Artificial Intelligence Analysis of Diabetic Retinopathy in Retinal Screening Events

Vought, Rita; Vought, Victoria; Shah, Megh; Szirth, Bernard; Bhagat, Neelakshi

doi:10.21203/rs.3.rs-2562995/v1

Cited by 5 publications

(5 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Establishing guidelines for ensuring AI trustworthiness remains an open question within the community, and various discussions on this matter can be found elsewhere. 30–32 Addressing these concerns is crucial to harness the full potential of LLMs while ensuring their responsible and reliable usage in regulatory research.…”

Section: Discussionmentioning

confidence: 99%

RxBERT: Enhancing drug labeling text mining and analysis with AI language modeling

Wu,

Gray,

Dang

et al. 2023

Exp Biol Med (Maywood)

View full text Add to dashboard Cite

The US drug labeling document contains essential information on drug efficacy and safety, making it a crucial regulatory resource for Food and Drug Administration (FDA) drug reviewers. Due to its extensive volume and the presence of free-text, conventional text mining analysis have encountered challenges in processing these data. Recent advances in artificial intelligence (AI) for natural language processing (NLP) have provided an unprecedented opportunity to identify key information from drug labeling, thereby enhancing safety reviews and support for regulatory decisions. We developed RxBERT, a Bidirectional Encoder Representations from Transformers (BERT) model pretrained on FDA human prescription drug labeling documents for an enhanced application of drug labeling documents in both research and drug review. RxBERT was derived from BioBERT with further training on human prescription drug labeling documents. RxBERT was demonstrated in several tasks using regulatory datasets, including those involved in the National Institutes of Technology Text Analysis Challenge Dataset (NIST TAC dataset), the FDA Adverse Drug Event Evaluation Dataset (ADE Eval dataset), and the classification of texts from submission packages into labeling sections (US Drug Labeling dataset). For all these tasks, RxBERT reached 86.5 F1-scores in both TAC and ADE Eval classification, respectively, and prediction accuracy of 87% for the US Drug Labeling dataset. Overall, RxBERT was shown to be as competitive or have better performance compared to other NLP approaches such as BERT, BioBERT, etc. In summary, we developed RxBERT, a transformer-based model specific for drug labeling that outperformed the original BERT model. RxBERT has the potential to be used to assist research scientists and FDA reviewers to better process and utilize drug labeling information toward the advancement of drug effectiveness and safety for public health. This proof-of-concept study also demonstrated a potential pathway to customized large language models (LLMs) tailored to the sensitive regulatory documents for internal application.

show abstract

Section: Discussionmentioning

confidence: 99%

RxBERT: Enhancing drug labeling text mining and analysis with AI language modeling

Wu,

Gray,

Dang

et al. 2023

Exp Biol Med (Maywood)

View full text Add to dashboard Cite

show abstract

“…Sentiment analysis represents a natural language processing technique employed to extract emotional or affective indicators from text. The aim is to understand emotions from textual data automatically, a method extensively researched in the healthcare and medical elds (Meghana and Chavali, 2023;Pan et al, 2021;Vought et al, 2023). Given that our incorporated data comprises primarily Chinese texts (albeit with a minor presence of English expressions that do not affect the overall context), models like Vader, which are tailored for emotion feature extraction in English texts, are not applicable (Valdez et al, 2020).…”

Section: Sentiment Analysismentioning

confidence: 99%

Unveiling Public Attitudes and Themes Towards Heart Failure in China on Baidu Tieba: A Data Mining Study

Yuan,

Wei,

et al. 2024

Preprint

View full text Add to dashboard Cite

Background The prevalence of heart failure is continually increasing, impacting various regions and populations. The application of online communities in healthcare has emerged as a significant area of research. However, the exploration of Chinese public attitudes and content regarding heart failure from a popular perspective remains uncharted. Objective Describing the sentimental attitudes and main themes of Posts by Users on the “Heart Failure Bar” in Baidu Tieba. Methods Data were processed using Python programming. Comments from the "Heart Failure Bar" in Baidu Tieba were collected, followed by data cleaning, preprocessing, saving, and analysis. Findings A total of 37,495 comments were included, with themes encompassing “symptom and experience sharing”, “concerns about the quality of life”, “seeking advice and providing recommendations”, and “sharing resource”. Of the 22,371 "sentiment sentences" subjected to sentiment analysis, 2,258 were positive, 5,004 moderately positive, 6,765 neutral, 5,316 moderately negative, and 3,028 negative. The average sentiment score of the texts was 0.36, indicating an overall moderately negative public attitude towards heart failure. Conclusions Users related to heart failure in China have a strong desire for more professional medical services. Regional disparities in medical standards present a significant issue. Online communities demonstrate potential in bridging gaps in healthcare services.

show abstract

“…In February 2019, just six months after the NSCAI was formed and eight months before the Interim Report, White House Executive Order 13859, “The American AI Initiative,” signaled the Trump administration’s eagerness to align with Silicon Valley and “remove barriers to AI innovation” (The White House, 2019). The accompanying “Guidance for Regulation of Artificial Intelligence Applications,” sounding a lot like Clinton-Gore, pledged “to reduce barriers to the development and adoption of AI technologies” by rolling back existing regulations, overriding state laws and avoiding new regulations (Vought, 2020, p. 1).…”

Section: Stage Two: the Concentration Of Computational Knowledge Prod...mentioning

confidence: 99%

Surveillance Capitalism or Democracy? The Death Match of Institutional Orders and the Politics of Knowledge in Our Information Civilization

Zuboff¹

2022

Organization Theory

View full text Add to dashboard Cite

Surveillance capitalism is what happened when US democracy stood down. Two decades later, it fails any reasonable test of responsible global stewardship of digital information and communications. The abdication of the world’s information spaces to surveillance capitalism has become the meta-crisis of every republic because it obstructs solutions to all other crises. The surveillance capitalist giants–Google, Apple, Facebook, Amazon, Microsoft, and their ecosystems–now constitute a sweeping political-economic institutional order that exerts oligopolistic control over most digital information and communication spaces, systems, and processes. The commodification of human behavior operationalized in the secret massive-scale extraction of human-generated data is the foundation of surveillance capitalism’s two-decade arc of institutional development. However, when revenue derives from commodification of the human, the classic economic equation is scrambled. Imperative economic operations entail accretions of governance functions and impose substantial social harms. Concentration of economic power produces collateral concentrations of governance and social powers. Oligopoly in the economic realm shades into oligarchy in the societal realm. Society’s ability to respond to these developments is thwarted by category errors. Governance incursions and social harms such as control over AI or rampant disinformation are too frequently seen as distinct crises and siloed, each with its own specialists and prescriptions, rather than understood as organic effects of causal economic operations. In contrast, this paper explores surveillance capitalism as a unified field of institutional development. Its four already visible stages of development are examined through a two-decade lens on expanding economic operations and their societal effects, including extraction and the wholesale destruction of privacy, the consequences of blindness-by-design in human-to-human communications, the rise of AI dominance and epistemic inequality, novel achievements in remote behavioral actuation such as the Trump 2016 campaign, and Apple-Google’s leverage of digital infrastructure control to subjugate democratic governments desperate to fight a pandemic. Structurally, each stage creates the conditions and constructs the scaffolding for the next, and each builds on what went before. Substantively, each stage is characterized by three vectors of accomplishment: novel economic operations, governance carve-outs, and fresh social harms. These three dimensions weave together across time in a unified architecture of institutional development. Later-stage harms are revealed as effects of the foundational-stage economic operations required for commodification of the human. Surveillance capitalism’s development is understood in the context of a larger contest with the democratic order—the only competing institutional order that poses an existential threat. The democratic order retains the legitimate authority to contradict, interrupt, and abolish surveillance capitalism’s foundational operations. Its unique advantages include the ability to inspire action and the necessary power to make, impose, and enforce the rule of law. While the liberal democracies have begun to engage with the challenges of regulating today’s privately owned information spaces, I argue that regulation of institutionalized processes that are innately catastrophic for democratic societies cannot produce desired outcomes. The unified field perspective suggests that effective democratic contradiction aimed at eliminating later-stage harms, such as “disinformation,” depends upon the abolition and reinvention of the early-stage economic operations that operationalize the commodification of the human, the source from which such harms originate. The clash of institutional orders is a death match over the politics of knowledge in the digital century. Surveillance capitalism’s antidemocratic economic imperatives produce a zero-sum dynamic in which the deepening order of surveillance capitalism propagates democratic disorder and deinstitutionalization. Without new public institutions, charters of rights, and legal frameworks purpose-built for a democratic digital century, citizens march naked, easy prey for all who steal and hunt with human data. Only one of these contesting orders will emerge with the authority and power to rule, while the other will drift into deinstitutionalization, its functions absorbed by the victor. Will these contradictions ultimately defeat surveillance capitalism, or will democracy suffer the greater injury? It is possible to have surveillance capitalism, and it is possible to have a democracy. It is not possible to have both.

show abstract

Artificial Intelligence Analysis of Diabetic Retinopathy in Retinal Screening Events

Cited by 5 publications

References 2 publications

RxBERT: Enhancing drug labeling text mining and analysis with AI language modeling

RxBERT: Enhancing drug labeling text mining and analysis with AI language modeling

Unveiling Public Attitudes and Themes Towards Heart Failure in China on Baidu Tieba: A Data Mining Study

Surveillance Capitalism or Democracy? The Death Match of Institutional Orders and the Politics of Knowledge in Our Information Civilization

Contact Info

Product

Resources

About