AleoMed

Inspiration

zkML is a fantastic case for applying ZKPs. We can run ML models on sensitive data such as DNA in a privacy-preserving manner.

What it does

Researchers can create ML data collection projects on our platform by uploading their trained ML models. We automatically generate and deploy Leo code for them.

Users can take part in ML research projects by providing their DNA fragment in FASTA file. The DNA data never leaves user's machine, ML model is run inside Aleo transition. The result of execution is then transferred to the researcher. The research participant gets rewarded with tokens.

How we built

We built frontend:

• displays all ongoing research projects
• allows a researcher to create a new project
• allows a user to participate in a DNA research project

We built a research project Aleo program template:

• it has a general logic around research
• a placeholder where transpiled ML code is injected

We built backend:

• stores and serves ongoing research projects
• generates LEO code using a serialized trained ML model combining transpiled and custom code

Challenges we ran into

Sometimes there were issues with blockchain stability

Accomplishments that we're proud of

• Implemented a great socially valuable use case of ZKP

What we learned

How to work with ML on Aleo

What's next for AleoMed

A grant would be cool :)

Built With

• aleo
• machine-learning
• python