Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


William A. Sethares has a total of 77 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Learning Label Hierarchy with Supervised Contrastive Learning

    by Ruixue Lian, William A. Sethares, Junjie Hu

    Supervised contrastive learning (SCL) frameworks treat each class as independent and thus consider all classes to be equally important. This neglects the common scenario in which label hierarchy exists, where fine-grained classes under the same category show more similarity than very different ones. This paper introduces a family of Label-Aware SCL methods (LASCL) that incorporates hierarchical information to SCL by leveraging similarities between classes, resulting in creating a more well-structured and discriminative feature space. This is achieved by first adjusting the distance between instances based on measures of the proximity of their classes with the scaled instance-instance-wise contrastive. An additional instance-center-wise contrastive is introduced to move within-class examples closer to their centers, which are represented by a set of learnable label parameters. The learned label parameters can be directly used as a nearest neighbor classifier without further finetuning. In this way, a better feature representation is generated with improvements of intra-cluster compactness and inter-cluster separation. Experiments on three datasets show that the proposed LASCL works well on text classification of distinguishing a single label among multi-labels, outperforming the baseline supervised approaches. Our code is publicly available.

  • Open Access

    Learning Label Hierarchy with Supervised Contrastive Learning

    by Ruixue Lian, William Sethares, Junjie Hu

    Ruixue Lian, William Sethares, Junjie Hu. Findings of the Association for Computational Linguistics: EACL 2024. 2024.


  • Open Access

    Assessing Speech Audibility via Syllabic-Rate Neural Responses in Adults and Children With and Without Hearing Loss

    by Varsha Pendyala, William Sethares, Vijayalakshmi Easwar

    An objective method for assessing speech audibility is essential to evaluate hearing aid benefit in children who are unable to participate in hearing tests. With consonant-vowel syllables, brainstem-dominant responses elicited at the voice fundamental frequency have proven successful for assessing audibility. This study aimed to harness the neural activity elicited by the slow envelope of the same repetitive consonant-vowel syllables to assess audibility. In adults and children with normal hearing and children with hearing loss wearing hearing aids, neural activity elicited by the stimulus /su∫i/ or /sa∫i/ presented at 55–75 dB SPL was analyzed using the temporal response function approach. No-stimulus runs or very low stimulus level (15 dB SPL) were used to simulate inaudible conditions in adults and children with normal hearing. Both groups of children demonstrated higher response amplitudes relative to adults. Detectability (sensitivity; true positive rate) ranged between 80.1 and 100%, and did not vary by group or stimulus level but varied by stimulus, with /sa∫i/ achieving 100% detectability at 65 dB SPL. The average minimum time needed to detect a response ranged between 3.7 and 6.4 min across stimuli and listener groups, with the shortest times recorded for stimulus /sa∫i/ and in children with hearing loss. Specificity was >94.9%. Responses to the slow envelope of non-meaningful consonant-vowel syllables can be used to ascertain audible vs. inaudible speech with sufficient accuracy within clinically feasible test times. Such responses can increase the clinical usefulness of existing objective approaches to evaluate hearing aid benefit.


  • Open Access

    Vision Backbone Enhancement via Multi-Stage Cross-Scale Attention

    by Liang Shang, Yanli Liu, Zhengyang Lou, Shuxue Quan, Nagesh Adluru, Bochen Guan, William A. Sethares

    Convolutional neural networks (CNNs) and vision transformers (ViTs) have achieved remarkable success in various vision tasks. However, many architectures do not consider interactions between feature maps from different stages and scales, which may limit their performance. In this work, we propose a simple add-on attention module to overcome these limitations via multi-stage and cross-scale interactions. Specifically, the proposed Multi-Stage Cross-Scale Attention (MSCSA) module takes feature maps from different stages to enable multi-stage interactions and achieves cross-scale interactions by computing self-attention at different scales based on the multi-stage feature maps. Our experiments on several downstream tasks show that MSCSA provides a significant performance boost with modest additional FLOPs and runtime.


  • Open Access

    The Watermark Imaging System: Revealing the Internal Structure of Historical Papers

    by Elisa Ou, Paul Messier, Ruixue Lian, Andrew Messier, William Sethares

    This paper introduces the Watermark Imaging System (WImSy) which can be used to photograph, document, and study sheets of paper. The WImSy provides surface images, raking light images, and transmitted light images of the paper, all in perfect alignment. We develop algorithms that exploit this alignment by combining several images together in a process that mimics both the “surface image removal” technique and the method of “high dynamic range” photographs. An improved optimization criterion and an automatic parameter selection procedure streamline the process and make it practical for art historians and conservators to extract the relevant information to study watermarks. The effectiveness of the method is demonstrated in several experiments on images taken with the WImSy at the Metropolitan Museum of Art in New York and at the Getty Museum in Los Angeles, and the results are compared with manually optimized images.


  • Exploring the Many Tunings of Balinese Gamelan

    by William Sethares, Wayne Vitale

    The study of gamelan tunings has a long history, both within Indonesia and among the many foreign researchers who have sought to reconcile those tuning systems with ideas and practices of Western music. Challenges arise because Indonesian gamelan and Western tuning systems are based on fundamentally different concepts and principles. This article introduces and describes the Gamelan Tuning Explorer (GTE), a Max patch intended to help facilitate, through visualization and adjustable parameters, greater understanding of certain core aspects of Balinese gamelan tuning—specifically, the nonstandardized tunings used for gamelan gong kebyar, a large orchestra of bronze-keyed metallophones, tuned gongs, and other instruments. The GTE incorporates measured tuning data from 47 complete gamelan gong kebyar; each encompasses about 150 keys and tuned gong chimes. Forty-two of these tuning data sets were collected in the mid 1970s by American ethnomusicologist Andrew Toth, who spent decades in Bali engrossed in research on this topic; five were remeasured by the authors in 2019 to document how they had changed over the ensuing decades. In addition to collecting this treasure trove of data, Toth identified several key conceptual ideas of Balinese gamelan tunings. These include the consistent use of ombak (the prominent beating of two paired notes played in “unison”); the relationship of paired-tuning ombak to octave treatment; the stretching and compression of octaves, both by register and by note; the Begbeg versus Tirus models of intervallic formation; and more. Toth's data and each of these tuning concepts have been analyzed in two previous articles (Vitale and Sethares 2021; Sethares and Vitale 2022). We seek to make these concepts more immediate and accessible with GTE by using a graphical, interactive interface. Each of these parameters can be explored by the simple manipulation of a slider or other interface element. The corresponding changes in the sound can be readily heard, and visualized with interactive Toth plots—graphic representations of gamelan tuning profiles developed by Toth and named in his honor. This article describes a series of observations enabled by GTE.


  • Open Access

    Predicting health-related quality of life change using natural language processing in thyroid cancer

    by Ruixue Lian, Vivian Hsiao, Juwon Hwang, Yue Ou, Sarah E. Robbins, Nadine P. Connor, Cameron L. Macdonald, Rebecca S. Sippel, William A. Sethares, David F. Schneider


  • Open Access

    Multi-Target PIR Indoor Localization and Tracking System with Artificial Intelligence

    by Xuan-Ying Chen, Chih-Yu Wen, William A. Sethares

    Pyroelectric infrared (PIR) sensors are low-cost, low-power, and highly reliable sensors that have been widely used in smart environments. Indoor localization systems may be wearable or non-wearable, where the latter are also known as device-free localization systems. Since binary PIR sensors detect only the presence of a subject’s motion in their field of view (FOV) without other information about the actual location, information from overlapping FOVs of multiple sensors can be useful for localization. This study introduces the PIRILS (pyroelectric infrared indoor localization system), in which the sensing signal processing algorithms are augmented by deep learning algorithms that are designed based on the operational characteristics of the PIR sensor. Expanding to the detection of multiple targets, the PIRILS develops a quantized scheme that exploits the behavior of an artificial neural network (ANN) model to demonstrate localization performance in tracking multiple targets. To further improve the localization performance, the PIRILS incorporates a data augmentation strategy that enhances the training data diversity of the target’s motion. Experimental results indicate system stability, improved positioning accuracy, and expanded applicability, thus providing an improved indoor multi-target localization framework.


  • Three-string inharmonic networks

    by Saba Goodarzi, William A. Sethares

    This paper studies the resonant frequencies of three-string networks by examining the roots of the relevant spectral equation. A collection of scaling laws are established which relate the frequencies to structured changes in the lengths, densities, and tensions of the strings. Asymptotic properties of the system are derived, and several situations where transcritical bifurcations occur are detailed. Numerical optimization is used to solve the inverse problem (where a desired set of frequencies is specified and the parameters of the system are adjusted to best realize the specification). The intrinsic dissonance of the overtones provides an approximate way to measure the inherent inharmonicity of the sound.


  • Open Access

    Cooperative Networked PIR Detection System for Indoor Human Localization

    by Chia-Ming Wu, Xuan-Ying Chen, Chih-Yu Wen, William A. Sethares

    Pyroelectric Infrared (PIR) sensors are low-cost, low-power, and highly reliable sensors that have been widely used in smart environments. Indoor localization systems can be categorized as wearable and non-wearable systems, where the latter are also known as device-free localization systems. Since the binary PIR sensor detects only the presence of a human motion in its field of view (FOV) without any other information about the actual location, utilizing the information of overlapping FOV of multiple sensors can be useful for localization. In this study, a PIR detector and sensing signal processing algorithms were designed based on the characteristics of the PIR sensor. We applied the designed PIR detector as a sensor node to create a non-wearable cooperative indoor human localization system. To improve the system performance, signal processing algorithms and refinement schemes (i.e., the Kalman filter, a Transferable Belief Model, and a TBM-based hybrid approach (TBM + Kalman filter)) were applied and compared. Experimental results indicated system stability and improved positioning accuracy, thus providing an indoor cooperative localization framework for PIR sensor networks.