Project Name
Homomorphic Encryption for Ocean Market
Project Category
Build & Integrate
Proposal Earmark
2nd/3rd Grant
Proposal Description
This is the second proposal of the project Homomorphic Encryption for Ocean Market.
Please inspect
to get more insights about the current progress.
The first proposal outlined (very) ambitious deliverables. We are still very optimistic to be able to deliver them, however the necessary effort was- and probably will be much higher than expected. We would like to introduce more fine-grained sub-targets/deliverables/milestones and show the increments transpantly to everyone.
Therefor in this proposal we would like to focus on a showcase which can be understood and executed easily by everyone. Our target is to truely lower the entry barrier for Homomorphic Encryption and give the possibility to every developer to test- and consider it for its own solutions without specific HE-knowledge.
The dataset which will be used for the demonstration is the well known MNIST-dataset. A plaintext ONNX-model will be used.
The three steps that will be executed are:
- Homomorphic Encryption
- Homomorphic Inference
- Homomorphic Decryption
The implementation will be based on at least on one (potentially both) of the following repositories which are still private but will be made public soon:
Grant Deliverables
- SMiLe research project repositories public available
- Open source demonstrator
- Step by step documentation
Project Description
Homomorphic Encryption based solution to preserve intellectual property and/or privacy of monetized data for participant X and AT THE SAME TIME preserve intellectual property and/or privacy of the monetized algorithm/model applied on that data for participant Y for applications in healthcare and finance.
Final Product
not set
Value Add Criteria
The Ocean Market currently supports
- monetization of data OR
- monetization of algorithms.
Data can be monetized by publishing a dataset. The seller loses intellectual property and/or privacy to the buyer by sharing his data.
Algorithms can be monetized with “Compute to Data”. The seller loses intellectual property and/or privacy by sharing his algorithm/model.
In both cases the buyer is protected.
Protecting BOTH parties, seller and buyer, AT THE SAME TIME (more precise for the same interaction) is currently not possible in a convenient manner, not even for the simplest algorithms/models.
This can lead to a blocking-scenario: If at least one party considers the potential damage bigger than the upside of interaction the interaction will not take place.
Homomorphic Encryption for Ocean Market solves this problem for a selected use-case to demonstrate it to a broader audience. It further creates an open-source-blueprint for applications in healthcare and finance.
Funding Requested
10000
Minimum Funding Requested
10000
Wallet Address
0x69E49561bbdB44CaF87ae6505f7330bAf8dD32BF