Search

Metal-assisted chemical etched (MACE; also known as MacEtch or MCCE) nanostructures are utilized widely in the solar cell industry due to their excellent optical properties combined with a simple and cost-efficient fabrication process. The photodetection community, on the other hand, has not shown much interest toward MACE due to its drawbacks, including insufficient surface passivation, increased junction recombination, and possible metal contamination, which are especially detrimental to p-n photodiodes. Here, we aim to change this by demonstrating how to fabricate high-performance MACE p-n photodiodes with above 90% external quantum efficiency (EQE) without external bias voltage at 200-1000 nm and dark current less than 3 nA/cm2 at -5 V using industrially applicable methods. The key is to utilize an induced junction created by an atomic layer deposited (ALD) highly charged Al2O3 thin film that simultaneously provides efficient field-effect passivation and full conformality over the MACE nanostructures. Achieving close to ideal performance demonstrates the vast potential of MACE nanostructures in the fabrication of high-performance low-cost p-n photodiodes.

End-to-end Spoken Language Understanding (E2E SLU) has attracted increasing interest due to its advantages of joint optimization and low latency when compared to traditionally cascaded pipelines. Existing E2E SLU models usually follow a two-stage configuration where an Automatic Speech Recognition (ASR) network first predicts a transcript which is then passed to a Natural Language Understanding (NLU) module through an interface to infer semantic labels, such as intent and slot tags. This design, however, does not consider the NLU posterior while making transcript predictions, nor correct the NLU prediction error immediately by considering the previously predicted word-pieces. In addition, the NLU model in the two-stage system is not streamable, as it must wait for the audio segments to complete processing, which ultimately impacts the latency of the SLU system. In this work, we propose a streamable multi-task semantic transducer model to address these considerations. Our proposed architecture predicts ASR and NLU labels auto-regressively and uses a semantic decoder to ingest both previously predicted word-pieces and slot tags while aggregating them through a fusion network. Using an industry scale SLU and a public FSC dataset, we show the proposed model outperforms the two-stage E2E SLU model for both ASR and NLU metrics., Accepted at ICASSP 2022

In this paper, an ultra-low-voltage, low-power pseudo-differential CMOS Schmitt trigger is presented. The bulk-driven and sub-threshold techniques are used to achieve low-voltage and low-power circuit. The regenerative current feedback is applied to provide the hysteresis of the op-amp based Schmitt trigger. The proposed Schmitt trigger is designed and simulated in 0.18 μm CMOS technology and it is operated in 0.4 V supply voltage with 150 nW power consumption.

The mechanism by which the frequency of an LC VCO is sensitive to the power supply is analyzed. It is shown that variations in both the common-mode and differential-mode components can give rise to periodic jitter in the presence of supply variations due to capacitive nonlinearities. A new compensation method that reduces this sensitivity is presented. Simulations are shown verifying that this method can reduce the periodic jitter by more than 80%.

An important application of 3-D integration technology is stacked processor-DRAM systems. One of the major design issues in 3-D processor-DRAM stacks is power delivery. Presence of hot spots, high density power regions, in processor die poses serious challenges to the design of power distribution network (PDN). In this paper, we investigate solutions to ensure power integrity in the hot spot regions.

The current article studies optimal intercore interconnect network in a NoC structure for 2D and 3D mesh, torus and hypercube topologies. Optimal wire width/spacing is calculated by numerically maximizing bandwidth times the reciprocal delay, which depends on the technology node and hop length. Through 3D integration and increasing tiers, optimal interconnect width and spacing in torus and hypercube topologies will decrease. The core-to-core channel width in all topologies will be obtained by assigning 20% of the power consumption to the routers. By increasing number of cores, channel width will decrease due to reduced power consumption of each core. This is more in hypercube topology, due to the fact that the number of router ports will increase along with an increase in the number of cores. In terms of the worst case delay, mesh topology is worse than the two other topologies. Also it is not scalable due to the increase in the number of cores. In all topologies, power consumption of the chip and the worst case delay will decrease by 3D integration and utilizing more tiers. Mesh and torus topologies make the least and the most use of wiring area, respectively. Bisection bandwidth increases in all topologies by 3D integration.

Data Mining (DM) is a powerful new technique to help companies discover the patterns and trends in their customers' preferences. It is also a well-known tool for customer relationship management (CRM). Data mining methodology has made a tremendous contribution for researchers wanting to extract hidden knowledge and information. This study has proposed a new procedure, based on an expanded RFM model, by including two additional parameters D and C. It constructs a model for clustering customer value based on RFMDC attributes and K-means algorithm. We evaluate the result and suggest suitable behavior policies for each cluster. The developed methodology has been implemented for Kalleh dairy company in Iran to illustrate the proposed procedure.

Forest roads have an important role in the forest management and economic of countries. The Caspian forest is the most important forests region in Iran. It's so important to construction of forest roads in the best place that can be useful. Haraz region is located inMazandaran province that is near to Alborz Mountain. Selection the best place for construction of a road in forest is an important problem. For this research three place are considered for evaluating in Haraz region that including: Kelerd, Pelet Cheshme and Mangel. In this paper we applied hybrid MCDM methods for evaluating the regions. AHP applied for calculating the weight of each criterion and sub criterion and then COPRAS-G method applied for evaluating the places for selecting the best place for constructing the forest road. Result showed that Kelerd region in the best place for this work.

The world manufacturing sector is facing massive challenges to survive in today's global and volatile market place. In an attempt to overcome these challenges, companies are adopting newer method to performance evaluation of industries and Small & Medium Size Enterprises (SME). The purpose of this study is to construct balanced score card (BSC) for steel industries. This research summarizes the evaluation indexes synthesized from the literatures relating to company performance first. Then, it introduces indexes for performance evaluation selected through experts' questionnaires. Furthermore, the relative weights of the chosen evaluation indexes are calculated by Fuzzy Analytic Hierarchy Process (FAHP). Weights of each criterion are described by linguistic terms which can be expressed in triangular fuzzy numbers. Fuzzy TOPSIS method, which is one of Multiple Criteria Decision Making (MCDM) analytical tool, is used to rank three mills' performance by calculating the distances to both the fuzzy positive-ideal solution (FPIS) and fuzzy negative-ideal solution (FNIS) simultaneously according to three case studies as empirical examples.

The similarities between unvoiced speech sounds and turbulent breath sounds were used to detect change in sound characteristics caused by narrowing of the upper airway (UA), similar to that occurring in obstructive sleep apnea (OSA). In 18 awake subjects, UA resistance (R AU ), an index of UA narrowing, was measured simultaneously with breath sounds recording. Linear Prediction Coding was applied on turbulent inspiratory sounds drawn from low and high R AU conditions and K-means was used to cluster the resulting coefficients. The resulting 2 clusters were tested for agreement with the underlying R AU status. Distinct clusters were formed when R UA increased relatively high but not in cases with lower rise in R UA (P

We present MPtracker, a new algorithm for tracking and separating the pitch frequencies of two speakers from their mixture. The pitch frequencies are detected by introducing a novel spectral distortion optimization which takes into account the sinusoidal modeling of the speech signal. The detected pitch frequencies are grouped, separated, and finally an interpolation method is applied to estimate missing pitch frequencies. We evaluated the performance of the proposed technique on 196 mixtures including 48 male-male, 48 female-female, and 96 male-female mixtures with target-to-interference ratios (TIR) ranging from 0 dB to +18 dB. The results show our simple but effective and fast technique significantly outperforms two widely-used approaches 1.

In this paper, performance and accuracy of both General Geometric Programming (GGP) and non-linear programming (NLP) algorithms, for optimization of low power VLSI circuits, have been studied and compared. An optimization procedure based on GGP and logical effort method has been proposed and employed for optimization of variety of sequential logic circuits. The results were compared to the NLP algorithm of Sequential Quadratic Programming (SQP). Experiments showed that the GGP algorithm with Logical Effort method exhibits higher precision and acceptable speed compared to NLP algorithms. In fact, GGP is 9 orders of magnitude more accurate but 24× slower than NLP. However, with increasing circuit complexity the GGP does not degrade like NLP. Consequently, for complex circuits GGP is a good substitution for the speed of NLP algorithms and precision of simple LP (Linear Programming) algorithms, like Logical Effort.

Automotive industries try to develop their technologies to improve quality and to minimize scrapes and wastes. In the paint shop vehicle, quality control generally relies on inspection. Industries try to develop their technologies to improve quality via proactive quality control. This paper investigates the predictability of the paint thickness to reduce defects, using ANFIS. A description of the automobile paint spray process is introduced and the inputs (as effective factors in paint spray process) are identified for each thin layer on a plate. A 50×80 sheet of metal is considered as a sample. In the present paper two models of ANFIS are presented. First model shows prediction of film thickness by using 6 inputs for bell, air layers, 7 input variables for dry film thickness or final paint thickness and 6 output points for three layers. Second model is predicting of paint appearing uniformity by average and standard deviation of film thickness.

Nowadays scholars pay attention on SMEs because of more benefits in compare with Large Enterprises, such as: Value added, Innovation, Entrepreneurs and flexibility. But the other face of the coin is lack of production resources in SMEs. One of the solutions for this problem in researches has been used the potential in contractors companies, in form of out sourcing of activities and projects. In this paper we studied the production strategies in Iran' SMEs by identification and prioritizing the production factors and implementing the Goal Decision making model. Case studies consist of 5 different companies in the same field of production; approach us to analyzing the companies' situation. The paper presented the Matrix, base on results from models and case studies. This matrix suggested the different SME's situation for selected production strategies

In model-based single channel speech separation, factorial hidden Markov models (FHMM) have been successfully applied to model the mixture signal Y(t) = X(t) + V(t) in terms of trained patterns of the speech signals X(t) and V(t). Nonetheless, when the test signals are scaled versions of the trained patterns (i.e. g x X(t) and g v V(t)), the performance of FHMM degrades significantly. In this paper, we introduce a modification to FHMM, called scaled FHMM, which compensates gain difference. In this technique, first, the scale factors are expressed in terms of the target-to-interference ratio (TIR). Then, an iteration quadratic optimization approach is coupled with FHMM to estimate TIR which with the decoded HMM sequences maximize the likelihood of the mixture signal. Experimental results, conducted on 180 mixtures with TIRs from 0 to 15 dB, show that the proposed technique significantly outperforms unscaled FHMM, and scaled/unscaled vector quantization speech separation techniques.

In this paper, we apply a modified integral equation technique to analyze rectangular dielectric waveguides with elimination of spurious modes. The two stage method applies interior problem by modal analysis to access the electric field in the waveguide in terms of the electric field on the boundaries. The exterior problem can be solved by integral equation technique to calculate the boundary fields regarding to interior fields. Using this approach, the eigen value problem can be formulated by finding the zeros of the relating determinant. Eliminating the spurious modes, the zeros of each mode are distinctly distinguishable. The consequent dispersion diagram is also has good agreements with the results of other methods for rectangular dielectric waveguide.

In most current model-based single channel separation techniques, it is assumed that the recording conditions are identical in the training phase and application phase. In this paper, we consider a general case in which training data and application data have different levels of energy and a technique is proposed to estimate the sources' gains which are required for the separation process. We use the periodogram of the speech signal as the selected feature for separation such that the sources' gains are estimated in terms of normalized periodograms of the sources and the mixture. The proposed technique is compared with a state-of-the-art technique which uses AR modeling of the speech signal and maximum likelihood for estimating gain and separating the sources. Experimental results show that our technique not only outperforms this technique in terms of SNR results and gain estimation accuracy but also reduces computational complexity.

Resources limitations and requirements and representing a lot of objectives, have made inevitable the necessity of applying optimal and efficient available capabilities in each organization. For this reason, ranking of goals along with organizational goals in each organization cannot be avoided and also most of successful companies pay more attention to this point. But in achieving this objective considering goals and the decision maker conditions is mandatory. Therefore, presenting a scientific solution and appropriate model that can satisfy the relations between goals and limitations seems to be necessary. In the complicated situation, considering all components and factors in decision-making obliges us to create a balance between goals and effective factors in achieving them. In this condition, not only all of the components and environmental variables cannot be considered one dimensionally, but also the best solution is considering all conditions and goals and making a decision in order to succeed in the desired utility of them. These problems have caused the nature of decision making include so many ambiguities, from the simplest to the most complicated affairs, and caused decisions to be taken in a fuzzy environment. As fuzzy logic has several applications in making decision in uncertain and fuzzy environments, this paper intends to model one of the strict and sophisticated decisions of managers using a mathematical model based on fuzzy inference system, that is, analysis and determining ranking of goals, from the available list in which the maximum use of available resources, considering the existing limitations, can be reached.

Manufactures and service provider institutes usually need queue theory to optimize their decisions about customers' waiting time. This will help them to specify resources that should be investigated and to provide customers' satisfaction. These two factors, resource allocation and customer satisfaction are so important for companies' survival essentially in a highly competitive environment. So describing queue systems performance in different environmental conditions is necessary. In this paper by analyzing actual queuing systems with layout constraints, a framework is introduced for specific conditions by using Markov chain concepts. This model can be a base for evaluating exponential queue systems with probability parameters. In this paper by providing a case study, it is tried to describe the proposed model. To do this, filling stations are analyzed in Tehran city.

Future is less predictable, but the remarkable point is that organizations can prepare themselves to encounter it which will result in the creation of competitive advantage for them. In order to develop in the insecure environment, organizations should leave their unidimensional view toward the future, and consider the probable events in the future in their planning, using a multi-dimensional viewpoint. The methodology that is introduced in this article is a modern one because it combines the general method of formulating strategies with the two common tools that are used in order to fight uncertainty, namely, scenario planning and fuzzy logic. Using the uncertain elements that exist in the environment, this method embarks on designing possible scenarios facing the organization, and with the help of the fuzzy information supplied by the experts, it sets to select the most robust strategy of the organization in the fuzzy inference system. This strategy is selected in such a way that under all possible conditions of the occurrence of scenarios, it is justifiable and not necessarily the optimal choice; and enjoys the best possible performance from among all the scenarios that are specified for the future. In this paper, we compare the common method of strategy formulation with the formulation of strategy through the robust methodology, then, we examine the results obtained from implementation of both methods in a case study in a real organization.

Model-based single channel speech separation techniques commonly use trained patterns of the individual speakers to separate the speech signals. In most recent proposed techniques, it is assumed that data used in the train and test phase have the same level of energy, a prerequisite which is hardly met in the real situations. Considering this limitation, we propose a technique which estimates the gain associated with the individual speakers from the mixture and thus obviate the need for this assumption. The basic idea is to express the probability density function (PDF) of the mixture in terms of the individual speakers' PDFs and corresponding gains. Then, those patterns and gains which maximize the mixture's PDF are selected and used to recover the speech signals. Experimental results conducted on a wide variety of mixtures with signal-to-signal ratios ranging from 0 to 18 dB show that the proposed technique estimates the speakers' gain with 95% accuracy within the range of the actual gain ±%20. Comparing the separated speech signals with the original ones in terms of SNR criterion with/without including the gain estimation stage, we observe a significant SNR improvement (on average 5.73 dB) for the gain included scenario.

We present an approach for separating two speech signals when only one single recording of their linear mixture is available. The log spectra of the sources are estimated from the mixture's log spectrum using minimum mean square error (MMSE) approach. The estimation is obtained from the assumption that the sources are modelled using a set of Gaussian subsources which are related to the mixture using MIXMAX approximation. The resulting estimator has a closed form and is expressed using the mean and variance of Gaussian subsources. In order to obtain the two most likely subsources which generate the mixture, we use the estimation-detection technique. We also show that the binary mask filtering which has been empirically - and with no mathematical justification - used in speech separation techniques is, in fact, a simplified form of the MMSE estimator. The proposed technique is compared with the binary mask when the input consists of male-male, female-female, and female-male mixtures. The experimental results in terms of segmental SNR show that the MMSE estimator outperforms binary mask filtering.

In this paper, we present a new technique for separating two speech signals from a single recording. For this purpose, we decompose the speech signal into the excitation signal and the vocal tract function and then estimate the components from the mixed speech using a hybrid model. We first express the probability density function (PDF) of the mixed speech's log spectral vectors in terms of the PDFs of the underlying speech signal's vocal tract functions. Then, the mean vectors of PDFs of the vocal tract functions are obtained using a Maximum Likelihood estimator given the mixed signal. Finally, the estimated vocal tract function along with the extracted pitch values are used to reconstruct estimates of the individual speech signals. We compare our model with both an underdetermined blind source separation and a CASA method. The experimental results show our model outperforms both techniques in terms of SNR improvement and the percentage of crosstalk suppression.

We present a generalized approach to speaker dependent model-based single channel speech separation techniques in which a priori knowledge of the underlying speakers is used to separate speech signals. For this purpose, we add an identification stage by which we first identify the underlying speakers in the mixture and then use the identified speakers' model to separate speech signals. The proposed technique not only preserves the advantages of model-based speaker dependent single channel speech separation algorithms (i.e. high separability) but also is able to separate the speech signals of an unlimited number of speakers given the speakers' models (i.e. generality). Evaluation results conducted on a database consisting of 100 mixed speech signals with target-to-interference ratios (TIR) ranging ?9 dB to +9 dB show significant performance improvements over those techniques which use a single model for all speakers.

In this paper, a new single channel speech separation technique based on vector quantization (VQ) and the MIXMAX approximation is presented. At the core of this approach are two trained codebooks of the quantized feature vectors of speakers, whereby the main evaluation for separation is performed. The performance of the VQ-based approach is evaluated by applying three separate features: log spectrum, modulated lapped transform (MLT) coefficients, and a fusion of pitch and envelop information. The experiments are conducted in two different scenarios: speaker-dependent and speaker independent. The results show that the log spectrum outperforms the other features for speaker-dependent scenario. However, for the speaker independent scenario, the best results are obtained from applying the pitch-envelop feature.

Preprocessing steps in MADM are very important and have a critical role in ranking operations. In many ranking algorithms, before running the models, it is necessary to do some modifications. These modifications are for example, Quantification, determination of utility of criteria, dimensionless operations for attributes, weighting methods and so on. Weighting methods try to define an importance of criteria in decision making process. Changing the weight in decision making process has a great influence in ranking results. Sometimes, determination of criteria weights are so difficult with conjoint of errors. There are many methods for this operation, but in literature we can not see an integrated model for classification of these methods. Methods such as entropy, LINMAP, eigenvector, smart, swing and so on, are the main weighting methods in this area. In this paper, our viewpoint for classification is multidimensional. To be or not to be decision making matrix, decision maker role in criteria weighting, number of criterion in weighting operation, and something like that. What is the weight for this criterion in decision making process? To answer, we must try to mine data from two different resources, first considering construction and layout of data in decision making matrix and second from opinion and perception of decision maker. We can see two main problems, first lack off the role of decision maker to finding weights by use decision making matrix and second educe of data from decision maker's opinion without attention to reliability and validity of data is very difficult. This paper tries to introduce a new approach for classification of current weighting methods and make a logical construction for this purpose. At the end, a new combinatorial model for reduce error and produce reliable and valid data for decision making is presented

The two main questions in this research are: 1) which biotechnology transfer methods relating to biotechnology have been used by the public institutes and private companies in Iran? and 2) how much successful were these methods? Regarding the importance of the issue, the method of implementing the research is based on the explanation and measurement. In this research the statistical community includes all the biotechnology projects in the commercial production scale. The research findings indicated that the most of the projects have chosen the modern technology transfer methods and processes, but their conformity, absorption, application, development, and promotion were respectively incomplete and imperfect. This finding differs from the complete technology transfer which could be successful. Perfect success of technology transfer in general and biotechnology in particular, is achieved when it leads to a complete transfer of the imported technology. Therefore, regarding the research findings and the comprehensive definition of biotechnology transfer, it is concluded that the success level of biotechnology transfer (based on the applied technology transfer methods) in the statistical community is incomplete.

In this paper, we study the effect of window size on the performance of model-based single channel speech separation techniques. The separation system consists of two trained codebooks of the log spectral vectors of each speaker where by the main process of separation is performed. Four Hamming windows with the durations of 20, 32, 64, and 128 msec are used for short-time processing of the data. The experimental results show, in contrast with other speech processing applications, that a longer window must be applied for the separation purpose. Improvements in SNR of 2-3 dB are reported by using an optimal window selection which can significantly improve the performance of current model-based speech separation techniques.