A5039901099- Topic Modeling
- Biomedical Text Mining and Ontologies
- Machine Learning in Healthcare
- Natural Language Processing Techniques
- Bone Tissue Engineering Materials
- Reconstructive Surgery and Microvascular Techniques
- Reconstructive Facial Surgery Techniques
- Facial Rejuvenation and Surgery Techniques
- Lymphatic System and Diseases
- Bone and Dental Protein Studies
- Bone Metabolism and Diseases
- Mesenchymal stem cell research
- Body Contouring and Surgery
- Electronic Health Records Systems
- AI in cancer detection
- Nonmelanoma Skin Cancer Studies
- Periodontal Regeneration and Treatments
- Anatomy and Medical Technology
- Speech and dialogue systems
- COVID-19 diagnosis using AI
- Nasal Surgery and Airway Studies
- 3D Printing in Biomedical Research
- Radiomics and Machine Learning in Medical Imaging
- Domain Adaptation and Few-Shot Learning
- Wound Healing and Treatments
Harvard University
2016-2025
Boston Children's Hospital
2016-2025
John Wiley & Sons (United States)
2023
Hudson Institute
2023
Liechtenstein Institute
2023
East Stroudsburg University
2022
Brandeis University
2022
RMIT University
2022
Université d'Orléans
2022
Centre National de la Recherche Scientifique
2022
This article discusses the requirements of a formal specification for annotation temporal information in clinical narratives. We discuss implementation and extension ISO-TimeML annotating corpus notes, known as THYME corpus. To reflect task heavily inference-based reasoning demands domain, new guideline has been developed, “the Guidelines to (THYME-TimeML)”. clarify what relations merit annotation, we distinguish between linguistically-derived inferentially-derived orderings text. also apply...
Seventy-three patients with bite wounds (16 clenched-fist injuries, 18 human wounds, and 39 animal bites) were cultured aerobically anaerobically. A total of 33 34 bites injuries had aerobic or facultative bacteria isolated from their wounds. 224 strains isolated, the most frequent isolate being alpha-hemolytic streptococci (50 strains). Staphylococcus aureus was Penicillin-resistant gram-negative rods infrequently (12 Anaerobic in 16 88 anaerobic common various Bacteroides species (36
Background: Despite a perceived interest in autologous fat transfer, there is no consensus as to the best technique or level of success. The purpose present study was determine national trends techniques for harvest, preparation, and application fat, well success by practitioners. Methods: Comprehensive surveys were sent 650 randomly selected members American Society Aesthetic Plastic Surgery. survey aimed at determining whether transfer commonly performed procedure and, if so, specific...
A review of published work in clinical natural language processing (NLP) may suggest that the negation detection task has been "solved." This proposes an optimizable solution does not equal a generalizable solution. We introduce new machine learning-based Polarity Module for detecting text, and extensively compare its performance across domains. Using four manually annotated corpora we show suffers when there is no in-domain development (for manual methods) or training data methods). Various...
Research objective To develop scalable informatics infrastructure for normalization of both structured and unstructured electronic health record (EHR) data into a unified, concept-based model high-throughput phenotype extraction. Materials methods Software tools applications were developed to extract information from EHRs. Representative convenience samples two EHR systems—Mayo Clinic Intermountain Healthcare—were used development validation. Extracted was standardized normalized meaningful...
Dmitriy Dligach, Timothy Miller, Chen Lin, Steven Bethard, Guergana Savova. Proceedings of the 15th Conference European Chapter Association for Computational Linguistics: Volume 2, Short Papers. 2017.
Precise phenotype information is needed to understand the effects of genetic and epigenetic changes on tumor behavior responsiveness. Extraction representation cancer phenotypes currently mostly performed manually, making it difficult correlate phenotypic data genomic data. In addition, are being produced at an increasingly faster pace, exacerbating problem. The DeepPhe software enables automated extraction detailed from electronic medical records patients. system implements advanced Natural...
A large percentage of medical information is in unstructured text format electronic record systems. Manual extraction from clinical notes extremely time consuming. Natural language processing has been widely used recent years for automatic texts. However, algorithms trained on data a single healthcare provider are not generalizable and error-prone due to the heterogeneity uniqueness documents. We develop two-stage federated natural method that enables utilization different hospitals or...
We aimed to mine the data in Electronic Medical Record automatically discover patients' Rheumatoid Arthritis disease activity at discrete rheumatology clinic visits. cast problem as a document classification task where feature space includes concepts from clinical narrative and lab values stored Record.The Training Set consisted of 2792 notes associated values. Test 1 included 1749 2 344 for which there were no The Apache Text Analysis Knowledge Extraction System was used analyze text...
Abstract Objectives To improve the accuracy of mining structured and unstructured components electronic medical record (EMR) by adding temporal features to automatically identify patients with rheumatoid arthritis (RA) methotrexate-induced liver transaminase abnormalities. Materials methods Codified information a string-matching algorithm were applied RA cohort 5903 from Partners HealthCare select 1130 potential toxicity. Supervised machine learning was as our key method. For features,...
Classic methods for clinical temporal relation extraction focus on relational candidates within a sentence. On the other hand, break-through Bidirectional Encoder Representations from Transformers (BERT) are trained large quantities of arbitrary spans contiguous text instead sentences. In this study, we aim to build sentence-agnostic framework task CONTAINS extraction. We establish new state-of-the-art result task, 0.684F in-domain (0.055-point improvement) and 0.565F cross-domain...
To develop an open-source temporal relation discovery system for the clinical domain. The is capable of automatically inferring relations between events and time expressions using a multilayered modeling strategy. It can operate at different levels granularity--from rough temporality expressed as event to document creation (DCT) containment fine-grained classic Allen-style relations.We evaluated our systems on 2 corpora. One subset Temporal Histories Your Medical Events (THYME) corpus, which...
Electronic health record (EHR) data is collected by individual institutions and often stored across locations in silos. Getting access to these difficult slow due security, privacy, regulatory, operational issues. We show, using ICU from 58 different hospitals, that machine learning models predict patient mortality can be trained efficiently without moving out of their silos a distributed strategy. propose new method, called Federated-Autonomous Deep Learning (FADL) trains part the model all...
Classifying whether concepts in an unstructured clinical text are negated is important unsolved task. New domain adaptation and transfer learning methods can potentially address this issue.We examine neural unsupervised methods, introducing a novel combination of with transformer-based to improve negation detection. We also want better understand the interaction between widely used bidirectional encoder representations from transformers (BERT) system methods.We use 4 datasets that annotated...
Abstract Animal tracking data are being collected more frequently, in greater detail, and on smaller taxa than ever before. These hold the promise to increase relevance of animal movement for understanding ecological processes, but this potential will only be fully realized if their accompanying location error is properly addressed. Historically, coarsely-sampled have proved invaluable large scale processes (e.g., home range, habitat selection, etc.), modern fine-scale unlock far...
<sec> <title>BACKGROUND</title> Electronic Health Records (EHRs) and routine documentation practices are crucial for providing comprehensive health records, diagnoses, treatments patients' daily care. However, the complexity verbosity of EHR narratives can overload healthcare providers risk diagnostic inaccuracies. </sec> <title>OBJECTIVE</title> This study aims to enhance proficiency Large Language Models (LLMs) with a medical Knowledge Graph in automated diagnosis generation by minimizing...
In a review of 250 cases lymphedema the lower extremity, 9 patients were noted to share unique similarities in their history and physical findings. Although these had mild swelling pretibial areas all referred with diagnosis legs, findings differed significantly from usual patient either congenital or acquired lymphedema. Notably, extremity was always bilateral symmetrical nature never involved feet. Skin changes characteristic not found, consistent fat pads present anterior lateral malleoli...