AI developers face significant hurdles in training their models on sufficiently large and diverse datasets because of privacy, security, and cost concerns. The Rhino Federated Computing Platform (FCP) addresses all of these concerns.
The Rhino FCP unlocks healthcare data collaborations using edge computing and federated learning, leaving data at rest at each site while allowing data pre-processing, harmonization, model training and validation - all with no data transfer.
Members of the rhino network include dozens of medical centers, universities and industry leaders collaborating on a variety of data modalities such as medical imaging, medical notes, histopathology, genomics, and proteomics.
What is Federated Computing?
Interoperability
Federated Data
Data is mapped to common data models and harmonized
Accessibility
Federated Analytics
Privacy preserving data analysis and queries to facilitate data discovery
Usability
Federated Computing
Edge compute applications including federated learning
With Federated Computing, code is sent to each edge node, executed and results are aggregated for creating powerful AI models while preserving patient privacy.
Why use Rhino versus...
Only use your own data
Downsides
Limited generalizability and potential to translate into impact
Rhino FCP Solution
Privacy-preserving data collaborations across sites with diverse patient data
Alternative 1
Alternative 2
Centralize data from different sources
Downsides
Slow due to data sharing concerns, expensive due to creating replicates of massive data sets and low-quality due to non-standard data processing methods
Rhino FCP Solution
Data remains federated on the edge, eliminating risks and costs of centralization
Build 'DIY' federation with open-source frameworks
Alternative 3
Downsides
Expensive, non-repeatable and non-scalable
Rhino FCP Solution
Start collaborating in days, not months, with an enterprise-hardened solution