Josep L. Rosselló

A New Stochastic Computing Methodology for Efficient Neural Network Implementation

This paper presents a new methodology for the hardware implementation of neural networks (NNs) based on probabilistic laws. The proposed encoding scheme circumvents the limitations of classical stochastic computing (based on unipolar or bipolar encoding) extending the representation range to any real number using the ratio of two bipolar-encoded pulsed signals. Furthermore, the novel approach presents practically a total noise-immunity capability due to its specific codification. We introduce different designs for building the fundamental blocks needed to implement NNs. The validity of the present approach is demonstrated through a regression and a pattern recognition task. The low cost of the methodology in terms of hardware, along with its capacity to implement complex mathematical functions (such as the hyperbolic tangent), allows its use for building highly reliable systems and parallel computing.

FPGA-based stochastic echo state networks for time-series forecasting

Hardware implementation of artificial neural networks (ANNs) allows exploiting the inherent parallelism of these systems. Nevertheless, they require a large amount of resources in terms of area and power dissipation. Recently, Reservoir Computing (RC) has arisen as a strategic technique to design recurrent neural networks (RNNs) with simple learning capabilities. In this work, we show a new approach to implement RC systems with digital gates. The proposed method is based on the use of probabilistic computing concepts to reduce the hardware required to implement different arithmetic operations. The result is the development of a highly functional system with low hardware resources. The presented methodology is applied to chaotic time-series forecasting.

High-Density Liquid-State Machine Circuitry for Time-Series Forecasting

Spiking neural networks (SNN) are the last neural network generation that try to mimic the real behavior of biological neurons. Although most research in this area is done through software applications, it is in hardware implementations in which the intrinsic parallelism of these computing systems are more efficiently exploited. Liquid state machines (LSM) have arisen as a strategic technique to implement recurrent designs of SNN with a simple learning methodology. In this work, we show a new low-cost methodology to implement high-density LSM by using Boolean gates. The proposed method is based on the use of probabilistic computing concepts to reduce hardware requirements, thus considerably increasing the neuron count per chip. The result is a highly functional system that is applied to high-speed time series forecasting.

Ultra-Fast Data-Mining Hardware Architecture Based on Stochastic Computing

Minimal hardware implementations able to cope with the processing of large amounts of data in reasonable times are highly desired in our information-driven society. In this work we review the application of stochastic computing to probabilistic-based pattern-recognition analysis of huge database sets. The proposed technique consists in the hardware implementation of a parallel architecture implementing a similarity search of data with respect to different pre-stored categories. We design pulse-based stochastic-logic blocks to obtain an efficient pattern recognition system. The proposed architecture speeds up the screening process of huge databases by a factor of 7 when compared to a conventional digital implementation using the same hardware area.

A surface site interaction point methodology for macromolecules and huge molecular databases.

Determining the position and magnitude of Surface Site Interaction Points (SSIP) is a useful technique for understanding intermolecular interactions. SSIPs have been used for the prediction of solvation properties and for virtual co‐crystal screening. To determine the SSIPs for a molecule, the Molecular Electrostatic Potential Surface (MEPS) is first calculated using ab initio methods such as Density Functional Theory. This leads to a high cost in terms of computation time and is not compatible with the analysis of huge molecular databases. Herein, we present a method for the fast estimation of SSIPs, which is based on the MEPS calculated from MMFF94 atomic partial charges. The results show that this method can be used to calculate SSIPs for large molecular databases with a much higher speed than the original ab initio methodology.

Noise-robust hardware implementation of neural networks

Efficient hardware implementations of neural networks are of high interest. Stochastic computing is an alternative to conventional digital logic that allows to exploit the intrinsic parallelism of neural networks using few hardware resources. We present a new stochastic methodology that extends the capabilities of classical stochastic computing. In particular, the present approach exhibits practically total immunity to noise. This is demonstrated evaluating the influence of the noise on the systems performance for a mathematical regression task.

Cyclic Reservoir Computing with FPGA Devices for Efficient Channel Equalization

The reservoir computation (RC) is a recurrent neural network architecture that is very suitable for time series prediction tasks. Its implementation in specific hardware can be very useful in relation to software approaches, especially when low consumption is an essential requirement. However, the hardware realization of RC systems is expensive in terms of circuit area and power dissipation, mainly due to the need of a large number of multipliers at the synapses. In this paper, we present an implementation of an RC network with cyclic topology (simple cyclic reservoir) in which we limit the available synapses’ weights, which makes it possible to replace the multiplications with simple addition operations. This design is evaluated to implement the equalization of a non-linear communication channel, and allows significant savings in terms of hardware resources, presenting an accuracy comparable to previous works.

An improved methodology to compute surface site interaction points using high density molecular electrostatic potential surfaces: An Improved Methodology to Compute Surface Site Interaction Points Using High Density Molecular Electrostatic Potential Surfaces

The theoretical calculation of Surface Site Interaction Points (SSIP) has been used successfully in some applications in the solid and liquid phase. In this work we propose a new set of optimizations for the search of SSIP using the Molecular Electrostatic Potential Surfaces (MEPS) calculated with Density Functional Theory and B3LYP/6‐31*G basis set. The measures that have been implemented are based on the search for the best agreement between experimental H‐bond donor and acceptor parameters (α and β) and the MEPS extremes exploring a range of electron density levels. Additionally, a parameterization as a function of atom types has been performed. The results show that the MEPS calculated at 0.01 au electron density level slightly improves the correlation with experimental data in comparison with the calculation over other density values. This fact is related to the bigger contribution of local electrostatics at higher density levels. The refinement has provided significant improvements to the correlation between theoretical and experimental data. Moreover, the proposed calculation over 0.01 au is six times faster on average than the computation at 0.002 au. The proposed methodology has been developed with the purpose to obtain high precision SSIP in a fast way and to improve their applications in virtual cocrystal screening, calculation of free energies in solution and molecular docking.

A Bayesian Target Predictor Method based on Molecular Pairing Energies estimation

Virtual screening (VS) is applied in the early drug discovery phases for the quick inspection of huge molecular databases to identify those compounds that most likely bind to a given drug target. In this context, there is the necessity of the use of compact molecular models for database screening and precise target prediction in reasonable times. In this work we present a new compact energy-based model that is tested for its application to Virtual Screening and target prediction. The model can be used to quickly identify active compounds in huge databases based on the estimation of the molecule’s pairing energies. The greatest molecular polar regions along with its geometrical distribution are considered by using a short set of smart energy vectors. The model is tested using similarity searches within the Directory of Useful Decoys (DUD) database. The results obtained are considerably better than previously published models. As a Target prediction methodology we propose the use of a Bayesian Classifier that uses a combination of different active compounds to build an energy-dependent probability distribution function for each target.

Stochastic hardware implementation of Liquid State Machines.

The hardware implementation of neural network models allows to efficiently exploit their inherent parallelism. Here, we focus on the Liquid State Machine (LSM) methodology to build recurrent Spiking Neural Networks (SNN), particularly suited to process time-dependent signals. We propose a low cost hardware implementation of LSM networks based on the use of stochastic computing (SC) concepts. The functionality of the present approach is demonstrated for a time-series prediction task.