Sodium-Controlled Interfacial Resistive Switching in Thin Film Niobium Oxide for Neuromorphic Applications.

A double layer 2-terminal device is employed to show Na-controlled interfacial resistive switching and neuromorphic behavior. The bilayer is based on interfacing biocompatible NaNbO3 and Nb2O5, which allows the reversible uptake of Na+ in the Nb2O5 layer. We demonstrate voltage-controlled interfacial barrier tuning via Na+ transfer, enabling conductivity modulation and spike-amplitude- and spike-timing-dependent plasticity. The neuromorphic behavior controlled by Na+ ion dynamics in biocompatible materials shows potential for future low-power sensing electronics and smart wearables with local processing.

Deep causal speech enhancement and recognition using efficient long-short term memory Recurrent Neural Network.

Long short-term memory (LSTM) has been effectively used to represent sequential data in recent years. However, LSTM still struggles with capturing the long-term temporal dependencies. In this paper, we propose an hourglass-shaped LSTM that is able to capture long-term temporal correlations by reducing the feature resolutions without data loss. We have used skip connections in non-adjacent layers to avoid gradient decay. In addition, an attention process is incorporated into skip connections to emphasize the essential spectral features and spectral regions. The proposed LSTM model is applied to speech enhancement and recognition applications. The proposed LSTM model uses no future information, resulting in a causal system suitable for real-time processing. The combined spectral feature sets are used to train the LSTM model for improved performance. Using the proposed model, the ideal ratio mask (IRM) is estimated as a training objective. The experimental evaluations using short-time objective intelligibility (STOI) and perceptual evaluation of speech quality (PESQ) have demonstrated that the proposed model with robust feature representation obtained higher speech intelligibility and perceptual quality. With the TIMIT, LibriSpeech, and VoiceBank datasets, the proposed model improved STOI by 16.21%, 16.41%, and 18.33% over noisy speech, whereas PESQ is improved by 31.1%, 32.9%, and 32%. In seen and unseen noisy situations, the proposed model outperformed existing deep neural networks (DNNs), including baseline LSTM, feedforward neural network (FDNN), convolutional neural network (CNN), and generative adversarial network (GAN). With the Kaldi toolkit for automated speech recognition (ASR), the proposed model significantly reduced the word error rates (WERs) and reached an average WER of 15.13% in noisy backgrounds.

Multi-Method Analysis of Histopathological Image for Early Diagnosis of Oral Squamous Cell Carcinoma Using Deep Learning and Hybrid Techniques.

Oral cancer is a fatal disease and ranks seventh among the most common cancers throughout the whole globe. Oral cancer is a type of cancer that usually affects the head and neck. The current gold standard for diagnosis is histopathological investigation, however, the conventional approach is time-consuming and requires professional interpretation. Therefore, early diagnosis of Oral Squamous Cell Carcinoma (OSCC) is crucial for successful therapy, reducing the risk of mortality and morbidity, while improving the patient's chances of survival. Thus, we employed several artificial intelligence techniques to aid clinicians or physicians, thereby significantly reducing the workload of pathologists. This study aimed to develop hybrid methodologies based on fused features to generate better results for early diagnosis of OSCC. This study employed three different strategies, each using five distinct models. The first strategy is transfer learning using the Xception, Inceptionv3, InceptionResNetV2, NASNetLarge, and DenseNet201 models. The second strategy involves using a pre-trained art of CNN for feature extraction coupled with a Support Vector Machine (SVM) for classification. In particular, features were extracted using various pre-trained models, namely Xception, Inceptionv3, InceptionResNetV2, NASNetLarge, and DenseNet201, and were subsequently applied to the SVM algorithm to evaluate the classification accuracy. The final strategy employs a cutting-edge hybrid feature fusion technique, utilizing an art-of-CNN model to extract the deep features of the aforementioned models. These deep features underwent dimensionality reduction through principal component analysis (PCA). Subsequently, low-dimensionality features are combined with shape, color, and texture features extracted using a gray-level co-occurrence matrix (GLCM), Histogram of Oriented Gradient (HOG), and Local Binary Pattern (LBP) methods. Hybrid feature fusion was incorporated into the SVM to enhance the classification performance. The proposed system achieved promising results for rapid diagnosis of OSCC using histological images. The accuracy, precision, sensitivity, specificity, F-1 score, and area under the curve (AUC) of the support vector machine (SVM) algorithm based on the hybrid feature fusion of DenseNet201 with GLCM, HOG, and LBP features were 97.00%, 96.77%, 90.90%, 98.92%, 93.74%, and 96.80%, respectively.

Comprehensive study of Raman optical response of typical substrates for thin-film growth under 633†nm and 785†nm laser excitation.

Raman spectroscopy is one of the most efficient and non-destructive techniques for characterizing materials. However, it is challenging to analyze thin films using Raman spectroscopy since the substrates beneath the thin film often obscure its optical response. Here, we evaluate the suitability of fourteen commonly employed single-crystal substrates for Raman spectroscopy of thin films using 633 nm and 785 nm laser excitation systems. We determine the optimal wavenumber ranges for thin-film characterization by identifying the most prominent Raman peaks and their relative intensities for each substrate and across substrates. In addition, we compare the intensity of background signals across substrates, which is essential for establishing their applicability for Raman detection in thin films. The substrates LaAlO3 and Al2O3 have the largest free spectral range for both laser systems, while Al2O3 has the lowest background levels, according to our findings. In contrast, the substrates SrTiO3 and Nb:SrTiO3 have the narrowest free spectral range, while GdScO3, NGO and MgO have the highest background levels, making them unsuitable for optical investigations.

Investigation of external quality factor and coupling coefficient for a novel SIR based microstrip tri-band bandpass filter.

In this article, a new method is developed to design a three-band miniaturized bandpass filter (BPF) that uses two asymmetrically coupled resonators with one step discontinuity and open-circuited uniform impedance resonator (UIR) to achieve Global Interoperability with Microwave Access (WiMAX) and Radio Frequency Identification (RFID) applications. First, a pair of asymmetrical step impedance resonators (ASIR) is used to implement a dual band filter, then a half wavelength uniform impedance resonator is added below to the transmission line to achieve a triple band response. The proposed filter resonates at frequencies of 3.7 GHz, 6.6 GHz, and 9 GHz with the fractional bandwidth of 7.52%, 5.1%, and 4.44%, respectively. By adjusting the physical length ratio (α) and the impedance ratio (R) of the asymmetric SIR, the proposed fundamental frequencies of the triple BPF are obtained. Moreover, the coupling coefficient (Ke) and external quality factor (Qe) are investigated between the resonators and the input/output ports of the transmission line and are calculated using full-wave EM simulator HFSS. In addition, five transmission zeros are introduced near the passbands to increase the filter selectivity. Finally, the proposed filter is designed and fabricated with a size of 13.69 × 25 mm (0.02 λg × 0.03 λg), where λg represents the guiding wavelength in the first passband. The simulated and measured results have a good correspondence, thus confirming the design concept.

Multiple nano-filaments based efficient resistive switching in TiO2 nanotubes array influenced by thermally induced self-doping and anatase to rutile phase transformation.

In this paper, the impact of thermally induced self-doping and phase transformation in TiO2 based resistive random-access memory (ReRAM) is discussed. Instead of a thin film, a vertically aligned one-dimensional TiO2 nanotube array (TNTA) was used as a switching element. Anodic oxidation method was employed to synthesize TNTA, which was thermally treated in the air at 350 °C followed by further annealing from 350 °C to 650 °C in argon. Au/TiO2 nanotube/Ti resistive switching devices were fabricated with porous gold (Au) top electrode. The x-ray diffraction results along with Raman spectra evidently demonstrate a change in phase of crystallinity from anatase to rutile, whereas photoluminescence spectra revealed the self-doping level in terms of oxygen vacancies (OV) and Ti interstitials (Tii) as the temperature of thermal treatment gets increased. The electrical characterizations establish the bipolar and electroforming free resistive switching in all the samples. Among those, the ReRAM sample S3 thermally treated at 550 °C displayed the most effective resistive switching properties with R OFF/R ON of 102 at a read voltage of -0.6 V and a SET voltage of -2.0 V. Moreover, the S3 sample showed excellent retention performance for over 106 s, where stable R OFF/R ON ≈ 107 was maintained throughout the experiment.