A5048061360- Topic Modeling
- Mental Health via Writing
- Protein Structure and Dynamics
- Natural Language Processing Techniques
- Genomics and Phylogenetic Studies
- Biomedical Text Mining and Ontologies
- Asymmetric Hydrogenation and Catalysis
- Machine Learning in Bioinformatics
- Music and Audio Processing
- Expert finding and Q&A systems
- Suicide and Self-Harm Studies
- Authorship Attribution and Profiling
- Artificial Intelligence in Healthcare and Education
- Library Science and Administration
- Machine Learning in Healthcare
- IoT and Edge/Fog Computing
- Data Quality and Management
- Educational Assessment and Pedagogy
- Food Security and Health in Diverse Populations
- Robotics and Automated Systems
- Methemoglobinemia and Tumor Lysis Syndrome
- Internet of Things and AI
- Digital Media Forensic Detection
- Mathematics, Computing, and Information Processing
- Chronic Disease Management Strategies
Emory University
2023-2024
Indian Institute of Technology Kharagpur
2008-2023
Birla Institute of Technology and Science, Pilani
2023
Millennium Engineering and Integration (United States)
2000
Boston University
1996-1997
The dataset described is an aspect-level sentiment analysis for therapies, including medication, behavioral and other created by leveraging user-generated text from Twitter. was constructed collecting Twitter posts using keywords associated with the therapies (often referred to as treatments). Subsequently, subsets of collected were manually reviewed, annotation guidelines developed categorize positive, negative, or neutral. contains a total 5364 mentioning 32 therapies. These are further...
Abstract Background There is growing concern around the use of sodium nitrite (SN) as an emerging means suicide, particularly among younger people. Given limited information on topic from traditional public health surveillance sources, we studied posts made to online suicide discussion forum, “Sanctioned Suicide,” which a primary source and procurement SN. Objective This study aims determine trends in SN purchase use, obtained via data mining subscriber forum. We also aim substances topics...
Substance use disorders (SUDs) are a growing concern globally, necessitating enhanced understanding of the problem and its trends through data-driven research. Social media unique important sources information about SUDs, particularly since data in such often generated by people with lived experiences. In this paper, we introduce Reddit-Impacts, challenging Named Entity Recognition (NER) dataset curated from subreddits dedicated to discussions on prescription illicit opioids, as well...
<sec> <title>BACKGROUND</title> The increasing use of social media to share lived and living experiences substance presents a unique opportunity obtain information on side-effects, usage patterns, opinions novel psychoactive substances (NPS). However, due the large volume data, obtaining useful insights through natural language processing (NLP) technologies such as models (LLMs) is challenging. </sec> <title>OBJECTIVE</title> To develop retrieval-augmented generation (RAG) architecture for...
In this paper, we introduce a novel approach for the detection and classification of shot boundaries based on interval type-2 fuzzy logic. We consider low level features (Fuzzy Histogram (FH) Fuzzy Co-occurrence Matrix (FCM)) evaluate rule base modelled by sets to detect cut boundaries. After cuts, recovery mechanism is used reduce false positives caused illumination variation (e.g., camera flash light). The rules gradual transition form unified various types transitions like dissolves,...
Abstract We have implemented an iterative algorithm for the identification of diagnostic patterns from sets multiple‐domain proteins, where domains need not be common to all proteins in defining set. Our was applied sequences gathered using a variety methods, including BLAST, keywords, and E.C. numbers. In cases, useful were obtained, possessing both high sensitivity specificity. The found correlate several cases with functional structural domains. Patterns generated large number sequence...
Abstract Inferring the gender of named entities present in a text has several practical applications information sciences. Existing approaches toward name identification rely exclusively on using distributions from labeled data. In absence such data, these methods fail. this article, we propose two‐stage model that is able to infer names without requiring explicit name‐gender labels. We use coreference resolution as backbone for our proposed model. To aid where existing contextual does not...
Named entity recognition (NER) is a popular language processing task with wide applications. Progress in NER has been noteworthy, as evidenced by the F1 scores obtained on standard datasets. In practice, however, end-user uses an model their dataset out-of-the-box, text that may not be pristine. this paper we present four model-agnostic adversarial attacks to gauge resilience of models such scenarios. Our experiments state-of-the-art methods five English datasets suggest are over-reliant...
ChatGPT is a conversational language model that can interact with users naturally and engagingly.It based on GPT, large-scale neural network generate coherent diverse text from given prompt.ChatGPT trained using reinforcement learning human feedback, which allows it to adapt different contexts preferences.However, also faces some limitations challenges, such as producing inaccurate or nonsensical answers, being sensitive input phrasing, over-verbose repetitive.In this paper, we propose view...
Objective: To detect and classify features of stigmatizing biased language in intensive care electronic health records (EHRs) using natural processing techniques. Materials Methods: We first created a lexicon regular expression lists from literature-driven stem words for linguistic patient labels, doubt markers, scare quotes within EHRs. The was further extended Word2Vec GPT 3.5, refined through human evaluation. These lexicons were used to search matches across 18 million sentences the...
The increasing use of social media to share lived and living experiences substance presents a unique opportunity obtain information on side effects, patterns, opinions novel psychoactive substances. However, due the large volume data, obtaining useful insights through natural language processing technologies such as models is challenging. This paper aims develop retrieval-augmented generation (RAG) architecture for medical question answering pertaining clinicians' queries emerging issues...
The increasing use of social media to share lived and living experiences substance presents a unique opportunity obtain information on side effects, patterns, opinions novel psychoactive substances. However, due the large volume data, obtaining useful insights through natural language processing technologies such as models is challenging. This paper aims develop retrieval-augmented generation (RAG) architecture for medical question answering pertaining clinicians' queries emerging issues...
Abstract Objective To detect and classify features of stigmatizing biased language in intensive care electronic health records (EHRs) using natural processing techniques. Materials Methods We first created a lexicon regular expression lists from literature-driven stem words for linguistic patient labels, doubt markers, scare quotes within EHRs. The was further extended Word2Vec GPT 3.5, refined through human evaluation. These lexicons were used to search matches across 18 million sentences...
<sec> <title>BACKGROUND</title> There is growing concern around the use of sodium nitrite (SN) as an emerging means suicide, particularly among younger people. Given limited information on topic from traditional public health surveillance sources, we studied posts made to a suicide discussion online forum, ‘Sanctioned Suicide’, primary source and procurement SN. </sec> <title>OBJECTIVE</title> This study aims determine trends in SN purchase usage, obtained data mining forum Suicide’. We also...
Music recommendation systems can significantly improve the listening and search experiences of a music library or application. There is simply too much on market for user to navigate tens millions songs effectively. Because high demand excellent recommendations, field Recommendation Systems (MRS) rapidly expanding. The main motivation developing rating-based system was extract relevant information from reviews instrumental music. In this study, we suggest an NSGA-II-based based interest,...
Almost every part of the world relies on textbooks as primary medium imparting education. The quality education, thus, is correlated with textbooks. In general, content in used less-developed countries not up to mark [1]. addition their intended purpose delivering information, also promote behaviours that adults wish pass next generation [7]. It is, important ensure are helpful effective learning and do condone undesirable social mores. task evaluating against these parameters trivial:...
Cardiovascular diseases (CVD) has emerged as one of the major causes for death in all over world. This paper displays a framework to remotely screen, health disease affected patients utilizing Machine (M2M) innovation which is part project called CySician . Real time patient monitoring system advantageous and society it will significantly reduce medical charges, waiting improve handling capability any hospital. In this pulse rate, ECG, body temperature, Body Mass Index(BMI) general clinical...