There’s nice snapshot of advancing work to develop improved neural network “synapse” technologies posted yesterday on IEEE Spectrum. Lower power, ease of use, manufacturability, and performance are all key parameters in the search for optimum neural network technologies. Several promising approaches were presented at last week’s IEEE International Electron Device Meeting in San Francisco. Currently there is a virtual alphabet soup of possibilities: resistive RAM, flash memory, magnetoresistive RAM (MRAM), electrochemical RAM (ECRAM) and phase change memory (PCM), among the contenders.
As pointed out in the IEEE Spectrum article (Searching for the Perfect Artificial Synapse for AI written by Samuel Moore), “Rather than use the logic and memory of ordinary CPUs to represent these (neural net connections), companies and academic researchers have been working on ways of representing them in arrays of different kinds of nonvolatile memories. That way, key computations can be made without having to move any data. AI systems based on resistive RAM, flash memory, MRAM, and phase change memory are all in the works, but they all have their limitations.”
IBM presented interesting work with ECRAM technology as noted in this excerpt from the article:
“The ECRAM cell looks a bit like a CMOS transistor. A gate sits atop a dielectric layer, which covers a semiconducting channel and two electrodes, the source and drain. However, in the ECRAM, the dielectric is lithium phosphorous oxynitride, a solid-state electrolyte used in experimental thin-film lithium-ion batteries. In an ECRAM, the part that would be the silicon channel in a CMOS transistor is made from tungsten trioxide, which is used in smart windows, among other things.
“To set the level of resistance—the synapse’s “weight” in neural networks terms—you pulse a current across the gate and source electrodes. When this pulse is of one polarity, it drives lithium ions into the tungsten layer, making it more conductive. Reverse the polarity, and the ions flee back into the lithium phosphate, reducing conductance.
“Reading the synapse’s weight just requires setting a voltage across the source and drain electrodes and sensing the resulting current. The separation of the read current path from the write current path is one of the advantages of ECRAM, says Jianshi Tang at IBM T.J. Watson Research Center. Phase change and resistive memories have to both set and sense conductance by running current through the same path. So reading the cell can potentially cause its conductance to drift.”
Moore’s article is worth a quick read. He notes, “There may be no perfect synapse for neuromorphic chips and deep learning devices. But it seems clear from the variety of new, experimental ones revealed at IEDM last week that there will be better ones than we have today.”
Link to IEEE Spectrum article: https://spectrum.ieee.org/tech-talk/semiconductors/devices/searching-for-the-perfect-neuron-for-ai
Illustration source: IEEE Spectrum