A5103078881- HIV Research and Treatment
- Topic Modeling
- HIV/AIDS drug development and treatment
- Cell Image Analysis Techniques
- HIV/AIDS Research and Interventions
- Web Data Mining and Analysis
- Genomics and Phylogenetic Studies
- AI in cancer detection
- Chromosomal and Genetic Variations
- Genetics, Bioinformatics, and Biomedical Research
- COVID-19 and healthcare impacts
- Infection Control and Ventilation
- COVID-19 epidemiological studies
- Computer Graphics and Visualization Techniques
- Natural Language Processing Techniques
- Advanced Malware Detection Techniques
- Spectroscopy Techniques in Biomedical and Chemical Research
- Machine Learning in Healthcare
- Text Readability and Simplification
- Advanced Vision and Imaging
- Explainable Artificial Intelligence (XAI)
- Advanced Neural Network Applications
- Adversarial Robustness in Machine Learning
- HIV, Drug Use, Sexual Risk
- Cellular Mechanics and Interactions
Brown University
2012-2021
Massachusetts Institute of Technology
2021
Miriam Hospital
2012
The science around the use of masks by general public to impede COVID-19 transmission is advancing rapidly. Policymakers need guidance on how should be used population combat pandemic. Here, we synthesize relevant literature inform multiple areas: 1) characteristics COVID-19, 2) filtering and efficacy masks, 3) estimated impacts widespread community mask use, 4) sociological considerations for policies concerning mask-wearing. A primary route likely via small respiratory droplets, known...
Abstract Deep learning, which describes a class of machine learning algorithms, has recently showed impressive results across variety domains. Biology and medicine are data rich, but the complex often ill-understood. Problems this nature may be particularly well-suited to deep techniques. We examine applications biomedical problems—patient classification, fundamental biological processes, treatment patients—and discuss whether will transform these tasks or if sphere poses unique challenges....
Izzeddin Gur, Ofir Nachum, Yingjie Miao, Mustafa Safdari, Austin Huang, Aakanksha Chowdhery, Sharan Narang, Noah Fiedel, Aleksandra Faust. Findings of the Association for Computational Linguistics: EMNLP 2023.
Molecular epidemiological evaluation of HIV-1 transmission networks can elucidate behavioral components that be targets for intervention. We combined phylogenetic and statistical approaches using pol sequences from patients diagnosed between 2004 2011 at a large HIV center in Rhode Island, following 75% the state's population. Phylogenetic trees were constructed maximum likelihood, putative clusters evaluated latent class analyses to determine association cluster size with underlying...
Next generation sequencing technologies have recently been applied to characterize mutational spectra of the heterogeneous population viral genotypes (known as a quasispecies) within HIV-infected patients. Such information is clinically relevant because minority genetic subpopulations HIV patients enable escape from selection pressures such immune response and antiretroviral therapy. However, methods for quasispecies sequence reconstruction next reads are not yet widely used remains an...
Aims: HIV-1 sequence diversity can affect host immune responses and phenotypic characteristics such as antiretroviral drug resistance. Current classification uses phylogeny-based methods to identify subtypes recombinants, which may overlook distinct subpopulations within subtypes. While local epidemic studies have characterized sequence-level clustering using phylogeny, identification of new genotype–phenotype associations are based on mutational correlations at individual positions. We...
Pre-trained large language models (LLMs) have recently achieved better generalization and sample efficiency in autonomous web automation. However, the performance on real-world websites has still suffered from (1) open domainness, (2) limited context length, (3) lack of inductive bias HTML. We introduce WebAgent, an LLM-driven agent that learns self-experience to complete tasks real following natural instructions. WebAgent plans ahead by decomposing instructions into canonical...
Next generation sequencing technologies have been successfully applied to HIV-infected patients in order obtain the mutational spectrum of heterogeneous viral populations within individuals, known as quasispecies. However, metage-nomics problem quasispecies sequence reconstruction from next reads is not-yet widely current practice and remains an emerging area research. Furthermore, majority research methodology HIV has focused on 454 sequencing, while many next-generation platforms are...
Abstract Background. Human immunodeficiency virus (HIV)-1 drug resistance mutations (DRMs) often accompany treatment failure. Although subtype differences are widely studied, DRM comparisons between subtypes either focus on specific geographic regions or include populations with heterogeneous treatments. Methods. We characterized patterns following first-line failure and their impact future in a global, multi-subtype reverse-transcriptase sequence dataset. developed hierarchical modeling...
Introduction Tenofovir‐containing regimens have demonstrated potential efficacy as pre‐exposure prophylaxis (PrEP) in preventing HIV‐1 infection. Transmitted drug resistance mutations associated with tenofovir, specifically the reverse transcriptase (RT) mutation K65R, may impact effectiveness of PrEP. The worldwide prevalence transmitted tenofovir different subtypes is unknown. Methods Sequences from treatment‐naïve studies and databases were aggregated analyzed by Stanford Database tools...
Large language models (LLMs) have shown exceptional performance on a variety of natural tasks. Yet, their capabilities for HTML understanding -- i.e., parsing the raw webpage, with applications to automation web-based tasks, crawling, and browser-assisted retrieval not been fully explored. We contribute (fine-tuned LLMs) an in-depth analysis under three tasks: (i) Semantic Classification elements, (ii) Description Generation inputs, (iii) Autonomous Web Navigation pages. While previous work...
Reducing computation cost, inference latency, and memory footprint of neural networks are frequently cited as research motivations for pruning sparsity. However, operationalizing those benefits understanding the end-to-end effect algorithm design regularization on runtime execution is not often examined in depth. Here we apply structured unstructured to attention weights transformer blocks BERT language model, while also expanding block sparse representation (BSR) operations TVM compiler....
In the past, computer vision systems for digitized documents could rely on systematically captured, high-quality scans. Today, transactions involving digital are more likely to start as mobile phone photo uploads taken by non-professionals. As such, document automation must now account captured in natural scene contexts. An additional challenge is that task objectives processing can be highly use-case specific, which makes publicly-available datasets limited their utility, while manual data...