The identification and modelling of a percussion ‘language,’ and the Emergence of Musical Concepts in a machine-learning experimental set-up

In experimental research into percussion ‘languages’, an interactive computer system, the Bol Processor, has been developed by the authors to analyse the performances of expert musicians and generate its own musical items that were assessed for quality and accuracy by the informants. The problem of transferring knowledge from a human expert to a machine in this context is the focus of this paper. A prototypical grammatical inferencer named QAVAID (Question Answer Validated Analytical Inference Device, an acronym also meaning ‘grammar’ in Arabic/Urdu) is described and its operation in a real experimental situation is demonstrated. The paper concludes on the nature of the knowledge acquired and the scope and limitations of a cognitive-computational approach to music. Bernard Bel is an electronics and computer engineer. A founder member of the International Society for Traditional Arts Research (ISTAR) he has for many years collaborated with ethnomusicologists and musicians on projects aimed at a scientific study of North Indian melodic and rhythmic systems. In 1981 he designed and constructed an accurate melodic movement analyser, and he subsequently developed software for the analysis of raga intonation. He went on to develop software for rhythmic analysis/synthesis in collaboration with Jim Kippen. Bernard Bel is now a member of the Groupe Représentation et Traitement des Connaissances, an AI laboratory at the CNRS, Marseille.     

Holographic storage in bleached emulsion of N divergent object beams generated by a two-dimensional regular array: analysis of the signal-to-noise ratio

The signal-to-noise ratio (SNR) of multiplexed holograms of N object waves recorded as volume-phase holograms in bleached silver halide emulsion are experimentally analyzed and compared with those of diffuse-object holograms. We introduce two types of SNR: one belonging to the spatial matrix itself, and the other, the overall SNR, for which the angular range of the matrix distribution was compared with the primary diffuse object. The experimental results obtained show that, even though the same range of spatial frequencies is maintained, the total number of waves that come from the matrix distribution affects the measurements, so the interrelation between the spatial-frequency distribution of noise and the number of storage waves is obtained, and the minimum SNR can be evaluated for diffuse objects.     

Protein‐Corona‐by‐Design in 2D: A Reliable Platform to Decode Bio–Nano Interactions for the Next‐Generation Quality‐by‐Design Nanomedicines

Hard corona (HC) protein, i.e., the environmental proteins of the biological medium that are bound to a nanosurface, is known to affect the biological fate of a nanomedicine. Due to the size, curvature, and specific surface area (SSA) 3‐factor interactions inherited in the traditional 3D nanoparticle, HC‐dependent bio–nano interactions are often poorly probed and interpreted. Here, the first HC‐by‐design case study in 2D is demonstrated that sequentially and linearly changes the HC quantity using functionalized graphene oxide (GO) nanosheets. The HC quantity and HC quality are analyzed using NanoDrop and label‐free liquid chromatography–mass spectrometry (LC‐MS) followed by principal component analysis (PCA). Cellular responses (uptake and cytotoxicity in J774 cell model) are compared using imaging cytometry and the modified lactate dehydrogenase assays, respectively. Cellular uptake linearly and solely correlates with HC quantity (R² = 0.99634). The nanotoxicity, analyzed by retrospective design of experiment (DoE), is found to be dependent on the nanomaterial uptake (primary), HC composition (secondary), and nanomaterial exposure dose (tertiary). This unique 2D design eliminates the size–curvature–SSA multifactor interactions and can serve as a reliable screening platform to uncover HC‐dependent bio–nano interactions to enable the next‐generation quality‐by‐design (QbD) nanomedicines for better clinical translation.     

Object-based urban structure type pattern recognition from Landsat TM with a Support Vector Machine

This study evaluates the potential of object-based image analysis in combination with supervised machine learning to identify urban structure type patterns from Landsat Thematic Mapper (TM) images. The main aim is to assess the influence of several critical choices commonly made during the training stage of a learning machine on the classification performance and to give recommendations for classifier-dependent intelligent training. Particular emphasis is given to assess the influence of size and class distribution of the training data, the approach of training data sampling (user-guided or random) and the type of training samples (squares or segments) on the classification performance of a Support Vector Machine (SVM). Different feature selection algorithms are compared and segmentation and classifier parameters are dynamically tuned for the specific image scene, classification task, and training data. The performance of the classifier is measured against a set of reference data sets from manual image interpretation and furthermore compared on the basis of landscape metrics to a very high resolution reference classification derived from light detection and ranging (lidar) measurements. The study highlights the importance of a careful design of the training stage and dynamically tuned classifier parameters, especially when dealing with noisy data and small training data sets. For the given experimental set-up, the study concludes that given optimized feature space and classifier parameters, training an SVM with segment-shaped samples that were sampled in a guided manner and are balanced between the classes provided the best classification results. If square-shaped samples are used, a random sampling provided better results than a guided selection. Equally balanced sample distributions outperformed unbalanced training sets.     

Power saving through state retention in IGZO‐TFT AMOLED displays for wearable applications

We present a qHD (960 × 540 with three sub‐pixels) top‐emitting active‐matrix organic light‐emitting diode display with a 340‐ppi resolution using a self‐aligned IGZO thin‐film transistor backplane on polyimide foil with a humidity barrier. The back plane process flow is based on a seven‐layer photolithography process with a CD = 4 μm. We implement a 2T1C pixel engine and use a commercial source driver IC made for low‐temperature polycrystalline silicon. By using an IGZO thin‐film transistor and leveraging the extremely low off current, we can switch off the power to the source and gate driver while maintaining the image unchanged for several minutes. We demonstrate that, depending on the image content, low‐refresh operation yields reduction in power consumption of up to 50% compared with normal (continuous) operation. We show that with the further increase in resolution, the power saving through state retention will be even more significant.     

Crystallographic Orientation in Pure and Aluminum-Doped Zinc Oxide Thin Films Prepared by Sol–Gel Technique

Novel sol–gel processes for the preparation of grain-oriented thin films of pure and aluminum-doped zinc oxide (AZO) are reported. Using various stabilizing ligands, oxide films with different degree of c -axis orientation could be prepared on amorphous substrates. The origin of orientation in sol–gel-derived ZnO and AZO films is investigated. The texture selection of ZnO films is independent of the heat-treatment schedule but sensitive to the precursor system. The electrical properties of the films do not depend significantly on the crystallographic orientation. The film visible transmittance remained higher than 90% even after vacuum annealing.     

Performance analysis of SSE and AVX instructions in multi-core CPUs and GPU computing on FDTD scheme for solid and fluid vibration problems

In this work a unified treatment of solid and fluid vibration problems is developed by means of the Finite-Difference Time-Domain (FDTD). The scheme here proposed takes advantage from a scaling factor in the velocity fields that improves the performance of the method and the vibration analysis in heterogenous media. Moreover, the scheme has been extended in order to simulate both the propagation in porous media and the lossy solid materials. In order to accurately reproduce the interaction of fluids and solids in FDTD both time and spatial resolutions must be reduced compared with the set up used in acoustic FDTD problems. This aspect implies the use of bigger grids and hence more time and memory resources. For reducing the time simulation costs, FDTD code has been adapted in order to exploit the resources available in modern parallel architectures. For CPUs the implicit usage of the advanced vectorial extensions (AVX) in multi-core CPUs has been considered. In addition, the computation has been distributed along the different cores available by means of OpenMP directives. Graphic Processing Units have been also considered and the degree of improvement achieved by means of this parallel architecture has been compared with the highly-tuned CPU scheme by means of the relative speed up. The speed up obtained by the parallel versions implemented were up to 3 (AVX and OpenMP) and 40 (CUDA) times faster than the best sequential version for CPU that also uses OpenMP with auto-vectorization techniques, but non includes implicitely vectorial instructions. Results obtained with both parallel approaches demonstrate that massive parallel programming techniques are mandatory in solid-vibration problems with FDTD.     

A review of conventional and knowledge based systems for machining price quotation

Price estimation for the preparation of quotations is critical process for companies that have to struggle to get orders by offering competitive pricing. For machining companies, said process is complex because of the large amount of variation that can occur. In this case, a manufacturing expert is generally in charge of this task. However, manufacturing experts have several other important tasks to which they must attend. The use of software systems that automate the estimation of costs and prices is common, knowledge-based systems are one of the primary alternatives for two fundamental reasons: they behave the same way an expert would and they save and maintain all knowledge within the company regardless of who has worked on them. This paper covers the principal cost and price estimation methods and reviews the knowledge-based systems that have been implemented in the area of machined part manufacturing. Recommendations as to how future knowledge-based systems for the estimation of this type of pricing should develop are also included.