CHARLES: A C++ fixed-point library for Photonic-Aware Neural Networks.

In this paper we present CHARLES (C++ pHotonic Aware neuRaL nEtworkS), a C++ library aimed at providing a flexible tool to simulate the behavior of Photonic-Aware Neural Network (PANN). PANNs are neural network architectures aware of the constraints due to the underlying photonic hardware, mostly in terms of low equivalent precision of the computations. For this reason, CHARLES exploits fixed-point computations for inference, while it supports both floating-point and fixed-point numerical formats for training. In this way, we can compare the effects due to the quantization in the inference phase when the training phase is performed on a classical floating-point model and on a model exploiting high-precision fixed-point numbers. To validate CHARLES and identify the most suited numerical format for PANN training, we report the simulation results obtained considering three datasets: Iris, MNIST, and Fashion-MNIST. Fixed-training is shown to outperform floating-training when executing inference on bitwidths suitable for photonic implementation. Indeed, performing the training phase in the floating-point domain and then quantizing to lower bitwidths results in a very high accuracy loss. Instead, when fixed-point numbers are exploited in the training phase, the accuracy loss due to quantization to lower bitwidths is significantly reduced. In particular, we show that for Iris dataset, fixed-training achieves a performance similar to floating-training. Fixed-training allows to obtain an accuracy of 90.4% and 68.1% with the MNIST and Fashion-MNIST datasets using only 6 bits, while the floating-training reaches an accuracy of just 25.4% and 50.0% when exploiting the same bitwidths.

Fast Approximations of Activation Functions in Deep Neural Networks when using Posit Arithmetic.

With increasing real-time constraints being put on the use of Deep Neural Networks (DNNs) by real-time scenarios, there is the need to review information representation. A very challenging path is to employ an encoding that allows a fast processing and hardware-friendly representation of information. Among the proposed alternatives to the IEEE 754 standard regarding floating point representation of real numbers, the recently introduced Posit format has been theoretically proven to be really promising in satisfying the mentioned requirements. However, with the absence of proper hardware support for this novel type, this evaluation can be conducted only through a software emulation. While waiting for the widespread availability of the Posit Processing Units (the equivalent of the Floating Point Unit (FPU)), we can already exploit the Posit representation and the currently available Arithmetic-Logic Unit (ALU) to speed up DNNs by manipulating the low-level bit string representations of Posits. As a first step, in this paper, we present new arithmetic properties of the Posit number system with a focus on the configuration with 0 exponent bits. In particular, we propose a new class of Posit operators called L1 operators, which consists of fast and approximated versions of existing arithmetic operations or functions (e.g., hyperbolic tangent (TANH) and extended linear unit (ELU)) only using integer arithmetic. These operators introduce very interesting properties and results: (i) faster evaluation than the exact counterpart with a negligible accuracy degradation; (ii) an efficient ALU emulation of a number of Posits operations; and (iii) the possibility to vectorize operations in Posits, using existing ALU vectorized operations (such as the scalable vector extension of ARM CPUs or advanced vector extensions on Intel CPUs). As a second step, we test the proposed activation function on Posit-based DNNs, showing how 16-bit down to 10-bit Posits represent an exact replacement for 32-bit floats while 8-bit Posits could be an interesting alternative to 32-bit floats since their performances are a bit lower but their high speed and low storage properties are very appealing (leading to a lower bandwidth demand and more cache-friendly code). Finally, we point out how small Posits (i.e., up to 14 bits long) are very interesting while PPUs become widespread, since Posit operations can be tabulated in a very efficient way (see details in the text).

Mission Planning and Decision Support for Underwater Glider Networks: A Sampling on-Demand Approach.

This paper describes an optimal sampling approach to support glider fleet operators and marine scientists during the complex task of planning the missions of fleets of underwater gliders. Optimal sampling, which has gained considerable attention in the last decade, consists in planning the paths of gliders to minimize a specific criterion pertinent to the phenomenon under investigation. Different criteria (e.g., A, G, or E optimality), used in geosciences to obtain an optimum design, lead to different sampling strategies. In particular, the A criterion produces paths for the gliders that minimize the overall level of uncertainty over the area of interest. However, there are commonly operative situations in which the marine scientists may prefer not to minimize the overall uncertainty of a certain area, but instead they may be interested in achieving an acceptable uncertainty sufficient for the scientific or operational needs of the mission. We propose and discuss here an approach named sampling on-demand that explicitly addresses this need. In our approach the user provides an objective map, setting both the amount and the geographic distribution of the uncertainty to be achieved after assimilating the information gathered by the fleet. A novel optimality criterion, called A η , is proposed and the resulting minimization problem is solved by using a Simulated Annealing based optimizer that takes into account the constraints imposed by the glider navigation features, the desired geometry of the paths and the problems of reachability caused by ocean currents. This planning strategy has been implemented in a Matlab toolbox called SoDDS (Sampling on-Demand and Decision Support). The tool is able to automatically download the ocean fields data from MyOcean repository and also provides graphical user interfaces to ease the input process of mission parameters and targets. The results obtained by running SoDDS on three different scenarios are provided and show that SoDDS, which is currently used at NATO STO Centre for Maritime Research and Experimentation (CMRE), can represent a step forward towards a systematic mission planning of glider fleets, dramatically reducing the efforts of glider operators.

Approaching the Ocean Color problem using fuzzy rules.

In this paper, we propose a fuzzy logic-based approach which exploits remotely sensed multispectral measurements of the reflected sunlight to estimate the concentration of optically active constituents of the sea water. The relation between the concentrations of interest and the subsurface reflectances is modeled by a set of fuzzy rules extracted automatically from the data through a two-step procedure. First, a compact initial rule base is generated by projecting onto the input variables the clusters produced by a fuzzy clustering algorithm. Then, a genetic algorithm is applied to optimize the rules. Appropriate constraints maintain the semantic properties of the initial model during the genetic evolution. Results of the application of the fuzzy model obtained from data simulated with an ocean color model over the channels of the Medium Resolution Imaging Spectrometer are shown and discussed.