Closed innopreneur closed 3 years ago
Issue Status: 1. Open 2. Started 3. Submitted 4. Done
This issue now has a funding of 33332.0 OCEAN attached to it as part of the oceanprotocol fund.
Issue Status: 1. Open 2. Started 3. Submitted 4. Done
Work has been started.
These users each claimed they can complete the work by 2 days, 23 hours from now. Please review their action plans below:
1) nazariyv has started work.
I have managed to set the environment up successfully. Next step is to play around with the API for data submissions and compute jobs. Once I am able to interact with that, I will create the pod with the outlined requirements. Although, I need to confirm exactly what some of those mean. For example, the require is that the output is not encrypted, and that it is not not a 1-1 copy :) Which means that it is a 1-1 copy :)
Project logo image attribution: Photo by Ivan Bandura on Unsplash 2) hellomalaysia has started work.
Streaming online 3) erieza09za08 has started work.
استشمار ومساهمة والحفاظ عليها وابلاغنا
Learn more on the Gitcoin Issue Details page.
Issue Status: 1. Open 2. Started 3. Submitted 4. Done
Work for 33332.0 OCEAN (2731.66 USD @ $0.08/OCEAN) has been submitted by:
@innoprenuer please take a look at the submitted work:
Issue Status: 1. Open 2. Started 3. Submitted 4. Done
The funding of 33332.0 OCEAN (2284.34 USD @ $0.07/OCEAN) attached to this issue has been approved & issued to @nazariyv.
DLP - Data Leakage Protection for Compute-to-Data
Prize Bounty - 33,333 OCEAN
Questions & Reviews
You can reach out to me for any queries via email to ecosystem@oceanprotocol.com or post your questions in our discord channel - Ocean Protocol's Discord
Challenge Description
Concept:
We recently released our much awaited feature - Compute to Data Ocean compute-to-data operates on Kubernates cluster. We have 2 main components for this -
More details on compute-to-data -
We currently employ algorithm whitelisting mechanism to safe guard data from hacks and leakage. This however requires compute provider to verify algorithm externally. We want you to build a feature within our compute infrastructure that automates data leakage protection mechanism. If properly implemented, this will replace our whitelisting feature and reduces efforts from compute provider side.
Ideally to accomplish this you will create a kubernates pod whose function is to check the output of the compute service and verify that output is not -
Encrypted
Not 1-1 copy
Not more than 10% of the dataset size
Does (not) contain keywords
calculate correlation between output and each feature (DP) - not more than a certain threshold - then throw an error
PIIs according to GDPR https://gdpr.eu/eu-gdpr-personal-data/
Have a dataset as sample (German credit score data) that you can use to prove success : https://www.openml.org/d/31
Submission Requirements
How to Submit?
Resources for Developers