Relative Performance Research Articles

Evaluation of Seed Yield Stability of Lentil Genotypes by Linear Mixed‐Effects Models and Multitrait Stability Index

ABSTRACTTo evaluate performance stability for 13 lentil genotypes, experiments were performed at three locations, Ardebil, Zanjan, and Maragheh, during three growing seasons (2011–2013). Environment, genotype, and interaction of genotype × environment were significantly effective on 100‐seed weight, plant height, number of days to flowering, number of days to maturity, rate of seed filling, and single seed weight. Nominal grain yield biplot identified stability of 2, 12, 1, 3, and 6 genotypes with regard to grain yield. According to biplot analysis, Genotypes 1, 3, and 10 not only produced the highest grain yield but also showed the greatest stability in yield production. Genotypes 1, 6, 7, and 13 were identified as highly stable and productive based on grain yield, mean weight stability index, and weighted average of absolute scores of best linear unbiased predictions (WAASB). Genotypes 1 and 7 were best genotypes according to multitrait stability index (MTSI). Harmonic mean of the relative performance of genotypic predicted value (HMRPGV) recognized Genotypes 6, 1, 5, and 3 as highly stable, productive, and adaptable genotypes. In general, Genotype 1 indicated highest grain yield and desirable agronomic traits compared to other genotypes tested, suggesting this genotype as new cultivar.

Lean Blowout Proximity Sensing in a Low Emission Liquid-Fueled Combustor

ABSTRACT Advanced aircraft engine combustor designs, intended for achieving low NOx emissions, necessitate lean combustion with small operating margins above the LBO boundary, posing a high risk of exhibiting lean blowout (LBO) problems. Employing active LBO controllers receiving input from LBO proximity sensors can allow for narrower margins, thereby yielding multiple performance gains. Although several previous studies investigated LBO proximity sensing, they predominantly focused on premixed gas-fueled combustors. The application of proximity sensing approaches to liquid-fueled combustor designs, relevant to aircraft engines have not received adequate attention. Therefore, this study investigates LBO proximity sensing in a low emission lean direct injection (LDI) liquid-fueled combustor design developed by NASA. The combustor operates at atmospheric pressure with preheated air. The proximity sensing is examined by analyzing both OH* and CH* chemiluminescence signals, with a focus on assessing their relative performances. The signal analysis approaches include spectral, partial extinction event (LBO precursor) detection, signal RMS, and the CH*/OH* ratio, with an emphasis on the first two methods. Both steady-state and transient operating conditions have been examined. The combustor did not exhibit large-scale flame partial extinction and re-ignition events near LBO, as commonly observed in gas-fueled combustors. However, an increase in flame unsteadiness and partial extinction events with very subtle signatures in optical signals, have been observed. Low-pass filtering has been observed to considerably enhance the precursor signatures enabling their improved detection. With a reduction in LBO margin, an increase in signal low-frequency power, RMS, and precursor event occurrence rate have been observed, which were utilized for providing LBO proximity measures. A key finding from this study is that CH* signals provide significantly superior performance compared to OH* for the proximity sensing in all the approaches, and the potential reasons are discussed.

Reduced benefit from long-term item frequency contributes to short-term memory deficits in dyslexia.

Dyslexia, a specific difficulty in acquiring proficient reading, is also characterized by reduced short-term memory (STM) capacity. Extensive research indicates that individuals with developmental dyslexia (IDDs) benefit less from exposure, and this hampers their long-term knowledge accumulation. It is well established that long-term knowledge has a great effect on performance in STM tasks, and thus IDDs' reduced benefit of exposure could potentially reduce their relative performance in such tasks, especially when frequent items, such as digit-words, are used. In this study we used a standard, widely used, STM assessment: the Digit Span subtest from the Wechsler Adult Intelligence Scale. The task was conducted twice: in native language and in second language. As exposure to native language is greater than exposure to second language, we predicted that IDDs' performance in the task administered in native language will reveal a larger group difference as compared to second language, due to IDDs' reduced benefit of item frequency. The prediction was confirmed, in line with the hypothesis that reduced STM in dyslexia to a large extent reflects reduced benefits from long-term item frequency and not a reduced STM per se.

Leveraging Item Parameter Drift to Assess Transfer Effects in Vocabulary Learning

ABSTRACT Longitudinal models typically emphasize between-person predictors of change but ignore how growth varies within persons because each person contributes only one data point at each time. In contrast, modeling growth with multi-item assessments allows evaluation of how relative item performance may shift over time. While traditionally viewed as a nuisance under the label of “item parameter drift” (IPD), IPD may be of substantive interest if it reflects how learning manifests on different items or subscales at different rates. In this study, we apply the Explanatory Item Response Model to assess IPD in a causal inference context. Simulation results show that when IPD is not accounted for, both parameter estimates and standard errors can be affected. We illustrate with an empirical application to the persistence of transfer effects from a content literacy intervention , revealing how researchers can leverage IPD to achieve a more fine-grained understanding of how vocabulary learning develops over time.

The predicting role of sustainable supplier selection on lead-time performance in public procurement: relational capability as a moderator

PurposeWith increasing concerns regarding the performance of public procurement, especially concerning timely delivery, this study aims to investigate the predictive role of sustainable supplier selection (SSS) in enhancing lead-time performance in public procurement, while also examining the moderating effect of relational capability.Design/methodology/approachThe study collected primary data from 179 accredited public organizations in Tanzania through structured questionnaires and used partial least squares structural equation modeling for data analysis.FindingsThe results reveal that SSS and relational capability influence lead-time performance in public procurement. Regarding the moderating effect of relational capability, the study establishes that relational capability strengthens the influence of SSS on lead-time performance. Therefore, based on the collected data from surveyed public organizations, relational capability positively moderates the relationship between SSS and lead-time performance in public procurement.Social implicationsThe study offer valuable implications for public procurement practitioners, academics and other stakeholders. Public organizations should consider contemplate integrating sustainability aspects into supplier selection processes. Furthermore, the study contributes to advancing efforts toward achieving sustainable development goals (i.e. SDG12) within Tanzania, particularly by advocating for sustainable public procurement practices aligned with national policies and priorities.Originality/valueThis study not only enhances the understanding of public procurement performance by investigating lead-time improvement, incorporating considerations of SSS and relational capability, but it also offers empirical evidence of the moderating role of relational capability within the unexplored Tanzanian context.

Relative Performance of Rice Varieties under Drip Irrigated Dry Direct-seeded and Flood Irrigated Transplanted Conditions

A field experiment was conducted to evaluate the performance of 3 different rice varieties (KNM-1638, JGL-24423 & DRR Dhan-42) under drip irrigated dry direct-seeded condition as well as conventionally flooded transplanted condition with 100 & 125% N. The experiment was performed in split-plot design with 3 varieties in the main plot and 2 levels of irrigation (1.0 Epan & 1.5 Epan) in combination with 2 doses of N (100 & 125% of RDF) in the sub-plot. An observation trial with these 3 varieties under transplanted condition with 100 & 125% N was taken up. The study was conducted during kharif seasons of 2022-23 and 2023-24 at College farm, College of Agriculture, Rajendranangar, Hyderabad, Telangana, India. In each variety, best performing irrigation level was found at 100% N & 125% N. Growth parameters, yield parameters & yield of best-performing treatments under dry direct-seeded (DDSR) condition was compared with conventionally flooded transplanted rice (CFTPR) with the same variety and same dose of N using one sample t-test. KNM-1638 recorded significantly higher yield at 5% level of significance under transplanted condition compared to dry direct-seeded condition both at 100% N (P=0.01) & 125% N (P=0.01). Grain yield was statistically comparable at 5% level of significance under dry direct-seeded and transplanted condition for JGL-24423 at 100% N (P= 0.18). However, at 125% N, transplanted rice recorded superior yields to that of direct-seeded rice at 5% level of significance (P = 0.01). DRR Dhan-42 registered comparable yields under direct-seeded and transplanted condition with 100 (P = 0.22) & 125% N (P = 0.07) at 5% level of significance. There observed 57-68% of saving in irrigation water in DDSR compared to CFTPR. Based on yield, economization of water and nitrogen, KNM-1638 is recommended for cultivation under transplanted condition with 100% N; whereas, JGL-24423 & DRR Dhan-42 can be successfully cultivated under dry direct-seeded condition with 100% N without yield penalty.

Intellectual Capital and Performance of Staff at Selected Institutions of Higher Learning in Uganda

Purpose: The study examined the relationship between intellectual capital and performance of staff at selected institutions in Uganda. The specific objectives were; i) To establish the relationship between human capital and performance of staff at selected institutions of higher learning in Uganda; ii) To establish the relationship between relational capital and performance of staff in selected institutions of higher learning in Uganda and iii) To establish the relationship between structural capital and performance of staff in selected institutions of higher learning in Uganda. Methodology: A cross-sectional design was used alongside a quantitative approach whereby a questionnaire was employed to collect the data from the 103 respondents who included lecturers from 10 selected institutions of higher learning. The results were computed using multiple regression and correlation analysis using SPSS and upon analysis, Findings: The outcomes exposed that; a) human capital has a positive but significant relationship with performance of staff at selected institutions of higher learning in Uganda (r = .536**, p<0.0001); b) there is a moderate relationship between relational capital and performance of staff at selected institutions of higher learning in Uganda (r = .374**, p>0.0001); c) there is a positive and very significant relationship between structural capital and performance of staff at selected institutions of higher learning in Uganda (r = .586**, p<0.0001). Unique Contribution to Theory, Policy and Practice: The research outcome undertakes that private institutions do not give attention to developing relational capital although it’s a very important aspect. It is also imperative to note that more resources are invested in the structural capital given its serious and observable impact on performance if it is put to utmost utilization. The study empirically differentiates the impacts of human, relational and structural capital on staff performance in institutions of higher learning offering context-specific insights and practically underscores the investment in structural and relational capital and continuous human capital development guiding resource allocation and policy formulation to enhance staff performance in Uganda.

Detection of Sarcopenia in a Community-Dwelling Older Population in China

Background and Purpose: Sarcopenia is a common muscle disease among the older population, posing an increased risk for functional decline and leading to loss of independence in daily living. Early detection of sarcopenia among older people before functional decline would be beneficial in enhancing their quality of life. The Asian Working Group for Sarcopenia (AWGS) 2019 recommends the use of 3 screening methods for community-based sarcopenia detection: calf circumference (CC), or the Strength, Assistance in walking, Rise from a chair, Climb stairs, and Falls (SARC-F) questionnaire, or the SARC-F in combination with CC (SARC-CalF) questionnaire. This study aims to compare the relative performance of these 3 methods as screening tools for sarcopenia within a community-dwelling older population. Methods: A total number of 700 community-dwelling older adults participated in the current study. Muscle mass, muscle strength, and physical performance were measured with bioelectrical impedance analysis, handgrip strength, and gait speed, respectively. The AWGS 2019 criteria were considered the criterion standard. The sensitivity/specificity, receiver operating characteristic (ROC) curve, and area under the receiver operating characteristic curve (AUROC) analyses were determined for CC, SARC-F, and SARC-CalF to determine their relative diagnostic performance. Results: Sarcopenia was identified in 21.4% of participants according to the AWGS2019 criteria. The overall prevalence of sarcopenia was 56.6%, 14.7%, and 22.9% according to CC, SARC-F, and SARC-CalF, respectively. Calf circumference showed the highest sensitivity but lowest specificity based on AWGS 2019 as the gold standard regardless of age, gender, and body mass index. The SARC-CalF showed better sensitivity but similar specificity than the SARC-F. The AUROC of CC was significantly better than that of SARC-F and SARC-CalF. The AUROCs of CC, SARC-F, and SARC-CalF were statistically significant in all populations, as well as in the categories of age, gender, and body mass index (P < .05). Conclusions: Calf circumference is useful in ruling out the presence of sarcopenia while the SARC-F is more effective in ruling in sarcopenia, especially in the context of population-based screening. Future studies should be carried out to investigate the value of population-based sarcopenia detection using these screening tools.

Latency-Aware Neural Architecture Performance Predictor With Query-to-Tier Technique

Neural Architecture Search (NAS) is a powerful tool for automating effective image and video processing DNN designing. The ranking of the accuracy has been advocated to design an efficient performance predictor for NAS. The previous contrastive method solves the ranking problem by comparing pairs of architectures and predicting their relative performance. However, it only focuses on the rankings between the two involved architectures and neglects the overall quality distributions of the search space, which may suffer generalization issues. On the contrary, we propose to let the performance predictor concentrate on the global quality level of specific architecture, and learn the tier embeddings of the whole search space automatically with learnable queries. The proposed method, dubbed as Neural Architecture Ranker with Query-to-Tier technique (NARQ2T), explores the quality tiers of the search space globally and classifies each individual to the tier they belong to. Thus, the predictor gains knowledge of the performance distributions of the search space which helps to generalize its ranking ability to the datasets more easily. Thanks to the encoder-decoder design, our method is able to predict the latency of the searched model without deteriorating the performance prediction. Meanwhile, the global quality distribution facilitates the search phase by directly sampling candidates according to the statistics of quality tiers, which is free of training a search algorithm, e.g., Reinforcement Learning or Evolutionary Algorithm, thus it simplifies the NAS pipeline and saves the computational overheads. The proposed NARQ2T achieves state-of-the-art performance on two widely used datasets for NAS research. Moreover, extensive experiments have validated the efficacy of the designed method.

Performance research of edge computing nodes based on heterogeneous environment

Abstract Edge computing realizes real-time parallel storage and computing in a distributed way, which can reduce the data transmission cost. The edge nodes, which are composed of different heterogeneous physical devices and hardware resources, assume the computing and storage capacity of the edge network. At present, the resource characteristics and performance of heterogeneous edge nodes are difficult to characterize. Therefore, a resource-aware performance feature detection model for heterogeneous edge compute nodes is proposed in this study. This model is based on a resource utilization matrix and resource relative performance quantization algorithm to extract the performance characteristics of the same type of resources of the edge nodes, to realize multi-dimensional and systematic feature detection and description of heterogeneous edge nodes. On this basis, a weight-based edge node task scheduling algorithm is designed. The final verification results show that the CPU relative performance ratios of Node 1-Node 2, Node 1-Node 3, and Node 2-Node 3 are 6.23, 3.81, and 0.81 respectively. The performance relationship among the three nodes is Node 1> Node 3> Node 2. The test group and the verification group tend to be consistent overall. This method has a good application prospect.

The Effect of Horizontal Pay Transparency on Employee Motivation When Pay Dispersion Is Performance Based and Non-Performance Based

ABSTRACT The movement toward pay transparency within firms has intensified. We investigate the effect of horizontal pay transparency on employee motivation in a broader information environment where (1) performance-based and non-performance-based pay dispersion coexist and (2) relative performance information allows employees to alleviate pay dispersion ambiguity in order to assess the basis of any pay dispersion that pay transparency reveals. Drawing on distributive justice theory, we predict and find that pay transparency differentially affects employees’ motivation based on what it reveals about their relative standing while working under the same performance incentives. Specifically, it depends on whether they are underpaid, overpaid, or their relative pay aligns with their relative performance. Our analysis provides additional insights related to these three scenarios, and in doing so, our study highlights the importance of considering employees’ broader information environment when considering the potential benefits and costs associated with pay transparency. Data Availability: Data are available from the authors upon request.

Deep Learning-Driven Compiler Enhancements for Efficient Matrix Multiplication

Matrix multiplication is a fundamental operation in many computational fields, requiring optimization to handle increasing data sizes efficiently. In this paper, the implementation of Deep Learning in Matrix multiplication is reviewed, which is considered important nowadays due to the growing complexity of matrix multiplication for gaming and complex programs. The current standard matrix multiplication and the time taken by it on different matrix sizes are described. The Tiled Matrix multiplication, which trims the matrix into various pieces and calculates the product for each piece, and thereafter combines the result, is also described. The times taken by both methods for different matrix sizes were compared. The main idea was to use Deep Neural Networks (DNN) to compare and rank code variants that are obtained in pieces and determine their relative performance. A tournament-based ranking system is used for assigning ranks to the code versions. The effectiveness of these techniques was evaluated on various matrix multiplication operations commonly found in deep learning workloads. Up to 8.844x speedup over the naive implementation for a matrix size of 1024 is achieved by this approach. The results demonstrate the effectiveness of combining compiler optimization techniques and deep learning models in optimizing matrix multiplication.

WITHDRAWN: A Technique for Enhancing Respiration Sensing in Interference Scenarios with Wi-Fi Signal

Wi-Fi-based non-contact respiratory monitoring technology plays a significant role in smart elderly care by eliminating the need for users to wear additional hardware devices. Although respiration sensing has shown impressive accuracy in ideal scenarios, achieving precise respiration monitoring in interfering scenarios, such as when other users are engaged in activities, remains challenging. This difficulty arises primarily from the weak nature of the respiratory signal, which is susceptible to interference from activities. To address this issue, this paper introduces the RespEnh system, a respiratory monitoring system that utilizes human location information to mitigate interference. Firstly, we enhance the quality of respiratory signal by leveraging the frequency diversity of Wi-Fi signal, effectively visualizing respiratory patterns even amidst interference. Secondly, we apply short time window shift to remove walking noise in the time domain while preserving respiratory signal characteristics. Finally, we integrate multi-antenna signals to enhance spatial domain performance. Experimental results demonstrate that our system achieves a relative anti-interference performance of 2.4m, even when breathing as far away as 6.4m. It effectively monitors person's respiration, with a detection rate of 80% within an error range of 1bpm.

Workplace stressor factors, profiles and the relationship to career stage in UK veterinarians, veterinary nurses and students.

Veterinary professionals experience higher psychological distress and lower wellbeing compared with the general population. Identifying workplace stressors is key to understanding and alleviating these difficulties. Identify the severity of workplace stressors in veterinary professions across different levels of professional experience. A cross-sectional quantitative design was utilised via administration of a veterinary stressor questionnaire that measured the severity of 93 generic and veterinary-specific workplace stressors. A total of 658 participants reported stressor severity score. Factor analysis revealed four main stressor categories: workload and job demands, client relations stress, performance and accountability, patient care challenges. An inverse relationship was found between experience level and stressor severity scores. Furthermore, the most severe stressors varied as a function of career stage. Stressor severity tends to decrease with increasing experience levels. Stressors can be categorised by theme and unique stressor profiles can be derived for practitioners according to experience level. This knowledge can assist in the design and delivery of workplace support initiatives. Identify key stressors across different levels of experience is important in focusing efforts to support people in the workplace.

Comparing standard‐ and reduced‐size passive integrated transponder (PIT) tags for monitoring juvenile wild spring Chinook Salmon

AbstractObjectivesAnnual migration monitoring can help to discern patterns and environmental factors that impact growth, survival, and movement timing in small fish. Mark–recapture methods form the basis for such monitoring, and the standard 12‐mm passive integrated transponder (PIT) tag has emerged as an essential tool for studies of juvenile salmonids. A smaller, 9‐mm PIT tag now provides the potential to conduct mark–recapture studies on smaller fish. We evaluated relative performance of the 9‐mm tag, which is similar in design to its 12‐mm predecessor.MethodsFor this comparison, we tagged and released approximately 8400 wild spring Chinook Salmon Oncorhynchus tshawytscha parr in Valley Creek, Idaho, from 2011 to 2013. Tag‐size cohorts were of similar average body size and were tagged in equal numbers. We estimated survival and detection probability for each cohort over two river segments.ResultsIn both segments, survival varied among years, but we observed no significant differences between tag‐size groups. At Valley Creek, average detection rates of fish with 9‐mm tags were a little less than one‐half the rates of fish with 12‐mm tags and were significantly lower in all 3 years. At Lower Granite Dam, detection rates were again lower for 9‐mm tags, but the differences were much smaller (3%–12%) and were not statistically significant.ConclusionWe found that 9‐mm tags can be as effective as 12‐mm tags and may allow for better inference to smaller (<55‐mm) non‐tagged fish. However, the lower detection rates of the 9‐mm tags could lead to less precise estimates, and site‐specific detection rates should be considered for studies that rely on these tags.

Performance measurement using intuitionistic fuzzy sets: A case study from debt collection centers

Performance measurement (PM) is a vital aspect of management, conducted periodically to assess the success or quality of a specific process or an organization. Performance encompasses both goals and relational models crucial for achieving those goals within a given time frame. From another perspective, performance measurement is about determining related performance indicators and providing a numerical summarized result which can be used for comparisons. Converting various indicators to an overall performance score can be modeled as a multi-criteria decision making problem. In this study, we focus on performance measurement of debt collection centers which are organizations specialized in collecting money from debtors. To this end, a decision model is developed and intuitionistic fuzzy Analytic Hierarchy Process is used to obtain the criteria weights. The proposed framework is shown in a real-world case study.

Global efficiency and profitability: Cooperatives as social innovation agents vs. Joint stock companies in the agri-food sector

Cooperativism, as an agent of social innovation, significantly contributes to fostering a dynamic economy and plays a fundamental role in rural areas. Analyses of profitability in cooperatives versus Investor-Owned Firms (IOFs) show disparate results depending on the methodology and indicators used. This disparity of results is largely due to the fact that common indicators are used in the business sphere that do not take into account the duality of investor-supplier members in cooperatives. This paper deals with a comparative analysis of the efficiency and profitability of cooperatives and IOFs, which takes into account the cooperative specificity as far as partner-producer duality is concerned. The populations analyzed are the olive oil producing companies in Andalusia and the methodology used is twofold: DEA, in the case of efficiency, and probit regression for the analysis of profitability. For the latter, a cooperative global profitability indicator has been defined. The results point to a superiority of the cooperatives in terms of both efficiency and overall profitability compared to the IOFs, which denotes a greater capacity to better remunerate the producer, especially in smaller companies. However, this superiority disappears as the cooperative gains dimension. Our findings show that the relative performance of cooperatives and IOFs varies by size. This reality warrants the attention of both society and governments, prompting the need for the development and implementation of public policies aimed at their growth and safeguarding. Additionally, it initiates a pathway for further research into other sectors of activity.

Capitalizing on Implementation Science to Advance Antimicrobial Stewardship and Health Equity in Treating Pediatric UTI’s

Background: Pediatric urinary tract infections represent the most common pediatric infection with increasing gram-negative antibiotic resistance. Overutilization of antimicrobials including third generation cephalosporins are known drivers of this resistance. Antimicrobial stewardship (AS) efforts have recently shown that antibiotic selection may be influenced by patient race. Implementation science (IS) can provide frameworks and strategies to improve antimicrobial utilization and equity. Methods: This was a pre/post study of 2 geographically different children’s hospitals general pediatric floors assessing the impact of a set of implementation strategies developed to improve provider knowledge of best practice antimicrobials (based on local susceptibilities for treatment of UTI’s) and influence uptake of best practice prescribing. IS strategies included provider education, local clinical champion and opinion leader involvement, leadership involvement, local policy changes, and stakeholder co-design of decision support tools (a clinical pathway, a specific ITI antibiogram, and a dynamic order set). No education was provided regarding racial differences in prescribing habits. Outcomes were measured utilizing a portion of the RE-AIM Framework of assessing adoption of “Right” antibiotic, order set adoption, and equitable reach (racial differences in prescribing). Results: Hospital A and B had a first-generation cephalosporin prescription rate of 29.7% (n=441) and 20.6% (n-557) pre implementation and 44.6% (n=84) and 47.5% (n=118) post (p < 0 .001). Both hospitals also saw a significant reduction in third-generation cephalosporins. In Hospital A, APRN’s were more likely to prescribe a first generation cephalosporin (52.4%) than a DO (42.1%) or MD (26.4%) pre-implementation (p=0.004). In Hospital B, APRN’s were less likely to prescribe first generation cephalosporins (5.4%) than a DO (28.9%) or MD (19.4%) pre-implementation (p=0.004). No statistical significance was seen post implementation for antimicrobial selection by provider type for either hospital. Based on race, both hospitals had Black and Other patients receiving more first-generation cephalosporins while white patients were more likely to receive third-generation cephalosporins (p=0.033) pre implementation. No statistical significance was seen post implementation for antimicrobial selection based on race. No improvement was seen in order set utilization. Conclusion: With order set utilization not improving with implementation of new dynamic order set, other strategies such as education, clinical champion and opinion leader involvement, and provision of a local UTI antibiogram were likely contributors to the improvement in best antimicrobial for treatment of UTI’s. Further mixed method research is warranted to improve understanding of the relative performance of our strategies, especially the lack of provider adoption of the novel dynamic order set.Disclosure: Jessica Snowden: Advisory board - Pfizer

Interband cascade light-emitting diodes grown on silicon substrates using GaSb buffer layer

Interband cascade light-emitting diodes (ICLEDs) offer attractive advantages for infrared applications, which would greatly expand if high-quality growth on silicon substrates could be achieved. This work describes the formation of threading dislocations in ICLEDs grown monolithically on GaSb-on-Silicon wafers. The epitaxial growth is done in two stages: the GaSb-on-Silicon buffer is grown first, followed by the ICLED growth. The buffer growth involves the nucleation of a 10-nm-thick AlSb buffer layer on the silicon surface, followed by the GaSb growth. The AlSb nucleation layer promotes the formation of 90° and 60° interfacial misfit dislocations, resulting in a highly planar morphology for subsequent GaSb growth that is almost 100% relaxed. The resulting GaSb buffer for growth of the ICLED has a threading dislocation density of ∼107/cm2 after ∼3 μm of growth. The fabricated LEDs showed variations in device performance, with some devices demonstrating comparable light–current–voltage curves to those for devices grown on GaSb substrates, while other devices showed somewhat reduced relative performance. Cross-sectional transmission electron microscopy observations of the inferior diodes indicated that the multiplication of threading dislocations in the active region had most likely caused the increased leakage current and lower output power. Enhanced defect filter layers on the GaSb/Si substrates should provide more consistent diode performance and a viable future growth approach for antimonide-based ICLEDs and other infrared devices.

Comparing human and machine's use of coarticulatory vowel nasalization for linguistic classificationa).

Anticipatory coarticulation is a highly informative cue to upcoming linguistic information: listeners can identify that the word is ben and not bed by hearing the vowel alone. The present study compares the relative performances of human listeners and a self-supervised pre-trained speech model (wav2vec 2.0) in the use of nasal coarticulation to classify vowels. Stimuli consisted of nasalized (from CVN words) and non-nasalized (from CVCs) American English vowels produced by 60 humans and generated in 36 TTS voices. wav2vec 2.0 performance is similar to human listener performance, in aggregate. Broken down by vowel type: both wav2vec 2.0 and listeners perform higher for non-nasalized vowels produced naturally by humans. However, wav2vec 2.0 shows higher correct classification performance for nasalized vowels, than for non-nasalized vowels, for TTS voices. Speaker-level patterns reveal that listeners' use of coarticulation is highly variable across talkers. wav2vec 2.0 also shows cross-talker variability in performance. Analyses also reveal differences in the use of multiple acoustic cues in nasalized vowel classifications across listeners and the wav2vec 2.0. Findings have implications for understanding how coarticulatory variation is used in speech perception. Results also can provide insight into how neural systems learn to attend to the unique acoustic features of coarticulation.