CM1K neuromorphic chip with 1024 neurons
The CM1K is a neuromorphic chip opening new frontiers for smart sensors and cognitive computing applications. It can solve pattern recognition problems from text and data analytics, vision, audition, and multi-sensory fusion with orders of magnitude less energy and complexity than modern microprocessors.
The CM1K chip features 1024 interconnected neurons working in parallel and capable of learning and recognizing patterns in a few microseconds. The neurons behave collectively as a K-Nearest Neighbor classifier or a Radial Basis Function and are trainable. They are especially suitable to cope with ill-defined and fuzzy data, high variability of context and even novelty detection. Last, but not least, multiple CM1K chips can be daisy-chained to scale a network from thousands to millions of neurons with the same simplicity of operation as a single chip.
The CM1K chip is a chain of 1024 identical neurons operating in parallel, but also interconnected together to make global
decisions. A neuron is a memory with some associated logic to compare an incoming pattern with the reference pattern stored in its memory and react (i.e. fire) according to its similarity range. A neuron also has a couple of attribute registers such as a context and category value. Once a pattern is broadcasted, the neurons communicate briefly with one another (for 16 clock cycles) to determine which one holds the closest match in its memory. The “Winner-Takes-All” neuron de-activates itself when its category is read, thus leaving the lead to the next “Winner-takes-All”, if applicable, and so on. A single CM1K matching a pattern of 256 bytes against 1024 models delivers the equivalent of 192 GiGaOps per second.
Learning is initiated by simply broadcasting a category after an input pattern. If it represents novelty, the next neuron available in the chain automatically stores the pattern and its category. If some firing neurons recognize the pattern but with a category other than the category to learn, they auto-correct their influence fields. This intrinsic inhibitory and excitatory behavior makes the CM1K chip a unique component for cognitive computing applications.
Furthermore, CM1K integrates a built-in recognition engine which can receive vector data directly through a digital input bus, broadcast it to the neurons and return the best-fit category 3 microseconds later. In the case of a video input signal, NeuroMem can optionally extract a 1D vector from 2D video data.
|Neuron memory size||256 bytes|
|Category register||15 bits|
|Distance register||16 bits|
|Context register||7 bits|
|Recognition status||Identified, Uncertain or Unknown|
|Classifiers||Radial Basis Function (RBF), K-Nearest Neighbor (KNN)|
|Distance Norms||L1 (Manhattan), Lsup|
|Logic module (optional use)||High speed recognition stage (for 1D & 2D vector data)|
|Communication module (optional use)||I2C serial controller at 100 or 400 Kbit per second|
BrainCard is a trainable pattern recognition board for IoT and smart appliances featuring a NeuroMem CM1K chip with 1024 neurons interfaced via SPI communication to Arduino and Raspberry Pi processor boards and ready to learn and recognize patterns.
NeuroStack has unique architecture featuring four NeuroMem CM1K chips and a Field Programmable Gate Array to perform high speed pattern learning and recognition.