Sekou L Remy | Clemson University (original) (raw)

Papers by Sekou L Remy

Background: Hypertension is a major risk factor of cardiovascular disease and a leading cause of ... more Background: Hypertension is a major risk factor of cardiovascular disease and a leading cause of morbidity and mortality globally. In Kenya, the rise of hypertension strains an already stretched health care system that has traditionally focused on the management of infectious diseases. Health care provision in this country remains fragmented, and little is known about the role of health information technology in care coordination. Furthermore, there is a dearth of literature on the experiences, challenges, and solutions for improving the management of hypertension and other noncommunicable diseases in the Kenyan private health care sector. Objective: The aim of this study is to assess stakeholders' perspectives on the challenges associated with the management of hypertension in the Kenyan private health care sector and to derive recommendations for the design and functionality of a digital health solution for addressing the care continuity and quality challenges in the management of hypertension. Methods: We conducted a qualitative case study. We collected data using in-depth interviews with 18 care providers and 8 business leads, and direct observations at 18 private health care institutions in Nairobi, Kenya. We analyzed the data thematically to identify the key challenges and recommendations for technology-enabled solutions to support the management of hypertension in the Kenyan private health sector. We subsequently used the generated insights to derive and describe the design and range of functions of a digital health wallet platform for enabling care quality and continuity. Results: The management of hypertension in the Kenyan private health care sector is characterized by challenges such as high cost of care, limited health care literacy, lack of self-management support, ineffective referral systems, inadequate care provider training, and inadequate regulation. Care providers lack the tools needed to understand their patients' care histories and effectively coordinate efforts to deliver high-quality hypertension care. The proposed digital health platform was designed to support hypertension care coordination and continuity through clinical workflow orchestration, decision support, and patient-mediated data sharing with privacy preservation, auditability, and trust enabled by blockchain technology. Conclusions: The Kenyan private health care sector faces key challenges that require significant policy, organizational, and infrastructural changes to ensure care quality and continuity in the management of hypertension. Digital health data interoperability solutions are needed to improve hypertension care coordination in the sector. Additional studies should investigate how patients can control the sharing of their data while ensuring that care providers have a holistic view of the patient during any encounter.

arXiv (Cornell University), Dec 1, 2017

The task of decision-making under uncertainty is daunting, especially for problems which have sig... more The task of decision-making under uncertainty is daunting, especially for problems which have significant complexity. Healthcare policy makers across the globe are facing problems under challenging constraints, with limited tools to help them make data driven decisions. In this work we frame the process of finding an optimal malaria policy as a stochastic multi-armed bandit problem, and implement three agent based strategies to explore the policy space. We apply a Gaussian Process regression to the findings of each agent, both for comparison and to account for stochastic results from simulating the spread of malaria in a fixed population. The generated policy spaces are compared with published results to give a direct reference with human expert decisions for the same simulated population. Our novel approach provides a powerful resource for policy makers, and a platform which can be readily extended to capture future more nuanced policy spaces.

In this paper, we evaluate the performance of feedback control with different publish and subscri... more In this paper, we evaluate the performance of feedback control with different publish and subscribe architectures. Specifically we look to evaluate the difference between ROS, MQTT, and Kafka under the three network conditions. Our results show that latency alone can not be used to assess performance of a real time control system.

In this paper, we describe the system components, present the implemented architecture, and show ... more In this paper, we describe the system components, present the implemented architecture, and show the effect of an interactive learning system in action. We evaluate the system's ability to learning with two datasets, one synthetic and the other from writing samples gathered from human subjects. With both datasets, the respective test and training sets are the same so as to permit the process of interactive learning to be observed as it occurs. At it's core, this learning approach transforms sensory input and actuator output into rank P = 1 spaces, and uses learn a probabilistic mapping between these two “states” to perform the target task. In the future P >1 will be used internally, and we conclude this work with a brief treatment on why we believe this to be a useful trajectory.

As machine learning and data science applications grow ever more prevalent, there is an increased... more As machine learning and data science applications grow ever more prevalent, there is an increased focus on data sharing and open data initiatives, particularly in the context of the African continent. Many argue that data sharing can support research and policy design to alleviate poverty, inequality, and derivative effects in Africa. Despite the fact that the datasets in question are often extracted from African communities, conversations around the challenges of accessing and sharing African data are too often driven by non-African stakeholders. These perspectives frequently employ a deficit narratives, often focusing on lack of education, training, and technological resources in the continent as the leading causes of friction in the data ecosystem. We argue that these narratives obfuscate and distort the full complexity of the African data sharing landscape. In particular, we use storytelling via fictional personas built from a series of interviews with African data experts to complicate dominant narratives and to provide counternarratives. Coupling these personas with research on data practices within the continent, we identify recurring barriers to data sharing as well as inequities in the distribution of data sharing benefits. In particular, we discuss issues arising from power imbalances resulting from the legacies of colonialism, ethno-centrism, and slavery, disinvestment in building trust, lack of acknowledgement of historical and present-day extractive practices, and Western-centric policies that are ill-suited to the African context. After outlining these problems, we discuss avenues for addressing them when sharing data generated in the continent. CCS CONCEPTS • Computing methodologies → Artificial intelligence; • Social and professional topics → Government technology policy.

is a researcher focused on removing barriers to effective use of Robotics and Cloud Computing in ... more is a researcher focused on removing barriers to effective use of Robotics and Cloud Computing in our homes, schools, and training centers. Dr. Remy is currently an Assistant Professor in Human-Centered Computing, and comes to Clemson from the University of Notre Dame where he was a Moreau Postdoctoral Fellow. He also had the pleasure of serving as a part-time instructor in Computer Science at Spelman College. A graduate of the Georgia Institute of Technology (ECE) and Morehouse College (CS), Remy leverages education in both engineering and liberal arts to enable change.

Machine learning techniques have currently been deployed in a number of real-world application ar... more Machine learning techniques have currently been deployed in a number of real-world application areas-from casino surveillance to fingerprint matching. That fact, coupled with advances in computer vision and human-computer interfaces, positions systems that can learn from human observation at the point where they can realistically and reliably be deployed as functional components in autonomous control systems. Healthcare applications though pose a unique challenge in that, although autonomous capability might be available, it might not be desired. And yet, based on recent studies focused on assessment of the changing demographics of the world, there is a need for technology that can deal with the shortcomings envisioned in the workforce. Traditional roles for robotics have focused on repetitive, hazardous or dull tasks. If we take the same stance on healthcare applications, we find that some therapeutic activities fall under this traditional classification due to the long-repetitive nature of the therapist-patient interaction. As such, in this paper, we discuss techniques that can be used to model exercise behavior by observing the patient during therapist-patient interaction. The ultimate goal is to monitor patient performance on repetitive exercises, possibly over the course of multiple days between therapy sessions.

ABSTRACT This work demonstrates improvements in a software tool that provides verbal feedback abo... more ABSTRACT This work demonstrates improvements in a software tool that provides verbal feedback about executed robot code. Designed for programming students with visual impairments, the tool is now multi-lingual and no longer requires locally installed text-to-speech software. These developments use cloud and web standards to provide greater flexibility in generating personalized verbal feedback.

Applications of robotics are quickly changing. Just as computer use evolved from research purpose... more Applications of robotics are quickly changing. Just as computer use evolved from research purposes to everyday functions, applications of robotics are making a transition to mainstream usage. With this change in applications comes a change in the user base of robotics, and there is a pronounced move to reduce the complexity of robotic control. The move to reduce complexity is linked to the separation of the role of robot designer and robot operator. For many target applications, the operator of the robot needs to be able to correct and augment its capabilities. One method to enable this is learning from human data, which has already been successfully applied to robotics. We assert that this learning process is only viable when the demonstrated human behavior is coherent. In this work we test the hypothesis that quantifying the coherence in the provided instruction can provide useful information about the progress of the learning process. We discuss results from the application of this method to reactive behaviors. Such behaviors permit the learning process to be computationally tractable in real-time. These results support the hypothesis that coherence is important for this type of learning and also show that this property can be used to provide an avenue for self regulation of the learning process.

A human-robot hybrid cell is developed for performing assembly in flexible manufacturing in colla... more A human-robot hybrid cell is developed for performing assembly in flexible manufacturing in collaboration between a robot and its human co-worker. Robot trust in human is considered, a computational model for the trust is derived, and a method to measure and display the trust in real-time is developed. The collaborative assembly includes robot-to-human handovers of payloads (assembly tools). A novel trust-based compliant handover motion planning strategy for the robot is derived. The robot varies its handover configuration and motion based on robot trust in human through kinematic redundancy with the aim of reducing potential impulse forces on human body through payload during handover. A comprehensive scheme is developed to evaluate the collaborative assembly including the trust-based handover strategy. The evaluation results show that consideration of robot trust in human during the assembly and adjustment in handover configuration and motion based on robot's trust levels in human significantly improve human-robot interaction and assembly performance through increasing safety, human trust in robot, handover success rate, and the overall assembly efficiency by 20%, 37.58%, 30% and 6.73% respectively and reducing cognitive workload by 25.63%, with a minor reduction in the handover efficiency by 1.87%.

National Conference on Artificial Intelligence, Mar 1, 2009

In this work we present research on three topics which have implications for future robotic appli... more In this work we present research on three topics which have implications for future robotic applications. Couched in learning from human provided examples, we study how robots can demonstrate learning curves akin to those observed for human students. Specifically we show how the parameters of robot learning curves relate to those parameters from learning curves generated by human students. Next we show how these parameters and learning process they represent are affected by the quality of instruction provided. Finally, we present a method to generate an estimate of the robot learning curve. This method is of merit since it is based on properties of interaction that can be extracted as learning occurs.

The recruitment stage in clinical trials is key in ensuring enrollment of a large and diverse num... more The recruitment stage in clinical trials is key in ensuring enrollment of a large and diverse number of participants. Recent trends in clinical trials recruitment strategies have leveraged social media, mobile, and web-based platforms to advertise trials to a broader and more diverse set of potential participants. We develop a method to improve clinical trials enrollment rates through novel models of communication that provide accurate and unbiased information about the clinical trials and provide awareness to target participants. The contributions of this paper are two-fold. First we propose a model to generate abstractive summaries for clinical trials based on sequence to sequence networks with attention policies. Second, we present a preliminary evaluation of the model in terms of learning, vocabulary development, choices of attention policies, and summarization outputs. Finally, we generate a dataset consisting of multi-sentence clinical trials summaries to be used for bench-marking and in future work.

arXiv (Cornell University), Oct 5, 2022

In this work we present a framework which may transform research and praxis in epidemic planning.... more In this work we present a framework which may transform research and praxis in epidemic planning. Introduced in the context of the ongoing COVID-19 pandemic, we provide a concrete demonstration of the way algorithms may learn from epidemiological models to scale their value for epidemic preparedness. Our contributions in this work are two fold: 1) a novel platform which makes it easy for decision making stakeholders to interact with epidemiological models and algorithms developed within the Machine learning community, and 2) the release of this work under the Apache-2.0 License. The objective of this paper is not to look closely at any particular models or algorithms, but instead to highlight how they can be coupled and shared to empower evidence-based decision making.

arXiv (Cornell University), May 6, 2018

PubMed, 2021

The use of epidemiological models for decision-making has been prominent during the COVID-19 pand... more The use of epidemiological models for decision-making has been prominent during the COVID-19 pandemic. Our work presents the application of nonparametric Bayesian techniques for inferring epidemiological model parameters based on available data sets published during the pandemic, towards enabling predictions under uncertainty during emerging pandemics. We present a methodology and framework that allows epidemiological model drivers to be integrated as input into the model calibration process. We demonstrate our methodology using the stringency index and mobility data for COVID-19 on an SEIRD compartmental model for selected US states. Our results directly compare the use of Bayesian nonparametrics for model predictions based on best parameter estimates with results of inference of parameter values across the US states. The proposed methodology provides a framework for What-If analysis and sequential decision-making methods for disease intervention planning and is demonstrated for COVID-19, while also applicable to other infectious disease models.

IEEE Transactions on Learning Technologies, 2012

The robotics field represents the integration of multiple facets of computer science and engineer... more The robotics field represents the integration of multiple facets of computer science and engineering. Robotics-based activities have been shown to encourage K-12 students to consider careers in computing and have even been adopted as part of core computer-science curriculum at a number of universities. Unfortunately, for students with visual impairments, there are still inadequate opportunities made available for teaching basic computing concepts using robotics-based curriculum. This outcome is generally due to the scarcity of accessible interfaces to educational robots and the unfamiliarity of teachers with alternative (e.g., nonvisual) teaching methods. As such, in this paper, we discuss the use of alternative interface modalities to engage students with visual impairments in robotics-based programming activities. We provide an overview of the interaction system and results on a pilot study that engaged nine middle school students with visual impairments during a two-week summer camp.

ABSTRACT As the Web has evolved, web-based capabilities or web services have become a significant... more ABSTRACT As the Web has evolved, web-based capabilities or web services have become a significant aspect of day-to-day routines for businesses and individuals, alike. Interactions with the Web and its services represent a significant portion of overall global power consumption. With the current national emphasis on sustainable resources and energy-efficiency, it is paramount that web processes be efficient in their use of energy. While many studies aim to reduce power consumption in network and computing hardware, our work focuses on models and frameworks to support energy-aware usage of web services (i.e. at the software and information technology (IT) process level). It will not be feasible to rely on measured power consumption when making web services usage decisions; instead predictive energy consumption models are more adequate for real-time decision support. In this paper, we introduce a model that isolates the power consumption of a particular web service within a regular server environment. The model assimilates key factors influencing the power consumption during web services executions such as server hardware characteristic, average CPU load, memory utilization, and hard drive access. Evaluative experiments demonstrate that our model can predict power consumption under varied domain-specific operations and conditions.

We have been involved in the development of several blockchain-based solutions that largely utili... more We have been involved in the development of several blockchain-based solutions that largely utilize workflows. Workflows are used to guide users from independent organizations to process and manage transactions, data and documents in a trusted, immutable, and transparent manner for all relevant entities on a given blockchain network. This work discusses our approach to automate the process of creating, updating, and using workflows for blockchain-based solutions. In particular, we present a workflow definition schema using existing templates. We also show how the workflow definition is used to automate the generation of graphical user interfaces and the possibility of generating associated blockchain smart contracts in the future.

Computation, networking and controls are in the midst of a true transformation. Previously it was... more Computation, networking and controls are in the midst of a true transformation. Previously it was assumed that delays in network communication would inhibit control systems from leveraging networked resources such as remote compute power. Now, industrial and residential applications of the Internet of Things force us to revisit principled deployment and support of distributed sensing and actuation. Web standards provide key resources in the development and support of flexible distributed systems, however the conventional wisdom is that their use adds to the network delay and uncertainty, and makes a difficult control problem intractable. We quantify the effect of an important set of web standards (both formal and informal) in the control of physical systems. The results reported in this paper indicate that although the use of web standards provides an effective path to flexible implementation of distributed control, there are software design choices that impact the observed performance in significant ways.

Springer eBooks, 2013

Cloud computing has the ability to transform simulation by providing access to computation remote... more Cloud computing has the ability to transform simulation by providing access to computation remotely. The transformations are not without cost however. The physics-based simulations required in robotics are sensitive to timing, and given the complexity of the operating environments, there are many reasons for a roboticist to be concerned. In this work we explore the impact of the cloud, web, and networking standards on the control of a simulated robot. Our results show that, on average, there is a noticeable impact on performance, but this impact is not statistically significant in five of the six considered scenarios. These results provide support for efforts that seek to use the cloud to support meaningful simulations. Our results are not globally applicable to robotics simulation. When using cloud-hosted simulations, roboticists yield fine tuned control of the environment, and as such there are some simulations are simply not viable candidates for this treatment.

Background: Hypertension is a major risk factor of cardiovascular disease and a leading cause of ... more Background: Hypertension is a major risk factor of cardiovascular disease and a leading cause of morbidity and mortality globally. In Kenya, the rise of hypertension strains an already stretched health care system that has traditionally focused on the management of infectious diseases. Health care provision in this country remains fragmented, and little is known about the role of health information technology in care coordination. Furthermore, there is a dearth of literature on the experiences, challenges, and solutions for improving the management of hypertension and other noncommunicable diseases in the Kenyan private health care sector. Objective: The aim of this study is to assess stakeholders' perspectives on the challenges associated with the management of hypertension in the Kenyan private health care sector and to derive recommendations for the design and functionality of a digital health solution for addressing the care continuity and quality challenges in the management of hypertension. Methods: We conducted a qualitative case study. We collected data using in-depth interviews with 18 care providers and 8 business leads, and direct observations at 18 private health care institutions in Nairobi, Kenya. We analyzed the data thematically to identify the key challenges and recommendations for technology-enabled solutions to support the management of hypertension in the Kenyan private health sector. We subsequently used the generated insights to derive and describe the design and range of functions of a digital health wallet platform for enabling care quality and continuity. Results: The management of hypertension in the Kenyan private health care sector is characterized by challenges such as high cost of care, limited health care literacy, lack of self-management support, ineffective referral systems, inadequate care provider training, and inadequate regulation. Care providers lack the tools needed to understand their patients' care histories and effectively coordinate efforts to deliver high-quality hypertension care. The proposed digital health platform was designed to support hypertension care coordination and continuity through clinical workflow orchestration, decision support, and patient-mediated data sharing with privacy preservation, auditability, and trust enabled by blockchain technology. Conclusions: The Kenyan private health care sector faces key challenges that require significant policy, organizational, and infrastructural changes to ensure care quality and continuity in the management of hypertension. Digital health data interoperability solutions are needed to improve hypertension care coordination in the sector. Additional studies should investigate how patients can control the sharing of their data while ensuring that care providers have a holistic view of the patient during any encounter.

arXiv (Cornell University), Dec 1, 2017

The task of decision-making under uncertainty is daunting, especially for problems which have sig... more The task of decision-making under uncertainty is daunting, especially for problems which have significant complexity. Healthcare policy makers across the globe are facing problems under challenging constraints, with limited tools to help them make data driven decisions. In this work we frame the process of finding an optimal malaria policy as a stochastic multi-armed bandit problem, and implement three agent based strategies to explore the policy space. We apply a Gaussian Process regression to the findings of each agent, both for comparison and to account for stochastic results from simulating the spread of malaria in a fixed population. The generated policy spaces are compared with published results to give a direct reference with human expert decisions for the same simulated population. Our novel approach provides a powerful resource for policy makers, and a platform which can be readily extended to capture future more nuanced policy spaces.

In this paper, we evaluate the performance of feedback control with different publish and subscri... more In this paper, we evaluate the performance of feedback control with different publish and subscribe architectures. Specifically we look to evaluate the difference between ROS, MQTT, and Kafka under the three network conditions. Our results show that latency alone can not be used to assess performance of a real time control system.

In this paper, we describe the system components, present the implemented architecture, and show ... more In this paper, we describe the system components, present the implemented architecture, and show the effect of an interactive learning system in action. We evaluate the system's ability to learning with two datasets, one synthetic and the other from writing samples gathered from human subjects. With both datasets, the respective test and training sets are the same so as to permit the process of interactive learning to be observed as it occurs. At it's core, this learning approach transforms sensory input and actuator output into rank P = 1 spaces, and uses learn a probabilistic mapping between these two “states” to perform the target task. In the future P >1 will be used internally, and we conclude this work with a brief treatment on why we believe this to be a useful trajectory.

As machine learning and data science applications grow ever more prevalent, there is an increased... more As machine learning and data science applications grow ever more prevalent, there is an increased focus on data sharing and open data initiatives, particularly in the context of the African continent. Many argue that data sharing can support research and policy design to alleviate poverty, inequality, and derivative effects in Africa. Despite the fact that the datasets in question are often extracted from African communities, conversations around the challenges of accessing and sharing African data are too often driven by non-African stakeholders. These perspectives frequently employ a deficit narratives, often focusing on lack of education, training, and technological resources in the continent as the leading causes of friction in the data ecosystem. We argue that these narratives obfuscate and distort the full complexity of the African data sharing landscape. In particular, we use storytelling via fictional personas built from a series of interviews with African data experts to complicate dominant narratives and to provide counternarratives. Coupling these personas with research on data practices within the continent, we identify recurring barriers to data sharing as well as inequities in the distribution of data sharing benefits. In particular, we discuss issues arising from power imbalances resulting from the legacies of colonialism, ethno-centrism, and slavery, disinvestment in building trust, lack of acknowledgement of historical and present-day extractive practices, and Western-centric policies that are ill-suited to the African context. After outlining these problems, we discuss avenues for addressing them when sharing data generated in the continent. CCS CONCEPTS • Computing methodologies → Artificial intelligence; • Social and professional topics → Government technology policy.

is a researcher focused on removing barriers to effective use of Robotics and Cloud Computing in ... more is a researcher focused on removing barriers to effective use of Robotics and Cloud Computing in our homes, schools, and training centers. Dr. Remy is currently an Assistant Professor in Human-Centered Computing, and comes to Clemson from the University of Notre Dame where he was a Moreau Postdoctoral Fellow. He also had the pleasure of serving as a part-time instructor in Computer Science at Spelman College. A graduate of the Georgia Institute of Technology (ECE) and Morehouse College (CS), Remy leverages education in both engineering and liberal arts to enable change.

Machine learning techniques have currently been deployed in a number of real-world application ar... more Machine learning techniques have currently been deployed in a number of real-world application areas-from casino surveillance to fingerprint matching. That fact, coupled with advances in computer vision and human-computer interfaces, positions systems that can learn from human observation at the point where they can realistically and reliably be deployed as functional components in autonomous control systems. Healthcare applications though pose a unique challenge in that, although autonomous capability might be available, it might not be desired. And yet, based on recent studies focused on assessment of the changing demographics of the world, there is a need for technology that can deal with the shortcomings envisioned in the workforce. Traditional roles for robotics have focused on repetitive, hazardous or dull tasks. If we take the same stance on healthcare applications, we find that some therapeutic activities fall under this traditional classification due to the long-repetitive nature of the therapist-patient interaction. As such, in this paper, we discuss techniques that can be used to model exercise behavior by observing the patient during therapist-patient interaction. The ultimate goal is to monitor patient performance on repetitive exercises, possibly over the course of multiple days between therapy sessions.

ABSTRACT This work demonstrates improvements in a software tool that provides verbal feedback abo... more ABSTRACT This work demonstrates improvements in a software tool that provides verbal feedback about executed robot code. Designed for programming students with visual impairments, the tool is now multi-lingual and no longer requires locally installed text-to-speech software. These developments use cloud and web standards to provide greater flexibility in generating personalized verbal feedback.

Applications of robotics are quickly changing. Just as computer use evolved from research purpose... more Applications of robotics are quickly changing. Just as computer use evolved from research purposes to everyday functions, applications of robotics are making a transition to mainstream usage. With this change in applications comes a change in the user base of robotics, and there is a pronounced move to reduce the complexity of robotic control. The move to reduce complexity is linked to the separation of the role of robot designer and robot operator. For many target applications, the operator of the robot needs to be able to correct and augment its capabilities. One method to enable this is learning from human data, which has already been successfully applied to robotics. We assert that this learning process is only viable when the demonstrated human behavior is coherent. In this work we test the hypothesis that quantifying the coherence in the provided instruction can provide useful information about the progress of the learning process. We discuss results from the application of this method to reactive behaviors. Such behaviors permit the learning process to be computationally tractable in real-time. These results support the hypothesis that coherence is important for this type of learning and also show that this property can be used to provide an avenue for self regulation of the learning process.

A human-robot hybrid cell is developed for performing assembly in flexible manufacturing in colla... more A human-robot hybrid cell is developed for performing assembly in flexible manufacturing in collaboration between a robot and its human co-worker. Robot trust in human is considered, a computational model for the trust is derived, and a method to measure and display the trust in real-time is developed. The collaborative assembly includes robot-to-human handovers of payloads (assembly tools). A novel trust-based compliant handover motion planning strategy for the robot is derived. The robot varies its handover configuration and motion based on robot trust in human through kinematic redundancy with the aim of reducing potential impulse forces on human body through payload during handover. A comprehensive scheme is developed to evaluate the collaborative assembly including the trust-based handover strategy. The evaluation results show that consideration of robot trust in human during the assembly and adjustment in handover configuration and motion based on robot's trust levels in human significantly improve human-robot interaction and assembly performance through increasing safety, human trust in robot, handover success rate, and the overall assembly efficiency by 20%, 37.58%, 30% and 6.73% respectively and reducing cognitive workload by 25.63%, with a minor reduction in the handover efficiency by 1.87%.

National Conference on Artificial Intelligence, Mar 1, 2009

In this work we present research on three topics which have implications for future robotic appli... more In this work we present research on three topics which have implications for future robotic applications. Couched in learning from human provided examples, we study how robots can demonstrate learning curves akin to those observed for human students. Specifically we show how the parameters of robot learning curves relate to those parameters from learning curves generated by human students. Next we show how these parameters and learning process they represent are affected by the quality of instruction provided. Finally, we present a method to generate an estimate of the robot learning curve. This method is of merit since it is based on properties of interaction that can be extracted as learning occurs.

The recruitment stage in clinical trials is key in ensuring enrollment of a large and diverse num... more The recruitment stage in clinical trials is key in ensuring enrollment of a large and diverse number of participants. Recent trends in clinical trials recruitment strategies have leveraged social media, mobile, and web-based platforms to advertise trials to a broader and more diverse set of potential participants. We develop a method to improve clinical trials enrollment rates through novel models of communication that provide accurate and unbiased information about the clinical trials and provide awareness to target participants. The contributions of this paper are two-fold. First we propose a model to generate abstractive summaries for clinical trials based on sequence to sequence networks with attention policies. Second, we present a preliminary evaluation of the model in terms of learning, vocabulary development, choices of attention policies, and summarization outputs. Finally, we generate a dataset consisting of multi-sentence clinical trials summaries to be used for bench-marking and in future work.

arXiv (Cornell University), Oct 5, 2022

In this work we present a framework which may transform research and praxis in epidemic planning.... more In this work we present a framework which may transform research and praxis in epidemic planning. Introduced in the context of the ongoing COVID-19 pandemic, we provide a concrete demonstration of the way algorithms may learn from epidemiological models to scale their value for epidemic preparedness. Our contributions in this work are two fold: 1) a novel platform which makes it easy for decision making stakeholders to interact with epidemiological models and algorithms developed within the Machine learning community, and 2) the release of this work under the Apache-2.0 License. The objective of this paper is not to look closely at any particular models or algorithms, but instead to highlight how they can be coupled and shared to empower evidence-based decision making.

arXiv (Cornell University), May 6, 2018

PubMed, 2021

The use of epidemiological models for decision-making has been prominent during the COVID-19 pand... more The use of epidemiological models for decision-making has been prominent during the COVID-19 pandemic. Our work presents the application of nonparametric Bayesian techniques for inferring epidemiological model parameters based on available data sets published during the pandemic, towards enabling predictions under uncertainty during emerging pandemics. We present a methodology and framework that allows epidemiological model drivers to be integrated as input into the model calibration process. We demonstrate our methodology using the stringency index and mobility data for COVID-19 on an SEIRD compartmental model for selected US states. Our results directly compare the use of Bayesian nonparametrics for model predictions based on best parameter estimates with results of inference of parameter values across the US states. The proposed methodology provides a framework for What-If analysis and sequential decision-making methods for disease intervention planning and is demonstrated for COVID-19, while also applicable to other infectious disease models.

IEEE Transactions on Learning Technologies, 2012

The robotics field represents the integration of multiple facets of computer science and engineer... more The robotics field represents the integration of multiple facets of computer science and engineering. Robotics-based activities have been shown to encourage K-12 students to consider careers in computing and have even been adopted as part of core computer-science curriculum at a number of universities. Unfortunately, for students with visual impairments, there are still inadequate opportunities made available for teaching basic computing concepts using robotics-based curriculum. This outcome is generally due to the scarcity of accessible interfaces to educational robots and the unfamiliarity of teachers with alternative (e.g., nonvisual) teaching methods. As such, in this paper, we discuss the use of alternative interface modalities to engage students with visual impairments in robotics-based programming activities. We provide an overview of the interaction system and results on a pilot study that engaged nine middle school students with visual impairments during a two-week summer camp.

ABSTRACT As the Web has evolved, web-based capabilities or web services have become a significant... more ABSTRACT As the Web has evolved, web-based capabilities or web services have become a significant aspect of day-to-day routines for businesses and individuals, alike. Interactions with the Web and its services represent a significant portion of overall global power consumption. With the current national emphasis on sustainable resources and energy-efficiency, it is paramount that web processes be efficient in their use of energy. While many studies aim to reduce power consumption in network and computing hardware, our work focuses on models and frameworks to support energy-aware usage of web services (i.e. at the software and information technology (IT) process level). It will not be feasible to rely on measured power consumption when making web services usage decisions; instead predictive energy consumption models are more adequate for real-time decision support. In this paper, we introduce a model that isolates the power consumption of a particular web service within a regular server environment. The model assimilates key factors influencing the power consumption during web services executions such as server hardware characteristic, average CPU load, memory utilization, and hard drive access. Evaluative experiments demonstrate that our model can predict power consumption under varied domain-specific operations and conditions.

We have been involved in the development of several blockchain-based solutions that largely utili... more We have been involved in the development of several blockchain-based solutions that largely utilize workflows. Workflows are used to guide users from independent organizations to process and manage transactions, data and documents in a trusted, immutable, and transparent manner for all relevant entities on a given blockchain network. This work discusses our approach to automate the process of creating, updating, and using workflows for blockchain-based solutions. In particular, we present a workflow definition schema using existing templates. We also show how the workflow definition is used to automate the generation of graphical user interfaces and the possibility of generating associated blockchain smart contracts in the future.

Computation, networking and controls are in the midst of a true transformation. Previously it was... more Computation, networking and controls are in the midst of a true transformation. Previously it was assumed that delays in network communication would inhibit control systems from leveraging networked resources such as remote compute power. Now, industrial and residential applications of the Internet of Things force us to revisit principled deployment and support of distributed sensing and actuation. Web standards provide key resources in the development and support of flexible distributed systems, however the conventional wisdom is that their use adds to the network delay and uncertainty, and makes a difficult control problem intractable. We quantify the effect of an important set of web standards (both formal and informal) in the control of physical systems. The results reported in this paper indicate that although the use of web standards provides an effective path to flexible implementation of distributed control, there are software design choices that impact the observed performance in significant ways.

Springer eBooks, 2013

Cloud computing has the ability to transform simulation by providing access to computation remote... more Cloud computing has the ability to transform simulation by providing access to computation remotely. The transformations are not without cost however. The physics-based simulations required in robotics are sensitive to timing, and given the complexity of the operating environments, there are many reasons for a roboticist to be concerned. In this work we explore the impact of the cloud, web, and networking standards on the control of a simulated robot. Our results show that, on average, there is a noticeable impact on performance, but this impact is not statistically significant in five of the six considered scenarios. These results provide support for efforts that seek to use the cloud to support meaningful simulations. Our results are not globally applicable to robotics simulation. When using cloud-hosted simulations, roboticists yield fine tuned control of the environment, and as such there are some simulations are simply not viable candidates for this treatment.