Rebecca Adaimi

Ms. Rebecca Adaimi

University of Texas at Austin, Austin, Texas, USA
orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 September 2022 in Sensors
Verified
Journal Article
Open access
PDF file
RARebecca Adaimi
Rebecca Adaimi
ETEdison Thomaz
Edison Thomaz
Continual learning (CL), also known as lifelong learning, is an emerging research topic that has been attracting increasing interest in the field of machine learning. With human activity recognition (HAR) playing a key role in enabling numerous real-world applications, an essential step towards the long-term deployment of such systems is to extend the activity model to dynamically adapt to changes in people’s everyday behavior. Current research in CL applied to the HAR domain is still under-explored with researchers exploring existing methods developed for computer vision in HAR. Moreover, analysis has so far focused on task-incremental or class-incremental learning paradigms where task boundaries are known. This impedes the applicability of such methods for real-world systems. To push this field forward, we build on recent advances in the area of continual learning and design a lifelong adaptive learning framework using Prototypical Networks, LAPNet-HAR, that processes sensor-based data streams in a task-free data-incremental fashion and mitigates catastrophic forgetting using experience replay and continual prototype adaptation. Online learning is further facilitated using contrastive loss to enforce inter-class separation. LAPNet-HAR is evaluated on five publicly available activity datasets in terms of its ability to acquire new information while preserving previous knowledge. Our extensive empirical results demonstrate the effectiveness of LAPNet-HAR in task-free CL and uncover useful insights for future challenges.

17 March 2019
Verified
Conference Paper
Open access
KCKeum San Chun
Keum San Chun
ASAshley B. Sanders
Ashley B. Sanders
RARebecca Adaimi
Rebecca Adaimi
NSNecole Streeper
Necole Streeper
DCDavid E. Conroy
David E. Conroy
ETEdison Thomaz
Edison Thomaz
Over the last decade, advances in mobile technologies have enabled the development of intelligent systems that attempt to recognize and model a variety of health-related human behaviors. While automated dietary monitoring based on passive sensors has been an area of increasing research activity for many years, much less attention has been given to tracking fluid intake. In this work, we apply an adaptive segmentation technique on a continuous stream of inertial data captured with a practical, off-the-shelf wrist-mounted device to detect fluid intake gestures passively. We evaluated our approach in a study with 30 participants where 561 drinking instances were recorded. Using a leave-one-participant-out (LOPO), we were able to detect drinking episodes with 90.3% precision and 91.0% recall, demonstrating the generalizability of our approach. In addition to our proposed method, we also contribute an anonymized and labeled dataset of drinking and non-drinking gestures to encourage further work in the field.

11 September 2022
Verified
Conference Paper
Open access
DLDawei Liang
Dawei Liang
RARebecca Adaimi
Rebecca Adaimi
RARebecca Adaimi
Rebecca Adaimi
RMRadu Marculescu
Radu Marculescu
ETEdison Thomaz
Edison Thomaz
Modern commercial wearable devices are widely equipped with inertial measurement units (IMU) and microphones. The motion and audio signals captured by these sensors can be used for recognizing a variety of user physical activities. Compared to motion data, audio data contains rich contextual information of human activities, but continuous audio sensing also poses extra data sampling burdens and privacy issues. Given such challenges, this paper studies a novel approach to augment IMU models for human activity recognition (HAR) with the superior acoustic knowledge of activities. Specifically, we propose a teacher-student framework to derive an IMU-based HAR model. Instead of training with motion data alone, an advanced audio-based teacher model is incorporated to guide the student HAR model. Once trained, the HAR model only takes as inputs motion data for inference. Based on a semi-controlled study with 15 participants, we show that an IMU model augmented with the proposed framework outperforms the original baseline model without augmentation (74.4% versus 70.0% accuracy) for recognizing 23 activities of daily living. We further discuss a few insights regarding the difference of model performance with and without our framework and possible trade-offs for actual deployment.

27 October 2025 in JMIR mHealth and uHealth
Verified
Journal Article
Open access
RARebecca Adaimi
Rebecca Adaimi
NTNina Thigpen
Nina Thigpen
ACAlicia Clausel
Alicia Clausel
NGNeta Gotlieb
Neta Gotlieb
KPKetan Patel
Ketan Patel
Massimiliano de ZambottiMassimiliano de Zambotti
Massimiliano de Zambotti
Background: Pregnancy and childbirth involve significant health challenges, including preventable maternal deaths, severe complications, and disparities tied to social determinants, emphasizing the need for improved maternal care. Pregnancy could benefit from a more comprehensive, continuous care model that captures dynamic changes and enhances maternal-fetal outcomes. Objective: This large-scale, real-world, high-density study aims to use wearable data to investigate maternal biobehavioral trajectories for pregnancies leading to loss, preterm, and term births, exploring how demographic factors like age and body mass index (BMI) affect these trajectories. Methods: Retrospective observational analysis of pregnancies from a sample of 10,318 and 18- to 51-year-old female Oura Ring users (324 preterm births, 5039 term births, 4955 pregnancies ending in loss before 20 weeks of gestation). Oura biobehavioral data were analyzed across a 64-week window encompassing 8 weeks prepregnancy, through pregnancy, and post partum, via generalized estimating equation (GEE) statistical modeling. Results: Gestational age emerged as a significant factor across all domains among term pregnancies (P<.001). During the first trimester, participants experienced marked sleep changes, peaking around week 9 and characterized by more time in bed (+30 min), asleep (+15 min), and awake (+15 min) compared with prepregnancy. Metrics declined and stabilized in the second trimester; by the third trimester, time in bed returned to baseline, while sleep remained reduced and wakefulness elevated. At birth, time in bed and wakefulness peaked, and sleep duration reached its minimum, with nighttime wake exceeding 3 SDs above baseline. Temperature changes were more pronounced, sustained, and occurred earlier than sleep changes—becoming evident by week 4, peaking at +0.3 °C above baseline by week 9, and showing a steady decline until birth. A secondary, modest increase (+0.1 °C) was observed near birth, followed by a decline postpartum. Heart rate (HR) increased steadily, peaking at +10 bpm above baseline at week 32, while HR variability declined by >15 milliseconds in a mirrored pattern. Respiratory rate peaked around week 9 and declined thereafter. Step count declined in the first trimester, with a ≈2000-step reduction at around week 8. After a slight rebound midpregnancy, activity declined again, reaching its lowest point near birth, with >2500 fewer steps than prepregnancy. Age and BMI showed significant but modest interaction effects (all P<.01). In pregnancies ending in loss, deviations emerged up to 2 weeks prior. Time in bed decreased starting ≈2 weeks before loss (P<.001), followed by reductions in sleep duration (P<.001), temperature trends (P<.001), respiratory rate (P=.019), HR (P=.005), and awake time (P=.033). Conclusions: These findings highlight the complex dynamics underlying changes in participants’ sleep, temperature trends, cardiorespiratory, and activity data throughout pregnancy, involving extensive adaptations. A deeper focus on normative changes can advance maternal-fetal medicine, improve clinical outcomes, and address scientific gaps.

5 August 2022 in Preprints.org
Verified
Preprint
Open access
PDF file
RARebecca Adaimi
Rebecca Adaimi
ETEdison Thomaz
Edison Thomaz
Continual learning (CL), a.k.a lifelong learning, is an emerging research topic that has been attracting increasing interest in the field of machine learning. With human activity recognition (HAR) playing a key role in enabling numerous real-world applications, an essential step towards the long-term deployment of such systems is to extend the activity model to dynamically adapt to changes in people’s everyday behavior. Current research in CL applied to HAR domain is still under-explored with researchers exploring existing methods developed for computer vision in HAR. Moreover, analysis has so far focused on task-incremental or class-incremental learning paradigms where task boundaries are known. This impedes the applicability of such methods for real-world systems. To push this field forward, we build on recent advances in the area of continual learning and design a lifelong adaptive learning framework using Prototypical Networks, LAPNet-HAR, that processes sensor-based data streams in a task-free data-incremental fashion and mitigates catastrophic forgetting using experience replay and continual prototype adaptation. Online learning is further facilitated using contrastive loss to enforce inter-class separation. LAPNet-HAR is evaluated on 5 publicly available activity datasets in terms of its ability to acquire new information while preserving previous knowledge. Our extensive empirical results demonstrate the effectiveness of LAPNet-HAR in task-free CL and uncover useful insights for future challenges.

8 July 2025 in JMIR Preprints
Verified
Preprint
Open access
RARebecca Adaimi
Rebecca Adaimi
NTNina Thigpen
Nina Thigpen
ACAlicia Clausel
Alicia Clausel
NGNeta Gotlieb
Neta Gotlieb
KPKetan Patel
Ketan Patel
Massimiliano de ZambottiMassimiliano de Zambotti
Massimiliano de Zambotti
BACKGROUND Pregnancy and childbirth involve significant health challenges, including preventable maternal deaths, severe complications, and disparities tied to social determinants, emphasizing the need for improved maternal care. Pregnancy could benefit from a more comprehensive, continuous care model that captures dynamic changes and enhances maternal-fetal outcomes. OBJECTIVE This large-scale, real-world, high-density study aims to use wearable data to investigate maternal biobehavioral trajectories for pregnancies leading to loss, preterm, and term births, exploring how demographic factors like age and body mass index (BMI) affect these trajectories. METHODS Retrospective observational analysis of pregnancies from a sample of 10,318 and 18- to 51-year-old female Oura Ring users (324 preterm births, 5039 term births, 4955 pregnancies ending in loss before 20 weeks of gestation). Oura biobehavioral data were analyzed across a 64-week window encompassing 8 weeks prepregnancy, through pregnancy, and post partum, via generalized estimating equation (GEE) statistical modeling. RESULTS Gestational age emerged as a significant factor across all domains among term pregnancies (P<.001). During the first trimester, participants experienced marked sleep changes, peaking around week 9 and characterized by more time in bed (+30 min), asleep (+15 min), and awake (+15 min) compared with prepregnancy. Metrics declined and stabilized in the second trimester; by the third trimester, time in bed returned to baseline, while sleep remained reduced and wakefulness elevated. At birth, time in bed and wakefulness peaked, and sleep duration reached its minimum, with nighttime wake exceeding 3 SDs above baseline. Temperature changes were more pronounced, sustained, and occurred earlier than sleep changes—becoming evident by week 4, peaking at +0.3 °C above baseline by week 9, and showing a steady decline until birth. A secondary, modest increase (+0.1 °C) was observed near birth, followed by a decline postpartum. Heart rate (HR) increased steadily, peaking at +10 bpm above baseline at week 32, while HR variability declined by >15 milliseconds in a mirrored pattern. Respiratory rate peaked around week 9 and declined thereafter. Step count declined in the first trimester, with a ≈2000-step reduction at around week 8. After a slight rebound midpregnancy, activity declined again, reaching its lowest point near birth, with >2500 fewer steps than prepregnancy. Age and BMI showed significant but modest interaction effects (all P<.01). In pregnancies ending in loss, deviations emerged up to 2 weeks prior. Time in bed decreased starting ≈2 weeks before loss (P<.001), followed by reductions in sleep duration (P<.001), temperature trends (P<.001), respiratory rate (P=.019), HR (P=.005), and awake time (P=.033). CONCLUSIONS These findings highlight the complex dynamics underlying changes in participants’ sleep, temperature trends, cardiorespiratory, and activity data throughout pregnancy, involving extensive adaptations. A deeper focus on normative changes can advance maternal-fetal medicine, improve clinical outcomes, and address scientific gaps. CLINICALTRIAL