Nanodesign

Nanodevices provide fantastic opportunities for electronic systems, offering new functions in an extremely energy efficient way. However,  they also less desirable features like variability, and the difficulty to achieve acceptable yield. Fully exploiting nanoelectronics potential thus calls for novel architectures and design approaches.

We are developing pioneering approaches to exploit nanoelectronics, able to process the huge flow of natural data modern systems are confronted to. The bioinspired road is of special interest to us. Biology is indeed able to exploit its own nanodevices — which in some aspects are similar to the ones we are fabricating — in a fantastically efficient way.

Examples of our projects:

1. Synaptic computation (joint activity with the Spintronics @ IEF group)

Our synapse work

We develop new models of computation where nanodevices with memory effects (“memristive” devices, spintronic devices or resistive RAMs) are used as “synapses”, i.e. active connections. Such systems excel at processing natural noisy data in a power efficient way.

This research includes the study of the device physics and of the modelling of memristive technology. It is performed in strong collaboration with technological groups and CEA LIST (Saclay).

Look how nanodevices may be exploited as synapses to perform complicated recognition tasks, how such devices should be modelled , and the advanced applications such systems might have.

Our PCM workWith CEA LETI and CEA LIST, we have introduced the Phase Change synapse.

 

2. Neuronal computation

This project aims at associating nanodevices in a way inspired by the biophysics of neurons, to design and fabricate “nanoneurons”.

The final goal is to provide ultra low power, adaptive sensors.

 

Stochastic synapses PEPS project (2011/2012)

In this Projet Exploratoire Premier Soutien CNRS/INSIS , we are exploring how unpredictable nanodevices may achieve advanced cognitive tasks.
With Unité Mixte de Physique CNRS/Thalès, Institut Matériaux Microélectronique Nanosciences de Provence, CEA LIST.
We are looking for a student (Look for our open positions!).

 

Probabilistic Computing  G3N project (2012)

In this project funded by Mission Interdisciplinarité CNRS, we are defining a total new way of computing in electronics using Bayesian probability instead of logic, and thanks to nanotechnology. With Unité Mixte de Physique CNRS/Thalès, LPPA Collège de France and LIG Grenoble.

We are looking for a student.

 

Xnet (2010-…)

Xnet logo With CEA LIST, we are developing the “Xnet” simulator for designing systems with all kind of memristive synapses. Please contact us for availability.

 

Flagship projet of the Nano-Saclay labex (2011/…)

With Spintronics @ IEFUnité Mixte de Physique CNRS/ThalèsCEA LEM and CEA LISTINRIA Saclay.

 

Do not miss the “Memristors for Computing” workshop November 19-21 in Fréjus, France.

Contact: Damien Querlioz