Fujitsu Deep Tensor® for Explainable AI


How Fujitsu Technology elicits New Insights from Graph Data

banner - sponsor (2)

Machines are beneficial to the extent that their actions can be expected to achieve our objectives.” – Russel, S. Human compatible: Artificial intelligence and the problem of control.

What should be the basis for AI trust?Explanation in AI systems is considered to be critical across all areas where machine learning is used.
In the case of a recommended medical treatment or a rejected application for a mortgage loan,
we may be entitled to clear explanations of how we arrived at a particular result.

So emerges the urgent need to make results and machine decisions transparent, understandable and explainable.

Our Partner, Fujitsu Laboratories developed Deep Tensor (DT), that provides users with information on the system’s prediction with an effective explanation for the AI system’s behavior. 

DT combines the proprietary AI technology Deep Tensor, a deep neural network that is especially suited to datasets with meaningful graph-like properties with Knowledge Graph (Neo4j).

Deep Tensor converts graph-structured data to a form of mathematical expression called a tensor and performs deep learning to achieve the highly accurate findings.

The technology is also able to run a reverse search of the deep-learning output to identify factors that had a significant impact on the results.
The knowledge graph consists of a huge amount of graph data that includes all sorts of knowledge. 

DT identifies the factors (partial graphs) that had a significant influence on an inference and coordinates these with partial graphs from a knowledge graph, building a series of pieces of information in the form of connections in the knowledge graph as the basis for the findings.
Knowledge Graph is used to present the basis for results obtained by AI.

Connecting inferences derived by Deep Tensor to Knowledge Graph, DT enables to understand the reasons behind AI-generated findings and to make them explainable.

This scheme solves the problem of black-box machine learning while maintaining the high inference accuracy achieved by Deep Tensor. Specifically, it dramatically improves Deep Learning efficiency and tells why it produced specific inferences.

DT is already in use in several mission-critical fields:

health management (health changes are detected from employee activity data for workstyle transformation)
risk management for investment decisions and loan products, detection of money laundering in finance
cybersecurity and network intrusion detection
genomic medicine and biochemistry for learning the structure and activities of chemical compounds

The follow figure shows an example in applying DT to genomic medicine, DT infers which gene mutation of the patient is associated with disease.

Fujitsu Partnership is part of our path to new growth.
We expand and enhance our offer for our Neo4j customers, thanks to a graph-based approach to artificial intelligence.

Neo4j graph database provides human-understandable representation of multi- dimensional data.
Accessible Deep Tensor integrates Deep Tensor with Neo4j, providing an end-to-end platform for Machine Learning.

Watch this video to discover how Deep Tensor works.

Resource links:

Explainable AI Through Combination of Deep Tensor and Knowledge Graph

Fujitsu Fuses Deep Tensor with Knowledge Graph to Explain Reason and Basis Behind AI-Generated Findings

Why AI Got the Answer -Explainable AI Showing Bases

Please read the article available on Neo4j blog ⏩ for information about Larus-Fujitsu and Neo4j Partnership.

For the additional information and to request a demo please contact us at