The Science Data Exchange

ComputeGenomics

AI needs frontier science, and science needs frontier AI. We're bridging the two by building the exchange and infrastructure layer that brings scientific data to scale, starting with the life-sciences.

Data Producers

  1. 01produce
  2. 02structure
  3. 03permission
  4. 04license

Do you produce scientific data?

GraphonMarkets will let firms that engage in scientific research fund science with science. We believe that now, more than ever, experimental scientific work is inherently commercially valuable. The AI industry is built on the output of human intelligence, and it is mathematically bound to being so. With the mass of readily accessible data largely consumed by this industry, demand for frontier data at the edge of human intelligence is accelerating.

Structure and Store

  • Vet and structure your data.We'll help you store and structure your data. After that, we'll vet the quality of your data and get you ready for commercialization.
  • Write detailed data cards. For each datastream you hook into GraphonMarkets, write a detailed data card. Make buyers aware of metadata, compliance info, and generation methods.

Configure Granular Permissions

  • Enforce access rights. You configure access keys that enforce custom usage rights on your datastreams. Keys enforce granular access limits and expiration dates.
  • Protect your data. Take control over export and compute permissions to keep your data in good hands. Control the visibility and exportability of your data, and set limits on the operations buyers are allowed to apply.

Monetize Access and Usage

  • Monetize access. The initial sale price of a key monetizes the privilege to access your data. Sellers will be able to bid on your keys, which will set the final sale price.
  • Monetize (and monitor) usage. After you have sold a key, collect recurring revenue based on usage. In addition, monitor detailed usage statistics to see how buyers are interacting with your data in real time.

Data Consumers

  1. 01license
  2. 02aggregate
  3. 03configure
  4. 04train

Do you consume scientific data?

GraphonMarkets will fix the data-constraint problem. Frontier data play a more prominent role in your competitive edge than ever before, and we want to help you bring your models to the next stratum of performance by giving you access to scientific datasets that are high-quality, clean, domain specific, and produced by the smartest humans. We know that scientific datasets are fragmented across labs, buried in inconsistent formats, and constrained by unclear access terms, but GraphonMarkets removes these barriers and incentivizes labs to bring their output online.

Source Specialized Data

  • Find differentiated datasets. Our platform incentivizes labs to bring their data online. Access high-quality, high-signal experimental data difficult or impossible to source through public repos.
  • Evaluate before you buy. Examine all aspects of a dataset to make sure it's right for your model. Read highly detailed data cards with deep metadata and compliance information. Download a sample to vet data in a way that fits your needs.

Find Suitable Access Terms

  • Know exactly what you're getting. GraphonMarkets will reduce the murkiness of compliance and consent around data acquisition. See usage rights, read limits, and export/compute restrictions up front.
  • Choose the right access model. Different labs will sell access under different constraints. So, you have the opportunity to purchase access under terms that fit your use case.

Train and Eval Across Labs

  • Own a secure environment. We will provide you an environment to help aggregate vast, complex train, val, and test sets compiled from datastreams across multiple labs.
  • Navigate heterogeneous permissions. Your aggregation environment is backed by a data-intelligence layer that acts on your keys and the associated data cards. Effortlessly combine datastreams with heterogeneous access permissions.

Contact

The Science Data Exchange

arya@graphonmarkets.comGraphonDataExch