Keeping AI accurate

Issue 9 2020 News

Fujitsu Laboratories announced the development of a technology to make AI models more robust against deception attacks. The technology protects against attempts to use forged attack data to trick AI models into making a deliberate misjudgement when AI is used for sequential data consisting of multiple elements.

With the use of AI technologies progressing in various fields in recent years, the risk of attacks that intentionally interfere with AI's ability to make correct judgments represents a source of growing concern. Many suitable conventional security resistance enhancement technologies exist for media data like images and sound. Their application to sequential data such as communication logs and service usage history remains insufficient, however, because of the challenges posed by preparing simulated attack data and the loss of accuracy.

To overcome these challenges, Fujitsu has developed a robustness enhancement technology for AI models applicable to sequential data. This technology automatically generates a large amount of data simulating an attack and combines it with the original training data set to improve resistance to potential deception attacks while maintaining the accuracy of judgment.

By applying this technology to an AI model developed by Fujitsu to judge the necessity of countermeasures against cyberattacks, it was confirmed that misjudgement of about 88% could be prevented. Find out more at

Share this article:
Share via emailShare via LinkedInPrint this page

Further reading:

From the editor's desk: Take identity seriously
Issue 8 2020, Technews Publishing , News
The list of companies subject to data breaches continues to grow, both in Africa and beyond. The dangerous part is that in Africa, including South Africa (at the moment), the companies do not have to ...

Free-flow smart weapons detection system
Issue 8 2020, XPro Security Solutions , News, Security Services & Risk Management, Products
Detecting people carrying weapons and preventing them from entering your venue is now possible, without sacrificing the visitor experience.

Net2 integrates with additional thermal scanners
Issue 8 2020, Paxton , News, Access Control & Identity Management
Paxton has updated Net2 to include six new thermal temperature scanning integrations to assist in managing the pandemic.

App Challenge winners
Issue 8 2020 , News, CCTV, Surveillance & Remote Monitoring
Security & Safety Things announced the winners of App Challenge for innovation in AI-enabled smart camera applications.

ONVIF introduces Profile M Release Candidate
Issue 7 2020 , News
ONVIF has introduced its Profile M Release Candidate for standardising the communication of metadata and event handling of analytics for smart applications.

First Distribution to distribute Video Storage Solutions
Issue 8 2020 , News, CCTV, Surveillance & Remote Monitoring, IT infrastructure
First Distribution has signed an agreement to distribute the entire Video Storage Solutions (VSS) product line of video surveillance appliances.

Ring Africa launches new cameras in SA
Issue 8 2020 , News, CCTV, Surveillance & Remote Monitoring, Smart Home Automation
Ring Africa has officially launched a range of indoor and outdoor home security cameras in South Africa.

VisionPass wins SIA award
Issue 8 2020, IDEMIA , News, Access Control & Identity Management
IDEMIA announced that VisionPass, its newly launched advanced facial recognition device, won the SIA New Product Showcase Award.

Liquid Telecom unveils Cybersecurity unit
Issue 8 2020 , News, Cyber Security
Liquid Telecom has unveiled its Cybersecurity unit, which offers end-to-end managed services for digital security solutions.

T-Systems and Fortinet launch cyber academy
Issue 7 2020 , News
T-Systems South Africa will be introducing cybersecurity education programmes to create a specialised Cyber Security Academy incorporated into the organisation’s existing ICT Academy.