Despite the global mental health crisis, access to screenings, professionals, and treatments remains high. In collaboration with licensed psychotherapists, we propose a C onversational AI T herapist with psychotherapeutic I nterventions (CaiTI), a platform that leverages large language models (LLMs) and smart devices to enable better mental health self-care. CaiTI can screen the day-to-day functioning using natural and psychotherapeutic conversations. CaiTI leverages reinforcement learning to provide personalized conversation flow. CaiTI can accurately understand and interpret user responses. When the user needs further attention during the conversation, CaiTI can provide conversational psychotherapeutic interventions, including cognitive behavioral therapy and motivational interviewing. Leveraging the datasets prepared by the licensed psychotherapists, we experiment and microbenchmark various LLMs’ performance in tasks along CaiTI’s conversation flow and discuss their strengths and weaknesses. With the psychotherapists, we implement CaiTI and conduct 14-day and 24-week studies. The study results, validated by therapists, demonstrate that CaiTI can converse with users naturally, accurately understand and interpret user responses, and provide psychotherapeutic interventions appropriately and effectively. We showcase the potential of CaiTI LLMs to assist the mental therapy diagnosis and treatment and improve day-to-day functioning screening and precautionary psychotherapeutic intervention systems.

Cardio exercise elevates both heart rate and respiration rate, resulting in distinct physiological changes that affect speech patterns, pitch, breathing sounds, and heart sounds. These variations, which occur post-exercise, are influenced by factors such as exercise intensity and individual fitness levels. A comprehensive audio dataset is critically needed to capture post-exercise physiological changes, as existing datasets focus mainly on resting speech, breathing, and heart sounds, neglecting the dynamic shifts following physical exertion. Current datasets fail to capture unique post-exercise variations like speech disfluencies, altered breathing patterns, and variable heart sound intensities, limiting model generalizability to post-exercise conditions. To address this gap, we recruited 59 subjects from diverse backgrounds to engage in cardio exercise, specifically running, reaching varied exertion levels to produce a rich dataset. Our dataset includes 250 sessions totaling 143 minutes of structured reading, 47 minutes of spontaneous speech, 71 minutes of breathing sounds, and 62.5 minutes of phonocardiogram (PCG) recordings. We designed and deployed preliminary case studies to show that speech changes post-cardio could serve as an indicator of exertion level. We envision this dataset as a foundational resource for designing models in speech and cardiorespiratory monitoring that are resilient to the physiological shifts induced by exercise. This dataset could advance natural language processing (NLP) applications, mobile health, and wearable sensing technologies by enabling resilient and accurate physiological monitoring in real-world conditions.

Representations derived from models such as BERT (Bidirectional Encoder Representations from Transformers) and Hu-BERT (Hidden units BERT), have helped to achieve state-of-the-art performance in dimensional speech emotion recognition. Despite their large dimensionality, and even though these representations are not tailored for emotion recognition tasks, they are frequently used to train large speech emotion models with high memory and computational costs. In this work, we show that there exist lower-dimensional subspaces within the these pre-trained representational spaces that offer a reduction in downstream model complexity without sacrificing performance on emotion estimation. In addition, we model label uncertainty in the form of grader opinion variance, and demonstrate that such information can improve the model’s generalization capacity and robustness. Finally, we compare the robustness of the emotion models against acoustic degradations and observed that the reduced dimensional representations were able to retain the performance similar to the full-dimensional representations without significant regression in dimensional emotion performance.

Over half a trillion recreational runners worldwide engage in running for psychological, health, and social benefits. Running metrics are essential for motivation, goal setting, performance improvement, health management, and injury prevention. Although wearable devices like fitness trackers and smartwatches offer various metrics, they often perform poorly on treadmills and can be uncomfortable or restrictive. In this work, we propose a non-contact, real-time, smartphone-based approach to estimate running metrics, including cadence, ground contact time (GCT), and balance, using the sound produced during treadmill running. In collaboration with a licensed running coach, we recruited over 50 subjects with varying levels of running expertise. We collected treadmill running sounds and ground-truth running metrics in different environments. We designed and developed a multi-task learning (MTL) machine learning mobile system to capture the treadmill running sounds and estimate running metrics in situ. Our proposed method shows comparable accuracy in estimating running metrics to commercial off-the-shelf (COTS) wearable devices.

Kangaroo Mother Care (KMC), involving chest-to-chest skin contact between an infant and caregiver, is proven to be an effective intervention for preterm and full-term infants. Accurate monitoring of KMC duration and infant's vital signs during KMC is clinically important. Existing monitoring methods, however, rely on manual efforts and require rigid sensors or wires/electrodes on the infant's body. We propose Joey, a fabric-based approach to continuously monitor KMC duration and two vital signs essential to an infant's well-being: heart rate and respiration rate. Joey is a soft fabric necklace worn by the caregiver. It leverages the transmission of electrocardiogram (ECG) signals across individuals during skin-to-skin contact. With a minimalist fabric sensor structure, Joey measures KMC duration via the presence of mixed ECG signals. It then isolates the infant's ECG from this mixture with a proposed signal extraction algorithm and employs a diffusion-based denoising model to mitigate motion artifacts, enabling reliable inference of infant's vital signs. We fabricate Joey prototypes with off-the-shelf hardware and evaluate its performance with user studies. Results demonstrate that Joey achieves an average F1 score of 96% for KMC duration measurement, and clinically-acceptable accuracy in infant's vital sign estimation with a mean absolute error of 2.3 beats per minute and 2.9 breaths per minute in estimating heart rate and respiration rate. Clinical interviews further confirm the usability of Joey's sensing fabric for infant skin. A demonstration video of Joey is available at: mobilex.cs.columbia.edu/joey

Kangaroo Mother Care (KMC), involving chest-to-chest skin contact between an infant and caregiver, is proven to be an effective intervention for preterm and full-term infants. Accurate monitoring of KMC duration and infant's vital signs during KMC is clinically important. Existing monitoring methods, however, rely on manual efforts and require rigid sensors or wires/electrodes on the infant's body. We propose Joey, a fabric-based approach to continuously monitor KMC duration and two vital signs essential to an infant's well-being: heart rate and respiration rate. Joey is a soft fabric necklace worn by the caregiver. It leverages the transmission of electrocardiogram (ECG) signals across individuals during skin-to-skin contact. With a minimalist fabric sensor structure, Joey measures KMC duration via the presence of mixed ECG signals. It then isolates the infant's ECG from this mixture with a proposed signal extraction algorithm and employs a diffusion-based denoising model to mitigate motion artifacts, enabling reliable inference of the infant's vital signs. We demonstrate Joey's sensing capability with hand-shaking experiments, showing the real-time mixed ECGs. A demonstration video of Joey for actual KMC practice is available at: mobilex.cs.columbia.edu/joey

Mapping 3D airflow fields is important for many HVAC, industrial, medical, and home applications. However, current approaches are expensive and time-consuming. We present Anemoi, a sub-$100 drone-based system for autonomously mapping 3D airflow fields in indoor environments. Anemoi leverages the effects of airflow on motor control signals to estimate the magnitude and direction of wind at any given point in space. We introduce an exploration algorithm for selecting optimal waypoints that minimize overall airflow estimation uncertainty. We demonstrate through microbenchmarks and real deployments that Anemoi is able to estimate wind speed and direction with errors up to 0.41 m/s and 25.1° lower than the existing state of the art and map 3D airflow fields with an average RMS error of 0.73 m/s.

We present SoFIT, an easily-deployed and privacy-preserving camera network system for occupant tracking. Unlike traditional camera network-based systems, SoFIT does not require a person to calibrate the network or provide real-world references. This enables anyone, including non-professionals, to install SoFIT. Once installed, SoFIT automatically localizes cameras within the network and generates the floor map leveraging movements of people using the space in daily life, before using the floor map and camera locations to track occupants throughout the environment. We demonstrate through a series of deployments that SoFIT can localize cameras with less than 4.8cm error, generate floor maps with 85% similarity to actual floor maps, and track occupants with less than 7.8cm error.

In this demonstration, in collaboration with licensed therapists, we introduce an AI therapist that takes advantage of the smart-home environment to screen day-to-day functioning and infer mental wellness of an occupant. Our system can assess a user's daily functioning and mental wellness based on a combination of direct conversation with users and information obtained from smart home devices using psychological rubrics proposed in [1]. We demonstrate that our system can converse with a user in a natural way (through a smartphone or smart speaker) and analyze a user's response semantically and sentimentally. In addition, we show that our system can provide preliminary interventions to help improve the user's wellness. In particular, when abnormal behavior is detected during the conversation or by smart home devices, the system provides psychotherapeutic consolations during the conversation and will check on the occupant's condition by actuating a home robot.

Vehicle accidents are one of the greatest cause of death and injury in urban areas for pedestrians, workers, and police alike. In this work, we present CSafe, a low power audio-wearable platform that detects, localizes, and provides alerts about oncoming vehicles to improve construction worker safety. Construction worker safety is a much more challenging problem than general urban or pedestrian safety in that the sound of construction tools can be up to orders of magnitude greater than that of vehicles, making vehicle detection and localization exceptionally difficult. To overcome these challenges, we develop a novel sound source separation algorithm, called Probabilistic Template Matching (PTM), as well as a novel noise filtering architecture to remove loud construction noises from our observed signals. We show that our architecture can improve vehicle detection by up to 12% over other state-of-art source separation algorithms. We integrate PTM and our noise filtering architecture into CSafe and show through a series of real-world experiments that CSafe can achieve up to an 82% vehicle detection rate and a 6.90° mean localization error in acoustically noisy construction site scenarios, which is 16% higher and almost 30° lower than the state-of-art audio wearable safety works.

Facial expression monitoring is crucial in fields including mental health care, driver assistant systems, and advertising. However, existing systems typically rely on cameras that capture entire faces, or contact-based bio-signal sensors, which are neither comfortable nor portable. In this demonstration, we present a wireless glasses system for non-contact facial expression monitoring. The system is composed of an IR camera and an embedded processing unit mounted on a 3D-printed glasses frame, and a novel data processing pipeline running across the glasses platform and a computer. Our system performs high-accuracy and real-time facial expression detection with a running time of up to 9 hours. We will show the fully-functioning wearable system in this demonstration.