Platform
Three ways to unlock data, without sharing it
Acuratio supports three complementary modes of privacy-preserving computation. Pick the one that matches how your data is partitioned.
Horizontal Learning
Train a ML model across distributed examples of the same data with Split Learning or Federated Averaging. Example: data localization, fraud, risk, forecasting.
Learn more →
Vertical Learning
Combine two or more data sources with a common set of users but different features. Built on Split Learning and Private Set Intersection. Example: pricing, retention.
Learn more →
Private Analytics
Keep individual information private while computing aggregate statistics or segmenting audiences. Powered by Private Join & Compute. Example: conversion, cohorts.
Learn more →
Which mode fits your setup?
A quick comparison of when to reach for each.
| Horizontal | Vertical | Private Analytics | |
|---|---|---|---|
| Shares features across parties | — | — | |
| Shares users across parties | — | ||
| Trains a Deep Learning model | — | ||
| Computes aggregate statistics | — | — | |
| Built on Split Learning | — | ||
| Built on Private Join & Compute (MPC) | — | — |