NeuroMem CHIPS

Non linear classifier

Identify, Classify, Detect novelty, cope with non linearity and ill defined problems, report cases of unknown and cases of uncertainty.

Model Generator

Real-time learning, high speed clustering, intrinsic deduplication, tracability of the knowledge built by the neurons

High performance

Fixed latency regardless of the number of neurons, order of micro-seconds per pattern, low power consumption (0.5 Watts), low clock frequency, small footprint

NeuroMem-CM1K-296x300

CM1K

NeuroMem network with 1024 neurons with 256 bytes of memory per neuron. Access through a proprietary parallel bus or an I2C bus. Optional use of a video recognition stage. 100-pins TQFP package 8×8 mm.

NM500

NeuroMem network with 576 neurons with 256 bytes of memory per neuron. Access through a proprietary parallel bus. Wafer-scaled package.

More to come …

Applications

DS_NM500-5
Image recognition

– Person identification,
– Part inspection,
– Microscopy,
– Medical imaging,
– Satellite and
– hyperspectral, etc

DS_NM500-6
Video

– Target tracking,
– Motion detection,
– Gesture recognition,
– Behavior analysis,
– Stereoscopy

DS_NM500-7
Signal recognition

– Speech, Voice,
– Audio,
– Biosensors
– Sonar, Radar…

DS_NM500-8
Data mining

– Fingerprint and iris,
– DNA, genomics,
– Meteorology, Physics,
– Environmental Sciences
– ext

DS_NM500-9
Text and Packets

– Cyber security
– Sentiment analytics,
– Social networking,
– Text analytics…