If you have a proposal idea but cannot create it yourself, submit a “Request for Proposal” idea in this thread for someone else in the OceanDAO community.
We encourage all ideas and contributors!
If you have a proposal idea but cannot create it yourself, submit a “Request for Proposal” idea in this thread for someone else in the OceanDAO community.
We encourage all ideas and contributors!
We’d love a platform where users can submit ideas on each dataset posted and how to improve its adoption or use-case - naturally with an incentive mechanism behind it in the form of data-tokens.
Please reach out to us to further build on this idea!
It would be great if there would be a rating agency that rates datasets on their usability for machine learning purposes.
My idea to do this would be to determine what the state of the art models on a certain dataset are. Then take the state of the art model and do transfer learning on the dataset on the Ocean marketplace via compute-to-data or if it is available for download then locally.
After that the newly trained model is evaluated on the test dataset of the original paper.
This gives an indication what this dataset adds for this model. This is a question that many people who want to use the dataset will have and is valuable as it takes time and compute to create. So selling this information directly without having to go through the steps will be valuable.
This could also be an add on to other data insight services.
As I don’t have time to set this up I am offering this idea but I would be happy to discuss more details if someone wants to pick this up.
Trustless Data Audit Service
Problem
Data Providers don’t have a trustless way to authenticate their datasets. Thus data providers have to rely on the traditional social channels (twitter) to promote their datasets for consumption.
Consumer have to rely on information in description about the datasets and its properties (and data providers’ word). Even if there is a sample attached, its not certain that the underlying dataset is according to the description. Though such dataset fraud will get caught soon after the first consume, but first consumer loses their funds by consuming fraudulent datasets.
Solution
Create an open-source audit service/dapp that allows data providers to self-authenticate their datasets trustlessly (and then they can attach these results to their dataset description) and consumers can trust these results (since its open source service and code can be verified). Such a dapp would improve trust factors between provider and consumers and will result in better consumption rates for high-quality datasets.
Ocean Token Dashboard
A dashboard to track Ocean tokens in various platforms - CEXs, DEXs, OPF wallets etc
and maybe produce some analytics on the movement of tokens between various platforms. This would provide transparency and trackability of OCEAN tokens to OPF and community.