|
|
|
|
Safety_Critical_Machine_Learning |
|
New patents filed in
2025/2026
on applications and Machine Learning platforms
using Neuromem® technology
for an Accountable, Measurable and Explainable AI
AMEN™ is not just an acronym, an advertising slogan, or anything
similar. AMEN™ is the goal of our research, the one we never lose sight of when
trying to solve real-life problems with AI. We work primarily in critical
sectors such as defense, aerospace, and biomedical, and we know how important
it is to be able to map human responsibility in an AI-based decision-making
system: measurability and explainability are essential. Our research, conducted
with these criteria, generates patents optimized for startups and companies
that want to implement them. Although the claims are
general, the patents are extremely detailed and
can be used as a design specification.
|
Soldier Advanced Vision Extender (SAVE-360™) SAVE-360™ is a non-invasive and
fully automated 360-degree vision system that can operate with automatic
vision technologies with very low energy consumption and long operational
autonomy. The system simultaneously provides only the essential data for
spatial understanding of dangerous situations. This patent is tailored for
Neuromem® technology. Patent Pending |
|
UNobtrusive Observer (UNO™) UNO™ is a vision system for monitoring
risk situations at workstations that is fully compliant with privacy
regulations and proof against any attempt to access temporary internal
images. This patent is tailored for Neuromem® technology. Patent Pending |
|
Built-In In Air Reverberation Test
(BI2ART™) BI2ART™ is a system for
monitoring the correct functioning of a USB ultrasound probe in the field
(hospitals, clinics, analysis laboratories) based on IAR (In Air
Reverberation) tests and fully integrated into the probe itself. It is the
first BIT (Built In Test) system for ultrasound probes. This patent is
tailored for Neuromem® technology. Patent Pending |
|
MemorY
To High Operational Speed (MYTHOS™) MYTHOS™ is a technology developed
to bridge the performance gap between shallow and deep networks. This
technology enables real-time operation with millions and even billions of
neural prototypes in an RBF neural classifier with Restricted Coulomb Energy
(RCE) learning. The MYTHOS™ patent is based on Neuromem® technology and
relates to the development of a machine learning hardware platform based on
multiple Neuromem®-based boards (e.g., M2 boards). (ARCHER™
technology) Patent Pending |
|
HYper-Dimensional RecOrd Generalization
ENgine (HYDROGEN™) HYDROGEN™ is an extension
technology of MYTHOS™ that enables the machine learning platform to scale in
input size by concentrating multiple asynchronous/plesiochronous data sources
on a single neural classifier with real-time learning capabilities. (ARCHER™ technology) Patent Pending |
|
Field Agnostic Tag Categorizer
(FatCat™) is a HW system for category scalability. Fuzzy Logic Average Table General REgression
ENgine (FlatGreen™) is a HW system for general regression with neuromorphic
RCE classifiers. (ARCHER™
technology) Patent Pending |
|
|
|
General Synaptics Aerospace_and_Defence_Machine_Learning_Company VAT NUMBER:_IT02670700992 REA NUMBER: GE-503104 Email:_luca.marchese@synaptics.org |
|
|
|
Copyright© 2026 General Synaptics |
