission can be made on the “FINAL SUBMISSION” page by separately uploading README,
Presentation, and the zip file that contains the scripts.

The Presentation file should discuss the developed methodology as well as all relevant results (max 10 page – no template).