No abstract

No abstract

No abstract

In this paper, we present an alternative approach to perform spike sorting of complex brain signals based on spiking neural networks (SNN). The proposed architecture is suitable for hardware implementation by using resistive random access memory (RRAM) technology for the implementation of synapses whose low latency (<1μs) enables real-time spike sorting. This offers promising advantages to conventional spike sorting techniques for brain-computer interfaces (BCI) and neural prosthesis applications. Moreover, the ultra-low power consumption of the RRAM synapses of the spiking neural network (nW range) may enable the design of autonomous implantable devices for rehabilitation purposes. We demonstrate an original methodology to use Oxide based RRAM (OxRAM) as easy to program and low energy (<75 pJ) synapses. Synaptic weights are modulated through the application of an online learning strategy inspired by biological Spike Timing Dependent Plasticity. Real spiking data have been recorded both intra- and extracellularly from an in-vitro preparation of the Crayfish sensory-motor system and used for validation of the proposed OxRAM based SNN. This artificial SNN is able to identify, learn, recognize and distinguish between different spike shapes in the input signal with a recognition rate about 90% without any supervision.

Ga–Sb alloys are potential candidates for phase change random access memory (PCRAM) applications. Ga–Sb alloys of variable compositions including the single stoichiometric GaSb and several Sb‐rich compositions were studied using resistivity versus temperature measurements, static laser testing, and time‐resolved X‐ray diffraction. It was found that the stoichiometric alloy has an unusual inverse optical contrast compared to typical phase change materials as the crystalline phase has lower reflectance compared to the amorphous phase. Sb‐rich alloys show a decrease in reflectance upon crystallization at lower temperature but an increase in reflectance at higher temperature from subsequent Sb segregation. Alloys very rich in Sb only show a positive change in reflectance upon crystallization typical for conventional phase change materials. magnified imageChange in reflectance of a Ga/Sb = 36:64 (in at.%) amorphous thin film crystallized by laser pulses of variable length and power. Unusual for a phase change material, this alloy shows both positive and negative contrast at different laser power levels.

IntroductionSoftware implementations of artificial Convolutional Neural Networks (CNNs), taking inspiration from biology, are at the state-of-the-art for Pattern Recognition (PR) applications and they are successfully used in commercial products [1]. However, they require power-hungry CPU/GPU to perform convolution operations based on computationally expensive sums of multiplications. This hinders their integration in portable devices. Some full CMOS-based hardware implementations of CNN have been suggested, but they still require the computation of multiplications [2]. In this work, we present for the first time to our knowledge a spike-based hardware implementation of CNN using HfO 2 based OxRAM devices as binary synapses. OxRAM devices are chosen for their low switching energy [3] and promising endurance performance [4]. We perform an experimental and theoretical study of the impact of programming conditions at both device and system levels. A complex visual pattern recognition application is demonstrated with a spike-based hierarchical CNN, inspired from the mammalian visual cortex organization. A high accuracy (pattern recognition rate >94%) is obtained for all the tested programming conditions, even if the variability associated to weaker programming conditions is larger. Electrical characterization and physical modelingWe tested 1T-1R OxRAM devices, integrated in standard 65nm CMOS technology [5] (Fig.1). The OxRAM device is composed of a 5 nm thick HfO x layer deposited by ALD embedded between a 10 nm thick Ti and a 35 nm TiN electrodes. OxRAM operating relies on formation/dissolution of an oxygen vacancy rich conductive filament (CF). Typical I-V characteristics are reported in Fig.1. Fig.2 shows the impact of the programming conditions on both the high resistance state (HRS) and the low resistance state (LRS) during pulsed cycling (T=100ns). The LRS is controlled by the compliance current during SET operation (I C SET ), while the HRS is tuned by means of the RESET voltage V BL . The corresponding LRS and HRS Cumulative Distribution Function (CDF) are reported in Fig. 3(a) and (b), respectively. While a bending of LRS CDF is observed when decreasing I C SET [6], the width of HRS CDFs corresponding to different reset voltages remains constant [7]. The values corresponding to the mean µ R , and standard deviation σ R , are also defined in Fig.3. A 2D multi-Trap Assisted Tunneling model (Fig.4) based on [8] is developed in order to interpret these results. The model is based on the hypothesis that during SET/RESET operations the CF is only partially disrupted/reformed in a L GAP region ( Fig.4) [9-10]. We model only this critical region, since the cell resistance value is mostly determined by its properties (gap length L GAP , cross-section D and Oxygen Vacancy concentration ). We simulate the LRS values assuming that an I C SET increase corresponds to a higher . The increase of the HRS for high V BL values is modeled tuning L GAP , while keeping constant (i.e. equal residual defect conc...

No abstract

Selector device is critical in high-density cross-point resistive switching memory arrays for suppressing the sneak leakage current path. GexSe1-x based ovonic threshold switch (OTS) selectors have recently demonstrated strong performance with high on-state current, nonlinearity and endurance. Detailed study of its reliability is still lacking and the understanding on the responsible mechanisms is limited. In this work, for the first time, the endurance degradation mechanism of Ge-rich GexSe1-x OTS is identified. Accumulation of slow defects that remain delocalized at off-state and GeSe segregation/crystallization during cycling lead to the recoverable and non-recoverable leakage current, respectively. Most importantly, a refreshing program scheme is developed to recover and prevent the OTS degradation and the endurance can be therefore improved by more than five orders without adding additional material elements or process steps.