Search

Sr-doped nickelate, Nd1-xSrxNiO3 (NSNO), perovskite thin films and Ruddlesden-Popper (RP) phases are actively investigated because of their physical properties, such as the metal-insulator transition and superconductivity. However, achieving epitaxial growth of NSNO perovskite and RP phase films in a sputtering system is challenging compared to pulsed laser deposition and molecular beam epitaxy, due to the difficulty in stabilizing nickel oxidation states and minimizing structural defects. Here, we used an off-axis radio frequency (RF) magnetron sputtering to fabricate epitaxial NSNO perovskite and RP phase thin films on SrTiO3 (001) substrates, systematically controlling the growth temperatures. We investigated the thermal stability of the perovskite phase and the structural and electronic characteristics of the RP phase films. These findings provide valuable insights into the synthesis of nickelate RP phase films using RF magnetron sputtering, paving the way for scalable thin films fabrication technologies., Comment: 13 pages, 4 figures

In this paper, we introduce the Actions and Objects Pathways (AOPath) for out-of-domain generalization in video question answering tasks. AOPath leverages features from a large pretrained model to enhance generalizability without the need for explicit training on the unseen domains. Inspired by human brain, AOPath dissociates the pretrained features into action and object features, and subsequently processes them through separate reasoning pathways. It utilizes a novel module which converts out-of-domain features into domain-agnostic features without introducing any trainable weights. We validate the proposed approach on the TVQA dataset, which is partitioned into multiple subsets based on genre to facilitate the assessment of generalizability. The proposed approach demonstrates 5% and 4% superior performance over conventional classifiers on out-of-domain and in-domain datasets, respectively. It also outperforms prior methods that involve training millions of parameters, whereas the proposed approach trains very few parameters.

Respiratory sound classification (RSC) is challenging due to varied acoustic signatures, primarily influenced by patient demographics and recording environments. To address this issue, we introduce a text-audio multimodal model that utilizes metadata of respiratory sounds, which provides useful complementary information for RSC. Specifically, we fine-tune a pretrained text-audio multimodal model using free-text descriptions derived from the sound samples' metadata which includes the gender and age of patients, type of recording devices, and recording location on the patient's body. Our method achieves state-of-the-art performance on the ICBHI dataset, surpassing the previous best result by a notable margin of 1.17%. This result validates the effectiveness of leveraging metadata and respiratory sound samples in enhancing RSC performance. Additionally, we investigate the model performance in the case where metadata is partially unavailable, which may occur in real-world clinical setting., Comment: Accepted INTERSPEECH 2024

Recently, deep end-to-end learning has been studied for intent classification in Spoken Language Understanding (SLU). However, end-to-end models require a large amount of speech data with intent labels, and highly optimized models are generally sensitive to the inconsistency between the training and evaluation conditions. Therefore, a natural language understanding approach based on Automatic Speech Recognition (ASR) remains attractive because it can utilize a pre-trained general language model and adapt to the mismatch of the speech input environment. Using this module-based approach, we improve a noisy-channel model to handle transcription inconsistencies caused by ASR errors. We propose a two-stage method, Contrastive and Consistency Learning (CCL), that correlates error patterns between clean and noisy ASR transcripts and emphasizes the consistency of the latent features of the two transcripts. Experiments on four benchmark datasets show that CCL outperforms existing methods and improves the ASR robustness in various noisy environments. Code is available at https://github.com/syoung7388/CCL., Comment: Accepted NAACL 2024

Recent advancements in AI have democratized its deployment as a healthcare assistant. While pretrained models from large-scale visual and audio datasets have demonstrably generalized to this task, surprisingly, no studies have explored pretrained speech models, which, as human-originated sounds, intuitively would share closer resemblance to lung sounds. This paper explores the efficacy of pretrained speech models for respiratory sound classification. We find that there is a characterization gap between speech and lung sound samples, and to bridge this gap, data augmentation is essential. However, the most widely used augmentation technique for audio and speech, SpecAugment, requires 2-dimensional spectrogram format and cannot be applied to models pretrained on speech waveforms. To address this, we propose RepAugment, an input-agnostic representation-level augmentation technique that outperforms SpecAugment, but is also suitable for respiratory sound classification with waveform pretrained models. Experimental results show that our approach outperforms the SpecAugment, demonstrating a substantial improvement in the accuracy of minority disease classes, reaching up to 7.14%., Comment: Accepted EMBC 2024

Despite the remarkable advances in deep learning technology, achieving satisfactory performance in lung sound classification remains a challenge due to the scarcity of available data. Moreover, the respiratory sound samples are collected from a variety of electronic stethoscopes, which could potentially introduce biases into the trained models. When a significant distribution shift occurs within the test dataset or in a practical scenario, it can substantially decrease the performance. To tackle this issue, we introduce cross-domain adaptation techniques, which transfer the knowledge from a source domain to a distinct target domain. In particular, by considering different stethoscope types as individual domains, we propose a novel stethoscope-guided supervised contrastive learning approach. This method can mitigate any domain-related disparities and thus enables the model to distinguish respiratory sounds of the recording variation of the stethoscope. The experimental results on the ICBHI dataset demonstrate that the proposed methods are effective in reducing the domain dependency and achieving the ICBHI Score of 61.71%, which is a significant improvement of 2.16% over the baseline., Comment: accepted to ICASSP 2024

Deep generative models have emerged as a promising approach in the medical image domain to address data scarcity. However, their use for sequential data like respiratory sounds is less explored. In this work, we propose a straightforward approach to augment imbalanced respiratory sound data using an audio diffusion model as a conditional neural vocoder. We also demonstrate a simple yet effective adversarial fine-tuning method to align features between the synthetic and real respiratory sound samples to improve respiratory sound classification performance. Our experimental results on the ICBHI dataset demonstrate that the proposed adversarial fine-tuning is effective, while only using the conventional augmentation method shows performance degradation. Moreover, our method outperforms the baseline by 2.24% on the ICBHI Score and improves the accuracy of the minority classes up to 26.58%. For the supplementary material, we provide the code at https://github.com/kaen2891/adversarial_fine-tuning_using_generated_respiratory_sound., Comment: accepted in NeurIPS 2023 Workshop on Deep Generative Models for Health (DGM4H)

We studied models of leptogenesis where three right-handed Majorana neutrinos are involved and the minimal-extended seesaw mechanism including an additional singlet field produces four light neutrinos. This study shows that the type of mass ordering and heavy Majorana scales can be determined by inputting the simplest orthogonal matrix into the Casas-Ibarra(CI) representation of seesaw. The CP asymmetry produced from the decays of heavy neutrinos and the dilution mass are predicted in terms of the mass and mixing elements of the fourth neutrino. Upon the choice of CI matrix, the existence of a light sterile neutrino is required to explain the high-energy lepton asymmetry in light of phenomenological measurements. Although there are several free parameters attributable to an additional neutrino, the model can be in part constrained by low-energy experiments such as sterile neutrino searches and neutrinoless double-beta decays, as well as the observed baryon asymmetry in the universe., Comment: 23 pages, 7 figures

A prompt severity assessment model of patients confirmed for having infectious diseases could enable efficient diagnosis while alleviating burden on the medical system. This study aims to develop a SARS-CoV-2 severity assessment model and establish a medical system that allows patients to check the severity of their cases and informs them to visit the appropriate clinic center based on past treatment data of other patients with similar severity levels. This paper provides the development processes of a severity assessment model using machine learning techniques and its application on SARS-CoV-2 patients. The proposed model is trained on a nationwide dataset provided by a Korean government agency and only requires patients' basic personal data, allowing them to judge the severity of their own cases. After modeling, the boosting-based decision tree model was selected as the classifier while mortality rate was interpreted as the probability score. The dataset was collected from all Korean citizens who were confirmed with COVID-19 between February, 2020 and July, 2021. The experiments achieved high model performance with an approximate precision of $0{\cdot}923$ and AUROC score of $0{\cdot}950$ [$95$% Tolerance Interval $0{\cdot}940$-$0{\cdot}958$, $95$% Confidence Interval $0{\cdot}949$-$0{\cdot}950$]. Moreover, our experiments identified the most important variables affecting the severity in the model via sensitivity analysis. The prompt severity assessment model for managing infectious people has been attained through using a nationwide dataset. It has demonstrated its superior performance by surpassing that of conventional risk assessments. With the model's high performance and easily accessible features, the triage algorithm is expected to be particularly useful when patients monitor their health status by themselves through smartphone applications., Comment: To appear in Infectious Diseases and Therapy

Improving lithium-ion batteries (LIBs) safety remains in a challenging task when compared with the tremendous progress made in their performance in recent years. Embedding thermo-responsive polymer switching materials (TRPS) into LIB cells has been proved to be a promising strategy to provide consistent thermal abuse protections at coin-cell level. However, it is unrealistic to achieve large-scale applications without further demonstration in high-capacity pouch cells. Here, we employed tungsten carbide (WC) as a novel conductive filler, and successfully overcame the intrinsic processing difficulty of polyethylene (PE) matrix in a scalable solvent-based method to obtain ultra-thin, uniform, highly conductive TRPS. Moreover, by integrating TRPS directly into LIB electrodes, no extra fabrication facilities or processes are required for making the cells. As a result, multi-layer pouch cells with consistent electrochemical performance and thermal abuse protection function were fabricated using industry relevant manufacturing techniques, which brings TRPS one step further to the real application scenarios., Comment: 20 pages, 6 figures

Mel-frequency filter bank (MFB) based approaches have the advantage of learning speech compared to raw spectrum since MFB has less feature size. However, speech generator with MFB approaches require additional vocoder that needs a huge amount of computation expense for training process. The additional pre/post processing such as MFB and vocoder is not essential to convert real human speech to others. It is possible to only use the raw spectrum along with the phase to generate different style of voices with clear pronunciation. In this regard, we propose a fast and effective approach to convert realistic voices using raw spectrum in a parallel manner. Our transformer-based model architecture which does not have any CNN or RNN layers has shown the advantage of learning fast and solved the limitation of sequential computation of conventional RNN. In this paper, we introduce a vocoder-free end-to-end voice conversion method using transformer network. The presented conversion model can also be used in speaker adaptation for speech recognition. Our approach can convert the source voice to a target voice without using MFB and vocoder. We can get an adapted MFB for speech recognition by multiplying the converted magnitude with phase. We perform our voice conversion experiments on TIDIGITS dataset using the metrics such as naturalness, similarity, and clarity with mean opinion score, respectively., Comment: Work in progress

One of the most important factors in training object recognition networks using convolutional neural networks (CNNs) is the provision of annotated data accompanying human judgment. Particularly, in object detection or semantic segmentation, the annotation process requires considerable human effort. In this paper, we propose a semi-supervised learning (SSL)-based training methodology for object detection, which makes use of automatic labeling of un-annotated data by applying a network previously trained from an annotated dataset. Because an inferred label by the trained network is dependent on the learned parameters, it is often meaningless for re-training the network. To transfer a valuable inferred label to the unlabeled data, we propose a re-alignment method based on co-occurrence matrix analysis that takes into account one-hot-vector encoding of the estimated label and the correlation between the objects in the image. We used an MS-COCO detection dataset to verify the performance of the proposed SSL method and deformable neural networks (D-ConvNets) as an object detector for basic training. The performance of the existing state-of-the-art detectors (DConvNets, YOLO v2, and single shot multi-box detector (SSD)) can be improved by the proposed SSL method without using the additional model parameter or modifying the network architecture., Comment: Submitted to International Conference on Image Processing (ICIP) 2018

Traditional pedestrian collision warning systems sometimes raise alarms even when there is no danger (e.g., when all pedestrians are walking on the sidewalk). These false alarms can make it difficult for drivers to concentrate on their driving. In this paper, we propose a novel framework for an end-to-end pedestrian collision warning system based on a convolutional neural network. Semantic segmentation information is used to train the convolutional neural network and two loss functions, such as cross entropy and Euclidean losses, are minimized. Finally, we demonstrate the effectiveness of our method in reducing false alarms and increasing warning accuracy compared to a traditional histogram of oriented gradients (HoG)-based system., Comment: 6 pages, 5 figures

A fibrous herringbone-modified helicoidal architecture is identified within the exocuticle of an impact-resistant crustacean appendage. This previously unreported composite microstructure, which features highly textured apatite mineral templated by an alpha-chitin matrix, provides enhanced stress redistribution and energy absorption over the traditional helicoidal design under compressive loading. Nanoscale toughening mechanisms are also identified using high load nanoindentation and in-situ TEM picoindentation., Comment: 6 figures. Published in Advanced Materials 2016

Directional transport properties at the nanoscale remain a challenge primarily due to issues associated with control over the underlying anisotropy and scalability to macroscopic scales. In this letter, we develop a facile approach based on template-guided fluidic assembly of high mobility building blocks - single walled carbon nanotubes (SWNTs) - to fabricate ultra-thin and anisotropic SWNT films. A major advancement is the complete control over the anisotropy in the assembled nanostructure, realized by three-dimensional engineering of dip-coated SWNT thin films into alternating hydrophilic and hydrophobic micro-line patterns with prescribed intra/inter-line widths and line thicknesses. Variations in the contact line profile results in an evaporation-controlled assembly mechanism that leads to the formation of an alternating, and more importantly, contiguous SWNT network. Evidently, the nanoscopic thickness modulations are direct reflections of the substrate geometry and chemistry. The nanostructured film exhibits significant anisotropy in their electrical and thermal transport properties as well as optical transparency, as revealed by characterization studies. The direct interplay between the anisotropy and the 3D micro-line patterns of the substrate combined with the wafer-level scalability of the fluidic assembly allows us to tune the transport properties for a host of nanoelectronic applications., Comment: 14 pages, 6 figures, submitted to Advanced Functional Materials