1 Scientific RepoRts | 6:31932 | DOI: 10.1038/srep31932 www.nature.com/scientificreports physical Realization of a supervised Learning system Built with organic Memristive synapses Yu-pu Lin 1,* , Christopher H. Bennett 2,* , théo Cabaret 1 , Damir Vodenicarevic 2 , Djaafar Chabi 2 , Damien Querlioz 2 , Bruno Jousselme 1 , Vincent Derycke 1 & Jacques-olivier Klein 2 Multiple modern applications of electronics call for inexpensive chips that can perform complex operations on natural data with limited energy. A vision for accomplishing this is implementing hardware neural networks, which fuse computation and memory, with low cost organic electronics. A challenge, however, is the implementation of synapses (analog memories) composed of such materials. In this work, we introduce robust, fastly programmable, nonvolatile organic memristive nanodevices based on electrografted redox complexes that implement synapses thanks to a wide range of accessible intermediate conductivity states. We demonstrate experimentally an elementary neural network, capable of learning functions, which combines four pairs of organic memristors as synapses and conventional electronics as neurons. our architecture is highly resilient to issues caused by imperfect devices. It tolerates inter-device variability and an adaptable learning rule ofers immunity against asymmetries in device switching. Highly compliant with conventional fabrication processes, the system can be extended to larger computing systems capable of complex cognitive tasks, as demonstrated in complementary simulations. Biology-inspired electronics is currently attracting increasing attention as modern applications of electronics, such as biomedical systems, ubiquitous sensing, or the future Internet-of-hings, require systems able to deal with signiicant volumes of data, with a limited power budget. In the common von Neumann architecture of computers, an order of magnitude more energy is spent accessing memory than conducting arithmetic opera- tions. Whilst, bio-inspired computing schemes that fuse memory and computing ofer signiicant energy savings 1 . A fundamental bio-inspired architecture is the artiicial neural network (ANN), a system where neurons are con- nected to each other through numerous synapses 2 . Emerging nanoscale memories known as memristive devices have been proposed as ideal hardware analogues for the latter, while the former can be realized with standard transistor devices. herefore, a promising way to realize neuromorphic electronics is to build a hybrid system pairing transistor “neurons” interconnected via arrays of memristive devices, each which mimics a synaptic function 3–7 . Memristive nanodevices can mimic synaptic weights via non-linear conductivity, controllable by apply- ing voltage biases above characteristic device thresholds 7,8 . Simulated memristive ANNs have demonstrated capability to solve computational tasks using diverse algorithms 9–13 . Few experimental demonstrations of complete memristive ANNs exist; those built so far generally exploit inorganic devices 14–19 or three terminal nanodevices 20,21 . However, memristive devices can also be made with organic materials that are fundamentally attractive 22,23 as they ofer unique advantages: low material costs, scalable fabrication via roll-to-roll imprint lithography, and compatibility with lexible substrates. hese properties pave the way towards integration with embedded sensors, bio-medical devices, and other internet of things applications 24,25 , yet oten come at the cost of slower programming relative to inorganic memristive devices or binary organic memory devices 26,27 . he only ANN with organic memristors uses polyaniline polymeric devices 28 , with programming durations too slow for applications (30 s per programming pulse). Here, we introduce the irst demonstrator circuit capable of learning with organically-composed memristive devices as synapses that works at speeds relevant for applications (100 μs 1 LICSEN, NIMBE, CEA, CNRS, Université Paris-Saclay, CEA Saclay 91191 Gif-sur-Yvette, France. 2 Institut d’Electronique Fondamentale, Université Paris-Sud/Paris-Saclay, CNRS, 91405 Orsay, France. *These authors contributed equally to this work. Correspondence and requests for materials should be addressed to or C.H.B. (email: christopher.bennett@u-psud.fr) or V.D. (email: vincent.derycke@cea.fr) or J.-O.K. (email: jacques-olivier.klein@u-psud.fr). received: 24 May 2016 accepted: 27 July 2016 Published: 07 September 2016 opeN