Kexin Feng
Research Subjects
Kexin Feng has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Theodora Chaspari
University of Colorado Boulder
5 shared publications - Amran Haroon
Computer Science and Engineering, Texas A&M University
2 shared publications - Charles Swanson
Biological and Agricultural Engineering, Texas A&M University
2 shared publications - Liuyi Jin
Computer Science and Engineering, Texas A&M University
2 shared publications - Radu Stoleru
Computer Science and Engineering, Texas A&M University
2 shared publications - Tian Liu
Computer Science and Engineering, Texas A&M University
2 shared publications
Latest Publications
- Open Access
A Comparison of Self-Reports and Unobtrusive Multimodal Measures of Stress in a Realistic Simulated Task Environment
Although a popular measurement method in different fields, self-report measures have multiple shortcomings that stem from their very inherent reliance on the participant’s subjective judgment or recall. Given these limitations, in the domain of stress research, unobtrusive measures have been proposed as alternatives. Using a sample of 78 university students who participated in a disaster response simulation in a synthetic task environment, the present study sought to examine whether unobtrusive measures of stress (i.e., physiological, computer interaction, and facial data) can serve as valid substitutes for self-report measures (i.e., state anxiety, task load, and momentary stress). To that end, the study compared the extent to which features from these different modalities accurately (1) classified participants into the stress condition to which they were assigned as well as (2) predicted their performance on a complex synthetic task. The results indicated that the unobtrusive measures individually predicted task performance better than self-report measures, whereas self-report measures predicted the stress condition to which participants were randomly assigned (i.e., no-stress, low-stress, and high-stress) better. Therefore, in spite of their limitations, self-report measures cannot be systematically substituted by unobtrusive measures; this may depend on the outcome of interest. Future research is needed to examine the generalizability of the findings and explore the criterion-related validity of other modalities of stress.
Open AccessA Comparison of Self-Reports and Unobtrusive Multimodal Measures of Stress in a Realistic Simulated Task Environment
Although a popular measurement method in different fields, self-report measures have multiple shortcomings that stem from their very inherent reliance on the participant’s subjective judgment or recall. Given these limitations, in the domain of stress research, unobtrusive measures have been proposed as alternatives. Using a sample of 78 university students who participated in a disaster response simulation in a synthetic task environment, the present study sought to examine whether unobtrusive measures of stress (i.e., physiological, computer interaction, and facial data) can serve as valid substitutes for self-report measures (i.e., state anxiety, task load, and momentary stress). To that end, the study compared the extent to which features from these different modalities accurately (1) classified participants into the stress condition to which they were assigned as well as (2) predicted their performance on a complex synthetic task. The results indicated that the unobtrusive measures individually predicted task performance better than self-report measures, whereas self-report measures predicted the stress condition to which participants were randomly assigned (i.e., no-stress, low-stress, and high-stress) better. Therefore, in spite of their limitations, self-report measures cannot be systematically substituted by unobtrusive measures; this may depend on the outcome of interest. Future research is needed to examine the generalizability of the findings and explore the criterion-related validity of other modalities of stress.
- Open Access
- 25 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Sensor Networks
- Vol. 22 (4) , 1-25
Robust Rainfall Estimation with Multimodal Sensing for Precision Residential Irrigation
Current state-of-the-art residential irrigation systems, such as WaterMyYard, rely on rainfall data from nearby weather stations to adjust irrigation amounts. However, the precision of rainfall data is compromised by the limited spatial resolution of the rain gauges and the significant variability of hyperlocal rainfall, resulting in substantial water waste. To improve irrigation efficiency, we developed a cost-effective irrigation system, dubbed ERIC, which employs machine learning models to estimate rainfall from commodity doorbell camera footage and optimizes irrigation schedules without human intervention. Specifically, we: (a) designed novel visual and audio features with lightweight neural network models to robustly infer rainfall from the camera at the edge, preserving user privacy; (b) built a complete end-to-end irrigation system on Raspberry Pi 4, costing only $75. We deployed the system to five locations (collecting more than 750 hours of video) with diverse environmental backgrounds and varying camera placements. Our realistic evaluation validates that ERIC achieves state-of-the-art rainfall estimation performance (~5 mm/day), saving 9,112 gallons/month of water, translating to $28.56/month in utility savings. Data and code are available at https://github.com/LENSS/ERIC-BuildSys2024.git. 1
- 25 April 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Sensor Networks
- Vol. 22 (4) , 1-25
Open AccessRobust Rainfall Estimation with Multimodal Sensing for Precision Residential Irrigation
Current state-of-the-art residential irrigation systems, such as WaterMyYard, rely on rainfall data from nearby weather stations to adjust irrigation amounts. However, the precision of rainfall data is compromised by the limited spatial resolution of the rain gauges and the significant variability of hyperlocal rainfall, resulting in substantial water waste. To improve irrigation efficiency, we developed a cost-effective irrigation system, dubbed ERIC, which employs machine learning models to estimate rainfall from commodity doorbell camera footage and optimizes irrigation schedules without human intervention. Specifically, we: (a) designed novel visual and audio features with lightweight neural network models to robustly infer rainfall from the camera at the edge, preserving user privacy; (b) built a complete end-to-end irrigation system on Raspberry Pi 4, costing only $75. We deployed the system to five locations (collecting more than 750 hours of video) with diverse environmental backgrounds and varying camera placements. Our realistic evaluation validates that ERIC achieves state-of-the-art rainfall estimation performance (~5 mm/day), saving 9,112 gallons/month of water, translating to $28.56/month in utility savings. Data and code are available at https://github.com/LENSS/ERIC-BuildSys2024.git. 1
- Open Access
- 29 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 132-142
ERIC: Estimating Rainfall with Commodity Doorbell Camera for Precision Residential Irrigation
Current state-of-the-art residential irrigation systems, such as WaterMyYard, rely on rainfall data from nearby weather stations to adjust irrigation amounts. However, the accuracy of rainfall data is compromised by the limited spatial resolution of rain gauges and the significant variability of hyperlocal rainfall, leading to substantial water waste. To improve irrigation efficiency, we developed a cost-effective irrigation system, dubbed ERIC, which employs machine learning models to estimate rainfall from commodity doorbell camera footage and optimizes irrigation schedules without human intervention. Specifically, we: a) designed novel visual and audio features with lightweight neural network models to infer rainfall from the camera at the edge, preserving user privacy; b) built a complete end-to-end irrigation system on Raspberry Pi 4, costing only $75. We deployed the system across five locations (collecting over 750 hours of video) with varying backgrounds and light conditions. Comprehensive evaluation validates that ERIC achieves state-of-the-art rainfall estimation performance (~ 5mm/day), saving 9,112 gallons/month of water, translating to $28.56/month in utility savings. Data and code are available at https://github.com/LENSS/ERIC-BuildSys2024.git.
- 29 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 132-142
Open AccessERIC: Estimating Rainfall with Commodity Doorbell Camera for Precision Residential Irrigation
Current state-of-the-art residential irrigation systems, such as WaterMyYard, rely on rainfall data from nearby weather stations to adjust irrigation amounts. However, the accuracy of rainfall data is compromised by the limited spatial resolution of rain gauges and the significant variability of hyperlocal rainfall, leading to substantial water waste. To improve irrigation efficiency, we developed a cost-effective irrigation system, dubbed ERIC, which employs machine learning models to estimate rainfall from commodity doorbell camera footage and optimizes irrigation schedules without human intervention. Specifically, we: a) designed novel visual and audio features with lightweight neural network models to infer rainfall from the camera at the edge, preserving user privacy; b) built a complete end-to-end irrigation system on Raspberry Pi 4, costing only $75. We deployed the system across five locations (collecting over 750 hours of video) with varying backgrounds and light conditions. Comprehensive evaluation validates that ERIC achieves state-of-the-art rainfall estimation performance (~ 5mm/day), saving 9,112 gallons/month of water, translating to $28.56/month in utility savings. Data and code are available at https://github.com/LENSS/ERIC-BuildSys2024.git.
- Open Access
Robust and Explainable Depression Identification from Speech Using Vowel-Based Ensemble Learning Approaches
This study investigates explainable machine learning algorithms for identifying depression from speech. Grounded in evidence from speech production that depression affects motor control and vowel generation, pre-trained vowel-based embeddings, that integrate semantically meaningful linguistic units, are used. Following that, an ensemble learning approach decomposes the problem into constituent parts characterized by specific depression symptoms and severity levels. Two methods are explored: a "bottom-up" approach with 8 models predicting individual Patient Health Questionnaire-8 (PHQ-8) item scores, and a "top-down" approach using a Mixture of Experts (MoE) with a router module for assessing depression severity. Both methods depict performance comparable to state-of-the-art baselines, demonstrating robustness and reduced susceptibility to dataset mean/median values. System explainability benefits are discussed highlighting their potential to assist clinicians in depression diagnosis and screening.
Open AccessRobust and Explainable Depression Identification from Speech Using Vowel-Based Ensemble Learning Approaches
This study investigates explainable machine learning algorithms for identifying depression from speech. Grounded in evidence from speech production that depression affects motor control and vowel generation, pre-trained vowel-based embeddings, that integrate semantically meaningful linguistic units, are used. Following that, an ensemble learning approach decomposes the problem into constituent parts characterized by specific depression symptoms and severity levels. Two methods are explored: a "bottom-up" approach with 8 models predicting individual Patient Health Questionnaire-8 (PHQ-8) item scores, and a "top-down" approach using a Mixture of Experts (MoE) with a router module for assessing depression severity. Both methods depict performance comparable to state-of-the-art baselines, demonstrating robustness and reduced susceptibility to dataset mean/median values. System explainability benefits are discussed highlighting their potential to assist clinicians in depression diagnosis and screening.
- Open Access
A Pilot Study on Clinician-AI Collaboration in Diagnosing Depression from Speech
This study investigates clinicians' perceptions and attitudes toward an assistive artificial intelligence (AI) system that employs a speech-based explainable ML algorithm for detecting depression. The AI system detects depression from vowel-based spectrotemporal variations of speech and generates explanations through explainable AI (XAI) methods. It further provides decisions and explanations at various temporal granularities, including utterance groups, individual utterances, and within each utterance. A small-scale user study was conducted to evaluate users' perceived usability of the system, trust in the system, and perceptions of design factors associated with several elements of the system. Quantitative and qualitative analysis of the collected data indicates both positive and negative aspects that influence clinicians' perception toward the AI. Results from quantitative analysis indicate that providing more AI explanations enhances user trust but also increases system complexity. Qualitative analysis indicates the potential of integrating such systems into the current diagnostic and screening workflow, but also highlights existing limitations including clinicians' reduced familiarity with AI/ML systems and the need for user-friendly and intuitive visualizations of speech information.
Open AccessA Pilot Study on Clinician-AI Collaboration in Diagnosing Depression from Speech
This study investigates clinicians' perceptions and attitudes toward an assistive artificial intelligence (AI) system that employs a speech-based explainable ML algorithm for detecting depression. The AI system detects depression from vowel-based spectrotemporal variations of speech and generates explanations through explainable AI (XAI) methods. It further provides decisions and explanations at various temporal granularities, including utterance groups, individual utterances, and within each utterance. A small-scale user study was conducted to evaluate users' perceived usability of the system, trust in the system, and perceptions of design factors associated with several elements of the system. Quantitative and qualitative analysis of the collected data indicates both positive and negative aspects that influence clinicians' perception toward the AI. Results from quantitative analysis indicate that providing more AI explanations enhances user trust but also increases system complexity. Qualitative analysis indicates the potential of integrating such systems into the current diagnostic and screening workflow, but also highlights existing limitations including clinicians' reduced familiarity with AI/ML systems and the need for user-friendly and intuitive visualizations of speech information.
- 3 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 14 (2) , 1627-1633
Few-Shot Learning in Emotion Recognition of Spontaneous Speech Using a Siamese Neural Network With Adaptive Sample Pair Formation
Speech-based machine learning (ML) has been heralded as a promising solution for tracking prosodic and spectrotemporal patterns in real-life that are indicative of emotional changes, providing a valuable window into one's cognitive and mental state. Yet, the scarcity of labelled data in ambulatory studies prevents the reliable training of ML models, which usually rely on "data-hungry" distribution-based learning. Leveraging the abundance of labelled speech data from acted emotions, this paper proposes a few-shot learning approach for automatically recognizing emotion in spontaneous speech from a small number of labelled samples. Few-shot learning is implemented via a metric learning approach through a siamese neural network, which models the relative distance between samples rather than relying on learning absolute patterns of the corresponding distributions of each emotion. Results indicate the feasibility of the proposed metric learning in recognizing emotions from spontaneous speech in four datasets, even with a small amount of labelled samples. They further demonstrate superior performance of the proposed metric learning compared to distribution-learning, which is implemented with feedforward neural network fine-tuning. Findings from this work provide a foundation for the ambulatory tracking of human emotion in spontaneous speech contributing to the real-life assessment of mental health degradation outcomes.
- 3 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 14 (2) , 1627-1633
Few-Shot Learning in Emotion Recognition of Spontaneous Speech Using a Siamese Neural Network With Adaptive Sample Pair Formation
Speech-based machine learning (ML) has been heralded as a promising solution for tracking prosodic and spectrotemporal patterns in real-life that are indicative of emotional changes, providing a valuable window into one's cognitive and mental state. Yet, the scarcity of labelled data in ambulatory studies prevents the reliable training of ML models, which usually rely on "data-hungry" distribution-based learning. Leveraging the abundance of labelled speech data from acted emotions, this paper proposes a few-shot learning approach for automatically recognizing emotion in spontaneous speech from a small number of labelled samples. Few-shot learning is implemented via a metric learning approach through a siamese neural network, which models the relative distance between samples rather than relying on learning absolute patterns of the corresponding distributions of each emotion. Results indicate the feasibility of the proposed metric learning in recognizing emotions from spontaneous speech in four datasets, even with a small amount of labelled samples. They further demonstrate superior performance of the proposed metric learning compared to distribution-learning, which is implemented with feedforward neural network fine-tuning. Findings from this work provide a foundation for the ambulatory tracking of human emotion in spontaneous speech contributing to the real-life assessment of mental health degradation outcomes.
- 30 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 13 (3) , 1168-1182
Exploring Individual Differences of Public Speaking Anxiety in Real-Life and Virtual Presentations
Public speaking is a vital skill for making good impressions, effectively exchanging ideas, and influencing others. Yet, public speaking anxiety (PSA) ranks as a top social phobia. Recent advancements in wearable devices and ubiquitous virtual reality (VR) interfaces can help measure and mitigate PSA. This research quantifies PSA through bio-behavioral markers related to individuals' physiological and acoustic characteristics. The effect of virtual reality (VR) training on alleviating PSA is measured through self-reported and bio-behavioral indices. Psychological (e.g., general trait anxiety, personality) and demographic (e.g., age, gender, highest education, native language) traits are examined as moderating factors between bio-behavioral indices and PSA, as well as moderating factors for measuring the VR effectiveness in mitigating PSA. These measures are also used as clustering criteria for stratifying participants in group-based models of PSA. Results indicate the significance of such traits to modeling PSA with the proposed group-based models yielding Spearman's correlation of 0:55 ( $p \lt 0:05$ ) between the actual and predicted outcome. Results further demonstrate that systematic exposure to public speaking in VR can alleviate PSA in terms of both self-reported ( $p \lt 0:05$ ) and physiological ( $p \lt 0:05$ ) indices. Findings from this study will enable researchers to better understand antecedents and causes of PSA and lay the foundation for personalized adaptive feedback for PSA interventions.- 30 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Affective Computing
- Vol. 13 (3) , 1168-1182
Exploring Individual Differences of Public Speaking Anxiety in Real-Life and Virtual Presentations
Public speaking is a vital skill for making good impressions, effectively exchanging ideas, and influencing others. Yet, public speaking anxiety (PSA) ranks as a top social phobia. Recent advancements in wearable devices and ubiquitous virtual reality (VR) interfaces can help measure and mitigate PSA. This research quantifies PSA through bio-behavioral markers related to individuals' physiological and acoustic characteristics. The effect of virtual reality (VR) training on alleviating PSA is measured through self-reported and bio-behavioral indices. Psychological (e.g., general trait anxiety, personality) and demographic (e.g., age, gender, highest education, native language) traits are examined as moderating factors between bio-behavioral indices and PSA, as well as moderating factors for measuring the VR effectiveness in mitigating PSA. These measures are also used as clustering criteria for stratifying participants in group-based models of PSA. Results indicate the significance of such traits to modeling PSA with the proposed group-based models yielding Spearman's correlation of 0:55 ( $p \lt 0:05$ ) between the actual and predicted outcome. Results further demonstrate that systematic exposure to public speaking in VR can alleviate PSA in terms of both self-reported ( $p \lt 0:05$ ) and physiological ( $p \lt 0:05$ ) indices. Findings from this study will enable researchers to better understand antecedents and causes of PSA and lay the foundation for personalized adaptive feedback for PSA interventions.- Open Access
- 28 February 2020
- Published by Frontiers Media SA in Frontiers in Computer Science
- Vol. 2
A Review of Generalizable Transfer Learning in Automatic Emotion Recognition
Automatic emotion recognition is the process of identifying human emotion from signals such as facial expression, speech, and text. Collecting and labelling such signals is often tedious and many times requires expert knowledge. An effective way to address challenges related to the scarcity of data and lack of human labels, is transfer learning. In this manuscript, we will describe fundamental concepts in the field of transfer learning and review work which has successfully applied transfer learning for automatic emotion recognition. We will finally discuss promising future research directions of transfer learning for improving the generalizability of automatic emotion recognition systems.
- 28 February 2020
- Published by Frontiers Media SA in Frontiers in Computer Science
- Vol. 2
Open AccessA Review of Generalizable Transfer Learning in Automatic Emotion Recognition
Automatic emotion recognition is the process of identifying human emotion from signals such as facial expression, speech, and text. Collecting and labelling such signals is often tedious and many times requires expert knowledge. An effective way to address challenges related to the scarcity of data and lack of human labels, is transfer learning. In this manuscript, we will describe fundamental concepts in the field of transfer learning and review work which has successfully applied transfer learning for automatic emotion recognition. We will finally discuss promising future research directions of transfer learning for improving the generalizability of automatic emotion recognition systems.