Flexible piezoelectric acoustic sensors have been developed to generate multiple sound signals with high sensitivity, shifting the paradigm of future voice technologies. Speech recognition based on ...advanced acoustic sensors and optimized machine learning software will play an innovative interface for artificial intelligence (AI) services. Collaboration and novel approaches between both smart sensors and speech algorithms should be attempted to realize a hyperconnected society, which can offer personalized services such as biometric authentication, AI secretaries, and home appliances. Here, representative developments in speech recognition are reviewed in terms of flexible piezoelectric materials, self‐powered sensors, machine learning algorithms, and speaker recognition.
Flexible piezoelectric acoustic sensors and machine learning for speech processing can change the paradigm of voice technologies for the hyperconnected society, offering personalized intelligent services such as biometric authentication, AI secretaries, and Internet‐of‐Things (IoT) appliances. The recent advances in the fields of self‐powered flexible acoustic sensors and machine learning algorithms for speech recognition are comprehensively summarized.
Self-supervised learning (SSL) has gained remarkable success, for which contrastive learning (CL) plays a key role. However, the recent development of new non-CL frameworks has achieved comparable or ...better performance with high improvement potential, prompting researchers to enhance these frameworks further. Assimilating CL into non-CL frameworks has been thought to be beneficial, but empirical evidence indicates no visible improvements. In view of that, this paper proposes a strategy of performing CL along the dimensional direction instead of along the batch direction as done in conventional contrastive learning, named Dimensional Contrastive Learning (DimCL). DimCL aims to enhance the feature diversity, and it can serve as a regularizer to prior SSL frameworks. DimCL has been found to be effective, and the hardness-aware property is identified as a critical reason for its success. Extensive experimental results reveal that assimilating DimCL into SSL frameworks leads to performance improvement by a non-trivial margin on various datasets and backbone architectures.
Self-supervised learning (SSL) has emerged as a promising approach for learning representations from unlabeled data. Momentum-based contrastive frameworks such as MoCo-v3 have shown remarkable ...success among the many SSL methods proposed in recent years. However, a significant gap in encoder representation exists between the online encoder (student) and the momentum encoder (teacher) in these frameworks, limiting the performance on downstream tasks. We identify this gap as a bottleneck often overlooked in existing frameworks and propose "residual momentum" that explicitly reduces the gap during training to encourage the student to learn representations closer to the teacher's. We also reveal that a similar technique, knowledge distillation (KD), to reduce the distribution gap with cross-entropy-based loss in supervised learning is useless in the SSL context and demonstrate that the intra-representation gap measured by cosine similarity is crucial for EMA-based SSLs. Extensive experiments on different benchmark datasets and architectures demonstrate the superiority of our method compared to state-of-the-art contrastive learning baselines. Specifically, our method outperforms MoCo-v3 0.7% top-1 in ImageNet, 2.82% on CIFAR-100, 1.8% AP, and 3.0% AP75 on VOC detection pre-trained on the COCO dataset; it also improves DenseCL with 0.5% AP (800ep) and 0.6% AP75 (1600ep). Our work highlights the importance of reducing the teacher-student intra-gap in momentum-based contrastive learning frameworks and provides a practical solution for improving the quality of learned representations.
The optimal design of electric machines requires a significant amount of computation time, since the design is usually performed based on finite-element analysis (FEA) for better accuracy. For this ...reason, the development of an optimization algorithm is regarded as a prerequisite to reduce the number of objective function call, which is directly related to computation time. The local search algorithm does not guarantee that the global optimum solution can be searched effectively. Therefore, in this paper, we introduce the novel optimization method, search region management (SRM), to improve the efficiency of the local search algorithm. SRM uses a guided random bit generation based on Gray code for the region selection when the local search iterates the optimum search process for a multi-start. In addition, the SRM saves and manages the information of previous searched points to prevent overlapping calculation and to increase the effectiveness of the local search during every iteration process. To clarify the effectiveness of the proposed method for electric machine design, it is applied to the optimal design of brushless dc motor coupled with FEA minimizing torque ripple.
Herein, we introduce "Look and Diagnose" (LAD), a hybrid deep learning-based system that aims to support doctors in the medical field for diagnosing effectively the Benign Paroxysmal Positional ...Vertigo (BPPV) disorder. Given the body postures of the patient in the Dix-Hallpike and lateral head turns test, the visual information of both eyes is captured and fed into LAD for analyzing and classifying into one of six possible disorders which the patient might be suffering from. The proposed system consists of two streams: (1) an RNN-based stream that takes raw RGB images of both eyes to extract visual features and optical flow of each eye followed by ternary classification to determine left/right posterior canal (PC) or other; and (2) pupil detector stream that detects the pupil when it is classified as Non-PC and classifies the direction and strength of the beating to categorize the Non-PC types into the remaining four classes: Geotropic BPPV (left and right) and Apogeotropic BPPV (left and right). Experimental results show that with the given body postures of the patient, the system is capable of accurately classifying given BPPV disorder into the six types of disorder with an accuracy of 91% on the validation set. The proposed method can successfully classify disorders with an accuracy of 93% for the Posterior Canal disorder and 95% for the Geotropic and Apogeotropic disorder, paving a potential direction for research with the medical data.
Background: Intraventricular haemorrhage (IVH) patients with acute obstructive hydrocephalus (AOH) who require external ventricular drainage (EVD) are at high risk for poor outcomes. Intraventricular ...fibrinolysis (IVF) with low-dose recombinant tissue plasminogen activator (rtPA) can be used to improve patient outcomes. Here, we evaluated the impact of IVF on the risk of death and the functional outcomes in IVH patients with AOH. Methods: This prospective cohort study included IVH patients with hypertensive intracranial haemorrhage complicated by AOH who required EVD. We evaluated the risk of death and the functional outcomes at 1 and 3 months, with a specific focus on the impact of combined EVD with IVF by low-dose rtPA. Results: Between November 30, 2011 and December 30, 2014, 80 patients were included. Forty-five patients were treated with EVD alone (EVD group) and 35 received IVF (EVD+IVF group). The 30- and 90-day mortality rates were lower in the EVD+IVF group than in the EVD group (42.2 vs. 11.4%, p = 0.003, and 62.2 vs. 20%, p < 0.001, respectively). The Graeb scores were significantly lower in the EVD+IVF group than in the EVD group (p ≤ 0.001) during the first 3 days and on day 7 after assignment. The 30-day good functional outcome (modified Rankin Scale mRS score 0–3) was also higher in the EVD+IVF group than in the EVD group (6.7 vs. 28.6%, p = 0.008). However, the 90-day good functional outcome (mRS score 0–3) did not significantly increase in the EVD+IVF group (30.8% in the EVD group vs. 51.6% in the EVD+IVF group, p = 0.112). Conclusions: In our prospective observational study, EVD+IVF was associated with a lower risk of death in IVH patients. EVD+IVF improved the chance of having a good functional outcome at 1 month; however, this result was no longer observed at 3 months.
To date, the role of bridging intravenous thrombolysis before mechanical thrombectomy (MTE) is controversial but still recommended in eligible patients. Different doses of intravenous alteplase have ...been used for treating patients with acute ischemic stroke from large-vessel occlusion (LVO-AIS) in Asia, largely due to variations in the risks for intracerebral hemorrhage (ICH) and treatment affordability. Uncertainty exists over the potential benefits of treating low-dose alteplase, as opposed to standard-dose alteplase, prior to MTE among patients with LVO-AIS.
The aim of the study was to compare outcomes of low- vs. standard-dose of bridging intravenous alteplase before MTE among LVO-AIS patients.
We performed a retrospective analysis of LVO-AIS patients who were treated with either 0.6 mg/kg or 0.9 mg/kg alteplase prior to MTE at a stroke center in Northern Vietnam. Multivariable logistic regression models, accounting for potential confounding factors including comorbidities and clinical factors (e.g., stroke severity), were used to compare the outcomes between the two groups. Our primary outcome was functional independence at 90 days following stroke (modified Rankin score; mRS ≤ 2). Secondary outcomes included any ICH incidence, early neurological improvement, recanalization rate, and 90-day mortality.
We analyzed data of 107 patients receiving bridging therapy, including 73 with low-dose and 34 with standard-dose alteplase before MTE. There were no statistically significant differences between the two groups in functional independence at 90 days (adjusted OR 1.02, 95% CI 0.29-3.52) after accounting for potential confounding factors. Compared to the standard-dose group, patients with low-dose alteplase before MTE had similar rates of successful recanalization, early neurological improvement, 90-day mortality, and ICH complications.
In the present study, patients with low-dose alteplase before MTE were found to achieve comparable clinical outcomes compared to those receiving standard-dose alteplase bridging with MTE. The findings suggest potential benefits of low-dose alteplase in bridging therapy for Asian populations, but this needs to be confirmed by further clinical trials.
Gender influences participation in food value chains (VCs) with implications for VC upgrading. This study investigated roles as well as differences in production activities, awareness, training, and ...attitudes between men and women in Vietnam's smallholder pig VCs. Data were gathered from a survey of 1,014 actors in different nodes along the chain, and the results showed that both men and women participated in all nodes of the VCs. Women were mainly in charge of routine husbandry activities (e.g., preparing feed, feeding animals, and cleaning pig pens) and participated in input supply (34.7%), pig production (60.2%), pork processing (63.6%), retailing (93.1%), and home preparation and cooking (100%). Men were more often responsible for tasks requiring strength, knowledge, and skills (e.g., disease management) and had greater involvement in larger-scale farming (60–80%) and slaughtering activities (98.0%). Selling of pigs was handled by both genders, but mainly men (73–80%), especially in larger farms. Likely challenges for upgrading pig VCs include limited training for producers, low concern for occupational health risks in all nodes, and misperceptions about food safety. In general, this study found no clear evidence of perceived gender inequality in the smallholder pig VCs in lowland Vietnam. Gendered upgrading in pig VCs should focus on improving women's ability to access veterinary services and animal disease management and on educating relevant VC actors about occupational health risks.
Flexible piezoelectric acoustic sensors (f-PAS) have attracted significant attention as a promising component for voice user interfaces (VUI) in the era of artificial intelligence of things (AIoT). ...The signal distortion issue of highly sensitive biomimetic f-PAS is one of the most challenging obstacle for real-life application, due to the fundamental difference compared with the conventional microphones. Here, a noise-robust flexible piezoelectric acoustic sensor (NPAS) is demonstrated by designing the multi-resonant bands outside the noise dominant frequency range. Broad voice coverage up to 8 kHz is achieved by adopting an advanced piezoelectric membrane (Nb-doped PZT; PNZT) with the optimized polymer ratio. Deep learning-based speech processing of multi-channel NPAS is demonstrated to show the outstanding improvement in speaker recognition and speech enhancement compared to a commercial microphone. Finally, the NPAS filtered the crowd condition noises, showing independent speaker’s speeches can be identified and digitalized simultaneously.
Display omitted
•We designed the multi-resonant bands outside the noise dominant frequency range.•Broad voice coverage up to 8 kHz was achieved by adopting Nb-doped PZT piezoelectric membrane.•Deep learning-based speech processing of our sensor showed the outstanding improvement compared to a commercial microphone.