Goto

Collaborating Authors

Results


On the Opportunities and Risks of Foundation Models

arXiv.org Artificial Intelligence

AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.


"Brilliant AI Doctor" in Rural China: Tensions and Challenges in AI-Powered CDSS Deployment

arXiv.org Artificial Intelligence

Artificial intelligence (AI) technology has been increasingly used in the implementation of advanced Clinical Decision Support Systems (CDSS). Research demonstrated the potential usefulness of AI-powered CDSS (AI-CDSS) in clinical decision making scenarios. However, post-adoption user perception and experience remain understudied, especially in developing countries. Through observations and interviews with 22 clinicians from 6 rural clinics in China, this paper reports the various tensions between the design of an AI-CDSS system ("Brilliant Doctor") and the rural clinical context, such as the misalignment with local context and workflow, the technical limitations and usability barriers, as well as issues related to transparency and trustworthiness of AI-CDSS. Despite these tensions, all participants expressed positive attitudes toward the future of AI-CDSS, especially acting as "a doctor's AI assistant" to realize a Human-AI Collaboration future in clinical settings. Finally we draw on our findings to discuss implications for designing AI-CDSS interventions for rural clinical contexts in developing countries.


Wound and episode level readmission risk or weeks to readmit: Why do patients get readmitted? How long does it take for a patient to get readmitted?

arXiv.org Machine Learning

The Affordable care Act of 2010 had introduced Readmission reduction program in 2012 to reduce avoidable re-admissions to control rising healthcare costs. Wound care impacts 15 of medicare beneficiaries making it one of the major contributors of medicare health care cost. Health plans have been exploring proactive health care services that can focus on preventing wound recurrences and re-admissions to control the wound care costs. With rising costs of Wound care industry, it has become of paramount importance to reduce wound recurrences & patient re-admissions. What factors are responsible for a Wound to recur which ultimately lead to hospitalization or re-admission? Is there a way to identify the patients at risk of re-admission before the occurrence using data driven analysis? Patient re-admission risk management has become critical for patients suffering from chronic wounds such as diabetic ulcers, pressure ulcers, and vascular ulcers. Understanding the risk & the factors that cause patient readmission can help care providers and patients avoid wound recurrences. Our work focuses on identifying patients who are at high risk of re-admission & determining the time period with in which a patient might get re-admitted. Frequent re-admissions add financial stress to the patient & Health plan and deteriorate the quality of life of the patient. Having this information can allow a provider to set up preventive measures that can delay, if not prevent, patients' re-admission. On a combined wound & episode-level data set of patient's wound care information, our extended autoprognosis achieves a recall of 92 and a precision of 92 for the predicting a patient's re-admission risk. For new patient class, precision and recall are as high as 91 and 98, respectively. We are also able to predict the patient's discharge event for a re-admission event to occur through our model with a MAE of 2.3 weeks.


Ethical Machine Learning in Health Care

arXiv.org Artificial Intelligence

The use of machine learning (ML) in health care raises numerous ethical concerns, especially as models can amplify existing health inequities. Here, we outline ethical considerations for equitable ML in the advancement of health care. Specifically, we frame ethics of ML in health care through the lens of social justice. We describe ongoing efforts and outline challenges in a proposed pipeline of ethical ML in health, ranging from problem selection to post-deployment considerations. We close by summarizing recommendations to address these challenges.


Precision Health Data: Requirements, Challenges and Existing Techniques for Data Security and Privacy

arXiv.org Artificial Intelligence

Precision health leverages information from various sources, including omics, lifestyle, environment, social media, medical records, and medical insurance claims to enable personalized care, prevent and predict illness, and precise treatments. It extensively uses sensing technologies (e.g., electronic health monitoring devices), computations (e.g., machine learning), and communication (e.g., interaction between the health data centers). As health data contain sensitive private information, including the identity of patient and carer and medical conditions of the patient, proper care is required at all times. Leakage of these private information affects the personal life, including bullying, high insurance premium, and loss of job due to the medical history. Thus, the security, privacy of and trust on the information are of utmost importance. Moreover, government legislation and ethics committees demand the security and privacy of healthcare data. Herein, in the light of precision health data security, privacy, ethical and regulatory requirements, finding the best methods and techniques for the utilization of the health data, and thus precision health is essential. In this regard, firstly, this paper explores the regulations, ethical guidelines around the world, and domain-specific needs. Then it presents the requirements and investigates the associated challenges. Secondly, this paper investigates secure and privacy-preserving machine learning methods suitable for the computation of precision health data along with their usage in relevant health projects. Finally, it illustrates the best available techniques for precision health data security and privacy with a conceptual system model that enables compliance, ethics clearance, consent management, medical innovations, and developments in the health domain.


Opinion: AI needs patients' voices in order to revolutionize health care

#artificialintelligence

"Listen to your patient; they are telling you the diagnosis," an aphorism attributed to Dr. William Osler, the founder of modern medicine, still holds true today. The disappearance of patients' stories from electronic health records could be one reason that artificial intelligence and machine learning have so far failed to deliver their promised revolution of health care. The medical industry's fascination with artificial intelligence is understandable. Advancements in medicine have dramatically improved patient outcomes, and there is every reason to believe that machine learning, deep learning, artificial intelligence, and the like will do the same. But before we jump on the AI bandwagon, I offer this caution: consider the source of the data it is dependent on.


Amazon Textract is now HIPAA eligible Amazon Web Services

#artificialintelligence

Today, Amazon Web Services (AWS) announced that Amazon Textract, a machine learning service that quickly and easily extracts text and data from forms and tables in scanned documents, is now eligible for healthcare and life science workloads that require HIPAA compliance. This launch builds upon the existing portfolio of AWS artificial intelligence services that are HIPAA-eligible, including Amazon Translate, Amazon Comprehend, Amazon Transcribe, Amazon Polly, Amazon SageMaker and Amazon Rekognition – that help customers retrieve data from documents more accurately to reach better healthcare decisions, operate more efficiently, and help identify medical and scientific trends. Critical healthcare information often lies within documents such as medical records and forms. Healthcare and life science organizations need to access data that is locked inside those documents in order to fulfil medical claims, streamline administrative processes, and process electronic health records. They routinely extract text and data from documents through manual data entry or simple optical character recognition (OCR) software.


Artificial intelligence needs patients' voice to remake health care - STAT

#artificialintelligence

"Listen to your patient; they are telling you the diagnosis," an aphorism attributed to Dr. William Osler, the founder of modern medicine, still holds true today. The disappearance of patients' stories from electronic health records could be one reason that artificial intelligence and machine learning have so far failed to deliver their promised revolution of health care. The medical industry's fascination with artificial intelligence is understandable. Advancements in medicine have dramatically improved patient outcomes, and there is every reason to believe that machine learning, deep learning, artificial intelligence, and the like will do the same. But before we jump on the AI bandwagon, I offer this caution: consider the source of the data it is dependent on.


Extract and visualize clinical entities using Amazon Comprehend Medical Amazon Web Services

#artificialintelligence

Amazon Comprehend Medical is a new HIPAA-eligible service that uses machine learning (ML) to extract medical information with high accuracy. This reduces the cost, time, and effort of processing large amounts of unstructured medical text. You can extract entities and relationships like medication, diagnosis, and dosage, and you can also extract protected health information (PHI). Using Amazon Comprehend Medical allows end users to get value from raw clinical notes that is otherwise largely unused for analytical purposes because it's difficult to parse. There is immense value associated with extracting information from these notes and integrating it with other medical systems like an Electronic Health Record (EHR) and a Clinical Trial Management System (CTMS).


Deep EHR: Chronic Disease Prediction Using Medical Notes

arXiv.org Machine Learning

Early detection of preventable diseases is important for better disease management, improved inter-ventions, and more efficient health-care resource allocation. Various machine learning approacheshave been developed to utilize information in Electronic Health Record (EHR) for this task. Majorityof previous attempts, however, focus on structured fields and lose the vast amount of information inthe unstructured notes. In this work we propose a general multi-task framework for disease onsetprediction that combines both free-text medical notes and structured information. We compareperformance of different deep learning architectures including CNN, LSTM and hierarchical models.In contrast to traditional text-based prediction models, our approach does not require disease specificfeature engineering, and can handle negations and numerical values that exist in the text. Ourresults on a cohort of about 1 million patients show that models using text outperform modelsusing just structured data, and that models capable of using numerical values and negations in thetext, in addition to the raw text, further improve performance. Additionally, we compare differentvisualization methods for medical professionals to interpret model predictions.