CEA-Leti scientists have exhibited a machine-learning technique exploiting what has been previously considered as “non-ideal” traits of resistive-RAM (RRAM) devices, overcoming barriers to developing RRAM-based edge-learning systems.
Reported in a paper published in the January issue of Nature Electronics titled, “In-situ learning using intrinsic memristor variability via Markov chain Monte Carlo sampling”, the research team demonstrated how RRAM, or memristor, technology can be used to create intelligent systems that learn locally at the edge, independent of the cloud.
“This highlights that, beyond being RRAM-compatible, Bayesian machine learning offers an alternative modeling method that appears well suited to the characteristics of edge learning,” the article says.
“Our system could be used as the foundation for the design and fabrication of a standalone and fully integrated RRAM-based MCMC sampling chip, for applications outside the laboratory,” the authors of the article concluded. That achievement will finally open the door to edge learning and an entirely new set of applications.
The learning algorithms used in current RRAM-based edge approaches cannot be reconciled with device programming randomness, or variability, as well as other intrinsic non-idealities of the technology.
To get around that problem, the team developed a method that actively exploits that memristor randomness, implementing a Markov Chain Monte Carlo (MCMC) sampling learning algorithm in a fabricated chip that acts as a Bayesian machine-learning model.
Machine-learning models are normally trained using general-purpose hardware based on a von Neumann architecture, which is unsuited for edge learning because of the energy required to continuously move information between separated processing and memory centers on-chip.
Intensive research in the microelectronics industry is currently focused on using RRAM as non-volatile analog devices in hardware-based artificial neural networks that can allow computation to be carried out in memory, to drastically reduce these energy requirements.
RRAM has been applied to in-memory implementations of backpropagation algorithms to implement in-situ learning on edge systems. However, because back propagation requires high-precision memory elements, previous work has largely focused on how RRAM randomness can be mitigated – often necessitating energy-intensive techniques.
The CEA-Leti-based team’s breakthrough was an approach capable of leveraging this randomness, instead of trying to prevent it and allowing in-situ learning to be realized in a highly efficient fashion through the application of nanosecond voltage pulses to nanoscale memory devices.