Article
Multidisciplinary Sciences
Adam R. Fishbein, Nora H. Prior, Jane A. Brown, Gregory F. Ball, Robert J. Dooling
Summary: Studies of acoustic communication often focus on vocalization categories and units, but there is also subtle variation in how signals are uttered. This study on zebra finches demonstrates their ability to easily discriminate between different renditions of vocal signals, suggesting that sensitivity to fine acoustic details may be a primary channel of information in their song, as well as a shared property across species' vocal communication systems.
SCIENTIFIC REPORTS
(2021)
Article
Automation & Control Systems
Sam St John, Matthew Alberts, Jaydeep Karandikar, Jamie Coble, Bradley Jared, Tony Schmitz, Christoph Ramsauer, David Leitner, Anahita Khojandi
Summary: In this study, a Random Forest classifier with Recursive Feature Elimination is used to analyze machining audio collected by a single microphone during down-milling operations. The classifier can accurately predict the stability of the machining process without the need for additional sensors. This low-cost approach enables real-time visualization of the machining process and helps machinists avoid unstable processes.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2023)
Article
Environmental Sciences
Weiping Cheng, Yongxin Shen
Summary: This paper collected and analyzed leak detection signals from cast iron pipelines in a water distribution system (WDS). Significant statistical distributions were found in the characteristics of the signals. Using ML and Random Forest models, the leak detection signals were successfully recognized and classified with high recall rate and low false positive rate.
Article
Engineering, Biomedical
Sara Hawi, Jana Alhozami, Raneem AlQahtani, Dannah AlSafran, Maram Alqarni, Lola El Sahmarany
Summary: In recent years, acoustic signals have become increasingly popular as biomarkers for detecting Parkinson's disease. This paper presents a novel method to investigate the effect of a dataset incorporating long-term and short-term features on the performance of a random forest model for Parkinson's disease detection. The results show that the combination of these features improves the detection accuracy compared to using independent sets of features.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2022)
Article
Biology
Fiona Backhouse, Anastasia H. Dalziell, Robert D. Magrath, Justin A. Welbergen
Summary: Most studies on acoustic communication have focused on short vocalization units, but little is known about the drivers of sequence structure. In this study, we investigate the organization, transmission, and function of vocal sequences in Albert's lyrebirds. We found that individual males organized their vocal units into stereotyped sequences, which were shared within and among populations, suggesting social transmission. Furthermore, the sequence structure was found to enhance the receiver's perception of repertoire complexity.
PROCEEDINGS OF THE ROYAL SOCIETY B-BIOLOGICAL SCIENCES
(2022)
Article
Automation & Control Systems
Amit K. Shukla, Shubham Srivastav, Sandeep Kumar, Pranab K. Muhuri
Summary: In an Industry 4.0 ecosystem, digital interconnectedness and automation integration increase productivity but also bring the risk of cyber-attacks. Efficient threat intelligence techniques or intrusion detection systems (IDSs) are needed to identify and monitor these attacks. This paper proposes a novel unsupervised IDS, UInDeSI4.0, which uses feature selection and isolation forest to detect network threats in an unsupervised manner. Experimental results show that UInDeSI4.0 provides better accuracy and minimal features compared to traditional IDSs.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Engineering, Multidisciplinary
Dengyu Xiao, Chengjin Qin, Honggan Yu, Yixiang Huang, Chengliang Liu, Jianwei Zhang
Summary: A Noisy Domain Adaptive marginal Stacking Denoising Auto-encoder (NDAmSDA) based on acoustic signals is proposed for domain adaptation between different noise levels. By addressing the domain shift problem and simplifying deep learning frameworks to improve training acceleration, and transferring classifiers between different noise levels, the model extends diagnostic models' capabilities in real-world scenarios.
Article
Ecology
Sandor Zsebok, Denes Schmera, Miklos Laczi, Gergely Nagy, Eva Vaskuti, Janos Torok, Laszlo Zsolt Garamszegi
Summary: Studying the diversity of animal signals is crucial for understanding the evolution of communication systems, but current methods for quantifying acoustic diversity have limitations. The authors propose a new framework that utilizes tools from community ecology to decompose acoustic diversity and characterize the complexity of animal communication systems. By applying different diversity estimates, they can reveal additional insights about the function and evolution of communication systems, beyond traditional methods.
METHODS IN ECOLOGY AND EVOLUTION
(2021)
Article
Multidisciplinary Sciences
Mathilde Martin, Tess Gridley, Simon Harvey Elwen, Isabelle Charrier
Summary: Communication is essential for the survival of animal species, with signals playing a crucial role in various social interactions. This study on the acoustic communication system of Cape fur seals revealed distinct call types that could convey information about age and/or sex, as well as the presence of acoustic partitioning in their repertoire. These vocalizations appear to enhance social interactions in noisy environments, providing potential advantages for discrimination among individuals.
ROYAL SOCIETY OPEN SCIENCE
(2021)
Article
Geochemistry & Geophysics
T. Dittmann, Y. Liu, Y. Morton, D. Mencin
Summary: This study uses high rate GNSS time series to detect seismic signals and applies machine learning classification algorithm, which outperforms existing methods. The model can rapidly and accurately detect seismic events, with wide application prospects.
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
(2022)
Article
Geochemistry & Geophysics
T. Dittmann, Y. Liu, Y. Morton, D. Mencin
Summary: This study successfully differentiated seismic events from noise through a machine learning classifier, improving the accuracy of seismic signal detection. The model demonstrated high efficiency in detecting seismic events and accurately identifying them in stand-alone mode.
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
(2022)
Article
Geosciences, Multidisciplinary
Faraz S. Tehrani, Giorgio Santinelli, Meylin Herrera Herrera
Summary: This study introduces an Object-Based Image Analysis methodology using Machine Learning to detect landslides in multiple regions. The Random Forest models show high precision and recall in identifying landslide segments, with RF1 outperforming RF2 in accuracy.
GEOMATICS NATURAL HAZARDS & RISK
(2021)
Article
Ecology
Yi-Jen Sun, Shih-Ching Yen, Tzu-Hao Lin
Summary: Soundscapes contain valuable acoustic information, and separating sound sources is crucial for assessing acoustic diversity.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Acoustics
Fenghua Li, Kai Wang, Xishan Yang, Bo Zhang, Yanjun Zhang
Summary: This article presents a machine learning-based method for passive ocean acoustic thermometry, which can estimate the averaged sound speed from noise cross-correlation functions within half an hour. An empirical equation is proposed to describe the relationships between features based on feature importance analysis. Comparisons of estimation results among different methods are conducted to demonstrate the advantage of the machine learning-based approach.
Article
Agriculture, Dairy & Animal Science
Meng-Meng Chen, Yu-Heng Zhang, Yi-Mei Tai, Xi Wang
Summary: This study provides insight into the importance of vocal communication in animal group cooperation and decision-making, and identifies the influence of sex and social centrality on vocalizations.
Article
Biology
Karan J. Odom, Marcelo Araya-Salas, Janelle L. Morano, Russell A. Ligon, Gavin M. Leighton, Conor C. Taff, Anastasia H. Dalziell, Alexis C. Billings, Ryan R. Germain, Michael Pardo, Luciana Guimaraes de Andrade, Daniela Hedwig, Sara C. Keen, Yu Shiu, Russell A. Charif, Michael S. Webster, Aaron N. Rice
Summary: Animals produce a wide variety of sounds with highly variable acoustic structures, and phylogenetic comparative analyses can help understand the causes and consequences of this variation. Acoustic and evolutionary analyses are becoming increasingly sophisticated, making it challenging to choose appropriate methods. By providing a roadmap for quantifying and comparing sound in a phylogenetic context, researchers can address some of the challenges in this field.
BIOLOGICAL REVIEWS
(2021)
Editorial Material
Multidisciplinary Sciences
Stephanie A. White, Timothy F. Wright
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Behavioral Sciences
Grace Smith-Vidaurre, Valeria Perez-Marrufo, Timothy F. Wright
Summary: The study found that invasive populations have smaller population sizes, simpler frequency modulation patterns in calls, and reduced individual identity content, possibly due to a reduction in population size following invasion. Frequency modulation patterns were simpler in urban habitats, indicating that urban environments may alter the social environment and influence the complexity of learned individual signatures.
Article
Zoology
Caitlin R. Gabor, Stephanie N. Kivlin, Jessica Hua, Nate Bickford, Martha O. Burford Reiskind, Timothy F. Wright
Summary: Global environmental changes are forcing organisms to respond at an unprecedented pace, but we still have limited understanding of why some species can respond to these changes while others cannot. The concept of multidimensional phenospace is introduced to help understand organismal evolutionary responses to environmental change. The paper describes five barriers that hinder our understanding of these responses and suggests solutions to overcome them. Examples of target species are provided to study the interaction between phenotypic plasticity and adaptive evolution in changing phenospace.
INTEGRATIVE AND COMPARATIVE BIOLOGY
(2022)
Review
Behavioral Sciences
Timothy F. Wright, Elizabeth P. Derryberry
Summary: Vocal learning is a complex cognitive trait that varies across species in terms of which vocalizations are learned, how much is learned, when it is learned, who it is learned from, the extent of the internal template, and how the template is integrated with social learning and innovation. By examining vocal learning through a multi-dimensional framework, researchers can accelerate understanding of why vocal learning has evolved and how brains became capable of this behavior.
NEUROSCIENCE AND BIOBEHAVIORAL REVIEWS
(2021)
Article
Biodiversity Conservation
Carlos Campos, Melinda A. Martinez, Daniel Acosta, Jose A. Diaz-Luque, Igor Berkunsky, Nadine L. Lamberski, Javier Cruz-Nieto, Michael A. Russello, Timothy F. Wright
Summary: Endangered parrot populations may exhibit weak genetic differentiation between wild subpopulations and between wild and captive populations. Similar levels of genetic diversity were detected in the wild and captive populations of both species, with private alleles found in captivity and in the wild in the thick-billed parrot. This suggests the potential benefits of reintroduction of genetic variation found in captivity for both species.
Article
Biology
Simeon Q. Smeele, Dalia A. Conde, Annette Baudisch, Simon Bruslund, Andrew Iwaniuk, Johanna Staerk, Timothy F. Wright, Anna M. Young, Mary Brooke McElreath, Lucy Aplin
Summary: Previous studies have shown a correlation between longevity and brain size across different animal taxa. This study focused on parrots, which are known for their long lifespan and cognitive complexity. Using a large-scale comparative analysis, the researchers found a consistent correlation between relative brain size and life expectancy in parrots. This correlation was best explained by the direct effect of relative brain size, rather than developmental time, clutch size or age at first reproduction.
PROCEEDINGS OF THE ROYAL SOCIETY B-BIOLOGICAL SCIENCES
(2022)
Article
Biology
Diego F. Beltran, Marcelo Araya-Salas, Juan L. Parra, F. Gary Stiles, Alejandro Rico-Guevara
Summary: This study investigates the variation of hummingbird dimorphism across ecological gradients and finds that morphological dimorphism is negatively correlated with elevation, while dichromatism and song complexity are positively correlated with elevation. The results suggest that flight constraints, predation pressure, and habitat effects play important roles in shaping sexually dimorphic traits.
PROCEEDINGS OF THE ROYAL SOCIETY B-BIOLOGICAL SCIENCES
(2022)
Article
Biology
Bushra Moussaoui, Samantha L. Overcashier, Gregory M. Kohn, Marcelo Araya-Salas, Timothy F. Wright
Summary: In some species, the ability to acquire new vocalizations persists into adulthood and may be an important mediator of social interactions. This study hypothesizes that vocal learning exhibits senescence and that this decline relates to age-dependent changes in social behaviour. The results suggest that many components of vocal learning are largely maintained into later adulthood in the budgerigar.
PROCEEDINGS OF THE ROYAL SOCIETY B-BIOLOGICAL SCIENCES
(2023)
Article
Agriculture, Dairy & Animal Science
Katarzyna Wojczulanis-Jakubas, Marcelo Araya-Salas
Summary: In this study, the foraging efficiency of long-billed hermit hummingbirds was examined in relation to their exploration, risk avoidance, and arousal behavioral traits. It was found that foraging efficiency was lower in high-risk conditions, but the effects of behavioral traits varied depending on the environmental conditions. More explorative individuals had higher foraging efficiency in low-risk conditions, while the opposite was true in high-risk conditions. Regardless of the conditions, foraging efficiency increased with bird arousal and decreased with risk avoidance. These findings highlight the importance of considering additional behavioral dimensions to better understand individual foraging strategies.
Article
Ecology
Marcelo Araya-Salas, Grace Smith-Vidaurre, Gloriana Chaverri, Juan C. Brenes, Fabiola Chirino, Jorge Elizondo-Calvo, Alejandro Rico-Guevara
Summary: Animal acoustic signals are widely used for research purposes due to their ease of registration and versatility. However, analyzing the large datasets generated by bioacoustics research can be challenging. The ohun R package aims to aid in automated sound event detection and evaluation, providing tools to optimize detection routines and compare performance among different approaches.
METHODS IN ECOLOGY AND EVOLUTION
(2023)