[FEEDBACK]: Leaking pipeline is missing

[robvanderveer]

Type

General Feedback

What would you like to report?

The risk of leaking training data or other confidentiality issues of the AI pipeline (code, model parameters) are missing.

Code of Conduct

• I agree to follow this project's Code of Conduct

[aryanxk02]

I can come up with some content associated with this topic, but where in the project do we integrate this exactly?
cc: @shsingh @robvanderveer

[shsingh]

hi @aryanxk02 if you want to take a stab at creating the content first let's just work on that for now. we can work out where it goes as part of this discussion. /cc @sagarbhure @mik0w @robvanderveer

[aryanxk02]

Hi @shsingh, this is a brief outline for the content associated with the above topic. Feel free the make changes wherever you think necessary :)

Introduction

• Significance of data and model confidentiality in AI.

Understanding AI pipeline

• Various stages of an AI pipeline, from data collection to model deployment.
• Importance of sensitivity of the data and model parameters at every stage.

Critical Issue 1: Data Leakage

• What is Data Leakage?
• Possibility of data leakage occurrence and consequences.
• Prevention: Best practices.
• Real world examples of data leakage incidents.

Critical Issue 2: Model Parameters

• Confidentiality of model parameters.
• Best practices to follow to protect the parameters.
• Importance of code review for model related code.

Conclusion and References

• Information in a nutshell and cite references.

cc: @robvanderveer @sagarbhure

[robvanderveer]

This is a good start. Have a look at the AI guide for further inspiration.
You need to add 'code, documentation and configuration' as well, as that is also intellectual property right.
If you add this as a new entry in the top 10, you can focus 'model stealing' on behavioral stealing and add a remark there that another way to steal a model is to steel the parameters and then refer to this pipeline entry.

[aryanxk02]

Can you please share the link to the AI guide so I can refer to the same.

[robvanderveer]

https://owasp.org/www-project-ai-security-and-privacy-guide/
also listen to: https://mlsecops.com/podcast/a-holistic-approach-to-understanding-the-ai-lifecycle-and-securing-ml-systems-protecting-ai-through-people-processes-technology
watch: https://www.brighttalk.com/webcast/19697/586526
and look at especially the annexes in : https://www.enisa.europa.eu/publications/securing-machine-learning-algorithms