Apheris Platform for federated & privacy preserving data science

Securely analyze data of multiple parties while keeping proprietary information private

Schedule a demo
Platform overview

Secure, flexible, private

Apheris enables the complete range of data science operations on distributed data while preserving privacy

An easy to use end-to-end solution

We carry the load of technology orchestration, compliance and security. Our contractual agreements help data partners move quickly while complying with all regulatory requirements.

Data science on not directly accessible data

Apheris enables your organization to carry out the complete range of data science operations on distributed, not directly accessible data and supports any data of any format.

Holistic approach for privacy & security

Your Data Scientists are guided via a clear and governed process with privacy assessments and security measures even for the most complex data science pipelines.

Our CEO Robin explains how Apheris helps you to unlock the full potential of your data

apheris explanation video

How our Platform works




1
Data Owners connect data
Data Owners connect data

Data Owners connect data

  • Connect any type of data to the Apheris Platform
  • Data always stays in full control of the Data Owners
  • Privacy technologies and automated policy enforcements guarantee full data protection
2
Data Scientist develops model
Data Scientist develops model

Data Scientist develops model

  • Data Scientist requests to see a virtual dataset
  • Apheris creates a virtual dataset that resembles the Data Owners’ data but does not contain any sensitive information
  • The Data Scientist develops an analytics request (often an AI Model) on the virtual dataset
3
Privacy approval
Privacy approval

Privacy approval

  • The Data Scientist's model is sent to the Apheris Privacy Guard
  • Apheris performs a privacy assessment and adds privacy measures prior to approval for computation
4
Federated and private training
Federated and private training

Federated and private training

  • The model is trained in a federated setup on multiple datasets of different Data Owners
  • All results are iteratively aggregated in one global model
5
Model release
Model release

Model release

  • The trained model gets released to the Data Scientist and / or Data Owners (depending on their contractual agreement)
  • Privacy of the data is fully preserved
6
Extract value & Get insights
Extract value & get insights

Extract value & get insights

  • Receive entirely new insights to benefit your business
  • Integrate new models into your current workflow and make your data science projects successful
  • Monetize your data and build data partnerships
  • Improve your R&D through collaborative data science projects
1
Data Owners connect data
Data Owners connect data
2
Data Scientist develops model
Data Scientist develops model
3
Privacy approval
Privacy approval
4
Federated and private training
5
Model release
Model release
6
Extract value & get insights
Extract value & get insights

Get started

Enter your email to schedule a demo

Submit
Logo of Apheris and Gaia-X

Apheris partners with Gaia-X to shape the future of data ecosystems

Compliance, federation, data sovereignty, and trust are integral parts of everything we do at Apheris. We are proud to announce our partnership with Gaia-X.

The Apheris Privacy Guard creates data science workflows with built-in privacy protection

Apheris enables the secure analysis of data across organizations while keeping proprietary information private. Computations are executed locally – data never leaves the local environment and data privacy is preserved throughout the entire process. We use cutting-edge technologies to provide mathematical guarantees for privacy preservation.

Secure multiparty computation

Compute a joint function on multiple private inputs, where no party learns anything extra about other parties’ inputs.

Differential privacy

Introduce random noise into results of queries on underlying confidential data, so that observers cannot reconstruct the original data.

Privacy preserving record linkage

Match data records that belong to the same entity (e.g. person) without revealing the identity of the entity.

Homomorphic encryption

Encrypt data such that computation is possible on the ciphertext – the decrypted result matches computation on plaintext.

Platform architecture

enterprise grade security
Compliance

Apheris services are GDPR ready and feature capabilities that enable our customers, and their data collaboration partners to comply with GDPR and other laws and regulations.

Enterprise scalability

The Apheris Platform is architected and built for big data processing and supports cloud, multi-cloud and hybrid environments.

Enterprise-grade security

With industry-wide highest security standards, the Apheris Platform has been proven to protect enterprise grade-datasets at the world’s largest organizations.

Platform Use Cases

We see four main data collaboration setups which can unlock the full potential of the data

1. Collaborative data ecosystem – everyone contributes and consumes

Our awesome features
2. Matching of data providers and consumers – AI model provider consumes third party data

Our awesome features
3. Pre-collaboration assessment – data and model providers test complementary assets while preserving IP

Our awesome features
4. Trained models off the shelf – securely commercialize data & data products

Our awesome features