Search

As global attention on sustainable and ethical food systems grows, animal welfare-friendly products (AWFP) are increasingly recognized as essential to addressing consumer and producer concerns. However, traditional research often neglects the interdependencies between production, retail, and consumption stages within the supply chain. This study examined how cross-stage interactions among producers, consumers, and retail intermediaries can promote AWFP adoption. By establishing a short value chain from production to consumption, we conducted a two-month choice experiment in the operational restaurant, employing a mixed-method approach to quantitatively and qualitatively assess stakeholder responses. The results revealed that providing information about AWFP practices significantly influenced consumer behavior, increasing both product selection and perceived value. Retailers recognized the potential for economic benefits and strengthened customer loyalty, while producers identified new revenue opportunities by re-fattening delivered cow. These coordinated changes - defined as synchronized actions and mutual reinforcement across production, retail, and consumption - generated positive feedback loops that motivated stakeholders to adopt AWFP practices. This research underscores the potential of strategically designed short value chain to foster cross-stage coordination and highlights their role as practical entry points for promoting sustainable and ethical food systems on a larger scale., Comment: 26 pages, 6 figures, 10 tables

Several attempts have been made to handle multiple source separation tasks such as speech enhancement, speech separation, sound event separation, music source separation (MSS), or cinematic audio source separation (CASS) with a single model. These models are trained on large-scale data including speech, instruments, or sound events and can often successfully separate a wide range of sources. However, it is still challenging for such models to cover all separation tasks because some of them are contradictory (e.g., musical instruments are separated in MSS while they have to be grouped in CASS). To overcome this issue and support all the major separation tasks, we propose a task-aware unified source separation (TUSS) model. The model uses a variable number of learnable prompts to specify which source to separate, and changes its behavior depending on the given prompts, enabling it to handle all the major separation tasks including contradictory ones. Experimental results demonstrate that the proposed TUSS model successfully handles the five major separation tasks mentioned earlier. We also provide some audio examples, including both synthetic mixtures and real recordings, to demonstrate how flexibly the TUSS model changes its behavior at inference depending on the prompts., Comment: Submitted to ICASSP 2025

As the global food system faces increasing challenges from sustainability, climate change, and food security issues, alternative food networks like Community-Supported Agriculture (CSA) play an essential role in fostering stronger connections between consumers and producers. However, understanding consumer engagement with CSA is fragmented, particularly in Japan where CSA participation is still emerging. This study aims to identify potential CSA participants in Japan and validate existing theories on CSA participation through a quantitative analysis of 2,484 Japanese consumers. Using choice experiments, Latent Class Analysis, and Partial Least Squares Structural Equation Modeling, we identified five distinct consumer segments. The "Sustainable Food Seekers" group showed the highest positive utility for CSA, driven primarily by "Food Education and Learning Opportunities" and "Contribution to Environmental and Social Issues." These factors were consistently significant across all segments, suggesting that many Japanese consumers value CSA for its educational and environmental benefits. In contrast, factors related to "Variety of Ingredients" were less influential in determining participation intentions. The findings suggest that promoting CSA in Japan may be most effective by emphasizing its role in environmental and social impact, rather than focusing solely on product attributes like organic certification, which is readily available in supermarkets. This reflects a key distinction between CSA adoption in Japan and in other cultural contexts, where access to organic produce is a primary driver. For "Sustainable Food Seekers," CSA offers a way to contribute to broader societal goals rather than just securing organic products., Comment: 29 pages, 5 figures

The goal of text-queried target sound extraction (TSE) is to extract from a mixture a sound source specified with a natural-language caption. While it is preferable to have access to large-scale text-audio pairs to address a variety of text prompts, the limited number of available high-quality text-audio pairs hinders the data scaling. To this end, this work explores how to leverage audio-only data without any captions for the text-queried TSE task to potentially scale up the data amount. A straightforward way to do so is to use a joint audio-text embedding model, such as the contrastive language-audio pre-training (CLAP) model, as a query encoder and train a TSE model using audio embeddings obtained from the ground-truth audio. The TSE model can then accept text queries at inference time by switching to the text encoder. While this approach should work if the audio and text embedding spaces in CLAP were well aligned, in practice, the embeddings have domain-specific information that causes the TSE model to overfit to audio queries. We investigate several methods to avoid overfitting and show that simple embedding-manipulation methods such as dropout can effectively alleviate this issue. Extensive experiments demonstrate that using audio-only data with embedding dropout is as effective as using text captions during training, and audio-only data can be effectively leveraged to improve text-queried TSE models., Comment: Submitted to ICASSP 2025

Reverberation as supervision (RAS) is a framework that allows for training monaural speech separation models from multi-channel mixtures in an unsupervised manner. In RAS, models are trained so that sources predicted from a mixture at an input channel can be mapped to reconstruct a mixture at a target channel. However, stable unsupervised training has so far only been achieved in over-determined source-channel conditions, leaving the key determined case unsolved. This work proposes enhanced RAS (ERAS) for solving this problem. Through qualitative analysis, we found that stable training can be achieved by leveraging the loss term to alleviate the frequency-permutation problem. Separation performance is also boosted by adding a novel loss term where separated signals mapped back to their own input mixture are used as pseudo-targets for the signals separated from other channels and mapped to the same channel. Experimental results demonstrate high stability and performance of ERAS., Comment: Accepted to Interspeech 2024

Time-frequency (TF) domain dual-path models achieve high-fidelity speech separation. While some previous state-of-the-art (SoTA) models rely on RNNs, this reliance means they lack the parallelizability, scalability, and versatility of Transformer blocks. Given the wide-ranging success of pure Transformer-based architectures in other fields, in this work we focus on removing the RNN from TF-domain dual-path models, while maintaining SoTA performance. This work presents TF-Locoformer, a Transformer-based model with LOcal-modeling by COnvolution. The model uses feed-forward networks (FFNs) with convolution layers, instead of linear layers, to capture local information, letting the self-attention focus on capturing global patterns. We place two such FFNs before and after self-attention to enhance the local-modeling capability. We also introduce a novel normalization for TF-domain dual-path models. Experiments on separation and enhancement datasets show that the proposed model meets or exceeds SoTA in multiple benchmarks with an RNN-free architecture., Comment: Accepted to IWAENC 2024

The last decade has witnessed significant advancements in deep learning-based speech enhancement (SE). However, most existing SE research has limitations on the coverage of SE sub-tasks, data diversity and amount, and evaluation metrics. To fill this gap and promote research toward universal SE, we establish a new SE challenge, named URGENT, to focus on the universality, robustness, and generalizability of SE. We aim to extend the SE definition to cover different sub-tasks to explore the limits of SE models, starting from denoising, dereverberation, bandwidth extension, and declipping. A novel framework is proposed to unify all these sub-tasks in a single model, allowing the use of all existing SE approaches. We collected public speech and noise data from different domains to construct diverse evaluation data. Finally, we discuss the insights gained from our preliminary baseline experiments based on both generative and discriminative SE methods with 12 curated metrics., Comment: 6 pages, 3 figures, 3 tables. Accepted by Interspeech 2024. An extended version of the accepted manuscript with appendix

Deep learning-based speech enhancement (SE) models have achieved impressive performance in the past decade. Numerous advanced architectures have been designed to deliver state-of-the-art performance; however, their scalability potential remains unrevealed. Meanwhile, the majority of research focuses on small-sized datasets with restricted diversity, leading to a plateau in performance improvement. In this paper, we aim to provide new insights for addressing the above issues by exploring the scalability of SE models in terms of architectures, model sizes, compute budgets, and dataset sizes. Our investigation involves several popular SE architectures and speech data from different domains. Experiments reveal both similarities and distinctions between the scaling effects in SE and other tasks such as speech recognition. These findings further provide insights into the under-explored SE directions, e.g., larger-scale multi-domain corpora and efficiently scalable architectures., Comment: 5 pages, 3 figures, 4 tables, Accepted by Interspeech 2024

In the context of the urgent need to establish sustainable food systems, Community Supported Agriculture (CSA), in which consumers share risks with producers, has gained increasing attention. Understanding the factors that influence consumer participation in CSA is crucial, yet the complete picture and interrelations of these factors remain unclear in existing studies. This research adopts a scoping review and the KJ method to elucidate the factors influencing consumer participation in CSA and to theorize the consumer participation. In particular, we focus on the dynamics of individual decision-making for participation, under the premise that individuals are embedded in socio-cultural environments. We examine the decision-making process based on the seesaw of expected gains and losses from participation, along with the reflexivity to the individual and the process of updating decision-making post-participation. Our study highlights how individual decision-making for participation is influenced by relationships with others within the embedded socio-cultural environment, as well as by attachment and connection to the community. It also shows that discrepancies between expectations and experiences post-participation, and the transformation of the social capital, promote the updating of decision-making processes. In addition, among the factors identified in this study for participation in CSA, the decision to participate was heavily influenced by expectations of variety of ingredients, suggesting that other factors such as food education and learning opportunities, contribution to environmental and social issues, and connections with people and nature had little impact. Although there are limitations, the insights gained from this study offer profound implications for stakeholders and provide valuable insights for more sustainable and efficient CSA practices., Comment: 45 pages, 4 figures

We propose a multi-task universal speech enhancement (MUSE) model that can perform five speech enhancement (SE) tasks: dereverberation, denoising, speech separation (SS), target speaker extraction (TSE), and speaker counting. This is achieved by integrating two modules into an SE model: 1) an internal separation module that does both speaker counting and separation; and 2) a TSE module that extracts the target speech from the internal separation outputs using target speaker cues. The model is trained to perform TSE if the target speaker cue is given and SS otherwise. By training the model to remove noise and reverberation, we allow the model to tackle the five tasks mentioned above with a single model, which has not been accomplished yet. Evaluation results demonstrate that the proposed MUSE model can successfully handle multiple tasks with a single model., Comment: 6 pages, 4 figures, 2 tables, accepted by ASRU2023

The past decade has witnessed substantial growth of data-driven speech enhancement (SE) techniques thanks to deep learning. While existing approaches have shown impressive performance in some common datasets, most of them are designed only for a single condition (e.g., single-channel, multi-channel, or a fixed sampling frequency) or only consider a single task (e.g., denoising or dereverberation). Currently, there is no universal SE approach that can effectively handle diverse input conditions with a single model. In this paper, we make the first attempt to investigate this line of research. First, we devise a single SE model that is independent of microphone channels, signal lengths, and sampling frequencies. Second, we design a universal SE benchmark by combining existing public corpora with multiple conditions. Our experiments on a wide range of datasets show that the proposed single model can successfully handle diverse conditions with strong performance., Comment: 6 pages, 3 figures, 5 tables, published in ASRU 2023 (corrected the results of noisy speech on CHiME-4 (Simu) in Table 4)

Speech signals, typically sampled at rates in the tens of thousands per second, contain redundancies, evoking inefficiencies in sequence modeling. High-dimensional speech features such as spectrograms are often used as the input for the subsequent model. However, they can still be redundant. Recent investigations proposed the use of discrete speech units derived from self-supervised learning representations, which significantly compresses the size of speech data. Applying various methods, such as de-duplication and subword modeling, can further compress the speech sequence length. Hence, training time is significantly reduced while retaining notable performance. In this study, we undertake a comprehensive and systematic exploration into the application of discrete units within end-to-end speech processing models. Experiments on 12 automatic speech recognition, 3 speech translation, and 1 spoken language understanding corpora demonstrate that discrete units achieve reasonably good results in almost all the settings. We intend to release our configurations and trained models to foster future research efforts., Comment: Submitted to IEEE ICASSP 2024

High-quality and intelligible speech is essential to text-to-speech (TTS) model training, however, obtaining high-quality data for low-resource languages is challenging and expensive. Applying speech enhancement on Automatic Speech Recognition (ASR) corpus mitigates the issue by augmenting the training data, while how the nonlinear speech distortion brought by speech enhancement models affects TTS training still needs to be investigated. In this paper, we train a TF-GridNet speech enhancement model and apply it to low-resource datasets that were collected for the ASR task, then train a discrete unit based TTS model on the enhanced speech. We use Arabic datasets as an example and show that the proposed pipeline significantly improves the low-resource TTS system compared with other baseline methods in terms of ASR WER metric. We also run empirical analysis on the correlation between speech enhancement and TTS performances., Comment: Submitted to ICASSP 2024

We propose an unsupervised approach for training separation models from scratch using RemixIT and Self-Remixing, which are recently proposed self-supervised learning methods for refining pre-trained models. They first separate mixtures with a teacher model and create pseudo-mixtures by shuffling and remixing the separated signals. A student model is then trained to separate the pseudo-mixtures using either the teacher's outputs or the initial mixtures as supervision. To refine the teacher's outputs, the teacher's weights are updated with the student's weights. While these methods originally assumed that the teacher is pre-trained, we show that they are capable of training models from scratch. We also introduce a simple remixing method to stabilize training. Experimental results demonstrate that the proposed approach outperforms mixture invariant training, which is currently the only available approach for training a monaural separation model from scratch., Comment: Interspeech2023, 5pages, 2figures, 2tables

As global demand for animal products is projected to increase significantly by 2050, driven by population growth and increased incomes, smart livestock technologies are essential for improving efficiency, animal welfare, and environmental sustainability. Conducted within the unique agricultural context of Japan, characterized by small-scale, family-run farms and strong government protection policies, our study builds upon traditional theoretical frameworks that often oversimplify farmers' decision-making processes. By employing a scoping review, expert interviews, and a Modified Grounded Theory Approach, our research uncovers the intricate interplay between individual farmer values, farm management policies, social relations, agricultural policies, and livestock industry trends. We particularly highlight the unique dynamics within family-owned businesses, noting the tension between an "advanced management mindset" and "conservatism." Our study reveals that technology adoption is a sequential and iterative process, influenced by technology availability, farmers' digital literacy, technology implementation support, and observable technology impacts on animal health and productivity. These insights highlight the need for tailored support mechanisms and policies to enhance technology uptake, thereby promoting sustainable and efficient livestock production system., Comment: 58 pages, 3 figures

This study developed a novel thermal control system to cool detectors of the General AntiParticle Spectrometer (GAPS) before its flights. GAPS is a balloon-borne cosmic-ray observation experiment. In its payload, GAPS contains over 1000 silicon detectors that must be cooled below $-40^{\circ}\mbox{C}$. All detectors are thermally coupled to a unique heat-pipe system (HPS) that transfers heat from the detectors to a radiator. The radiator is designed to be cooled below $-50^{\circ}\mbox{C}$ during the flight by exposure to space. The pre-flight state of the detectors is checked on the ground at 1 atm and ambient room temperature, but the radiator cannot be similarly cooled. The authors have developed a ground cooling system (GCS) to chill the detectors for ground testing. The GCS consists of a cold plate, a chiller, and insulating foam. The cold plate is designed to be attached to the radiator and cooled by a coolant pumped by the chiller. The payload configuration, including the HPS, can be the same as that of the flight. The GCS design was validated by thermal tests using a scale model. The GCS design is simple and provides a practical guideline, including a simple estimation of appropriate thermal insulation thickness, which can be easily adapted to other applications., Comment: 8 pages, 14 figures, 3 tables

Human pose estimation, particularly in athletes, can help improve their performance. However, this estimation is difficult using existing methods, such as human annotation, if the subjects wear loose-fitting clothes such as ski/snowboard wears. This study developed a method for obtaining the ground truth data on two-dimensional (2D) poses of a human wearing loose-fitting clothes. This method uses fast-flushing light-emitting diodes (LEDs). The subjects were required to wear loose-fitting clothes and place the LED on the target joints. The LEDs were observed directly using a camera by selecting thin filmy loose-fitting clothes. The proposed method captures the scene at 240 fps by using a high-frame-rate camera and renders two 30 fps image sequences by extracting LED-on and -off frames. The temporal differences between the two video sequences can be ignored, considering the speed of human motion. The LED-on video was used to manually annotate the joints and thus obtain the ground truth data. Additionally, the LED-off video, equivalent to a standard video at 30 fps, confirmed the accuracy of existing machine learning-based methods and manual annotations. Experiments demonstrated that the proposed method can obtain ground truth data for standard RGB videos. Further, it was revealed that neither manual annotation nor the state-of-the-art pose estimator obtains the correct position of target joints., Comment: Extended abstract of WACV2023 workshop on Computer Vision 4 Winter Sports

We present Self-Remixing, a novel self-supervised speech separation method, which refines a pre-trained separation model in an unsupervised manner. The proposed method consists of a shuffler module and a solver module, and they grow together through separation and remixing processes. Specifically, the shuffler first separates observed mixtures and makes pseudo-mixtures by shuffling and remixing the separated signals. The solver then separates the pseudo-mixtures and remixes the separated signals back to the observed mixtures. The solver is trained using the observed mixtures as supervision, while the shuffler's weights are updated by taking the moving average with the solver's, generating the pseudo-mixtures with fewer distortions. Our experiments demonstrate that Self-Remixing gives better performance over existing remixing-based self-supervised methods with the same or less training costs under unsupervised setup. Self-Remixing also outperforms baselines in semi-supervised domain adaptation, showing effectiveness in multiple setups., Comment: Accepted by ICASSP2023, 5pages, 2figures, 2tables