Arpita Roy

ORCID: 0009-0007-7472-174X
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Topic Modeling
  • Spam and Phishing Detection
  • Natural Language Processing Techniques
  • Advanced Malware Detection Techniques
  • Software Engineering Research
  • User Authentication and Security Systems
  • Biomedical Text Mining and Ontologies
  • COVID-19 Clinical Research Studies
  • SARS-CoV-2 and COVID-19 Research
  • Misinformation and Its Impacts
  • Biometric Identification and Security
  • Herpesvirus Infections and Treatments
  • Poxvirus research and outbreaks
  • Hepatitis C virus research
  • Bacillus and Francisella bacterial research
  • Impact of Technology on Adolescents
  • Digital Mental Health Interventions
  • Data Quality and Management
  • Consumer Retail Behavior Studies
  • Angiogenesis and VEGF in Cancer
  • Hate Speech and Cyberbullying Detection
  • Music and Audio Processing
  • Advanced Steganography and Watermarking Techniques
  • Rheumatoid Arthritis Research and Therapies
  • Network Security and Intrusion Detection

Sharda University
2022-2023

Koneru Lakshmaiah Education Foundation
2020-2023

Jain University
2022

University of Maryland, Baltimore County
2016-2021

West Middlesex University Hospital
2020

University of Maryland, Baltimore
2019

In recent years pre-trained language models (PLM) such as BERT have proven to be very effective in diverse NLP tasks Information Extraction, Sentiment Analysis and Question Answering. Trained with massive general-domain text, these capture rich syntactic, semantic discourse information the text. However, due differences between general specific domain text (e.g., Wikipedia versus clinic notes), may not ideal for domain-specific extracting clinical relations). Furthermore, it require...

10.18653/v1/2021.emnlp-main.435 article EN cc-by Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing 2021-01-01

Arpita Roy, Youngja Park, Taesung Lee, Shimei Pan. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.

10.18653/v1/d19-1067 article EN 2019-01-01

In medical imaging, skin lesion prediction and classification is highly crucial while predicting malignancy. Various prevailing deep learning-based CAD diagnosis approaches show poor performance. It incredibly challenging to diagnose lesions with complex features like artefacts, boundary analysis, low contrast images foreground background images, constraint training datasets. Also, it relies on the appropriate tuning of millions parameters that causes generalization, overfitting, massive...

10.52783/jisem.v10i38s.6840 article EN Journal of Information Systems Engineering & Management 2025-04-22

The discovery of the SARS-CoV-2 Omicron (B.1.1.529) variant has sparked alarm globally because its rapid rate infection and trespassing acquired immunity due to vaccination or natural infection. This heavily mutated is rapidly spreading around world. Infected individuals with may suffer from flu-like symptoms, infected Delta frequently report low oxygen levels, high pulse rates, a loss smell taste. Also, causes asymptomatic mild disease so far, not any severe illness as like Delta, this new...

10.1177/2632010x221124908 article EN cc-by-nc Clinical Pathology 2022-01-01

The rapid evolution of digital technologies has necessitated robust and scalable testing solutions to ensure high-quality software delivery across multiple domains. Karate Framework emerges as a comprehensive solution that seamlessly integrates API UI test automation, enabling organizations achieve enhanced efficiency, reliability, agility in their processes. By leveraging Karate’s intuitive syntax versatile features, engineers can design reusable cases facilitate cross-functional efforts...

10.63345/ijrmeet.org.v13.i4.18 article EN 2025-04-01

Problematic internet use (PIU) by children and adolescents is a concern for many parents. Several factors, including students' education level, the method of instruction, dependence on internet, their intended could all be contributing factors to PIU depression. Disturbed mental health may attributed cancellation physical classes because COVID-19 outbreak. This study aimed assess association pandemic with depressive symptoms in adolescent students.

10.1002/hsr2.1008 article EN cc-by-nc-nd Health Science Reports 2022-12-22

Word embedding is a Natural Language Processing (NLP) technique that automatically maps words from vocabulary to vectors of real numbers in an space. It has been widely used recent years boost the performance vari-ety NLP tasks such as Named Entity Recognition, Syntac-tic Parsing and Sentiment Analysis. Classic word methods Word2Vec GloVe work well when they are given large text corpus. When input texts sparse many specialized domains (e.g., cybersecurity), these often fail produce...

10.48550/arxiv.1709.07470 preprint EN other-oa arXiv (Cornell University) 2017-01-01

Security of the data is one major concerns now a days. To secure text–based authentication vulnerable to attacks like shoulder surfing, hidden cameras and hackers. avoid these attacks, biometric schemes are useful by means face recognition, voice thumb impression detection, fingerprint detection etc which more costly as it requires some external equipments. In this paper, we proposed security system based on fuzzy rule intelligent can be used for process user's behavior. The behavioral...

10.1080/09720529.2020.1728894 article EN Journal of Discrete Mathematical Sciences and Cryptography 2020-02-17

Monkeypox (MPOX) is a zoonotic viral disease caused by an Orthopox DNA virus named mpox (MPOXV). Scientists first discovered MPOXV in monkey transported from Singapore to Denmark for research purposes. However, at first, the virus's primary animal reservoir was rodents. In 1970, 9-year-old child Democratic Republic of Congo diagnosed with MPOX. Since then, 11 African nations have recorded human instances MPOX.1 Between 1970 and 1979, six countries reported only 48 confirmed MPOX cases. more...

10.1002/hsr2.1030 article EN cc-by Health Science Reports 2022-12-29

Word embedding, a process to automatically learn the mathematical representations of words from unlabeled text corpora, has gained lot attention recently. Since are basic units natural language, more precisely we can represent morphological, syntactic and semantic properties words, better support downstream Natural Language Processing (NLP) tasks. traditional word embeddings mainly designed capture relatedness between co-occurred in predefined context, it may not be effective encoding other...

10.24963/ijcai.2020/686 article EN 2020-07-01

Illicit drug use is a serious problem around the world. Social media has increasingly become an important tool for analyzing patterns and monitoring emerging abuse trends. Accurately retrieving illicit drug-related social posts step in this research. Frequently, hashtags are used to identify retrieve on specific topic. However highly ambiguous. Posts with same not always Moreover, evolving, especially those related drugs. New street names introduced constantly avoid detection. In paper, we...

10.1109/bibm.2016.7822752 article EN 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 2016-12-01

We describe the systems developed by UMBC team for 2018 SemEval Task 8, SecureNLP (Semantic Extraction from CybersecUrity REports using Natural Language Processing). participated in three of sub-tasks: (1) classifying sentences as being relevant or irrelevant to malware, (2) predicting token labels sentences, and (4) attribute Malware Attribute Enumeration Characterization vocabulary defining malware characteristics. achieve F1 score 50.34/18.0 (dev/test), 22.23 (test-data), 31.98...

10.18653/v1/s18-1142 article EN cc-by 2018-01-01

Recently, text embedding techniques such as Word2Vec and BERT have produced state-of-the-art results in a wide variety of NLP tasks. As result, traditional features frequently used Information Extraction (IE) POS tags, dependency relations semantic types received less attention. In this paper, we investigate whether can be combined with word sentence embeddings to improve relation extraction. We explored diverse feature sets different neural network architectures evaluated our models on...

10.1109/ictai50040.2020.00072 article EN 2020-11-01

Nowadays, teens and young adults spend a significant amount of time on social media. According to the national survey American attitudes substance abuse, who media sites are at increased risk smoking, drinking illicit drug use. Reducing teens’ exposure use-related posts may help minimize their future use addiction. In this paper, we present method for automated detection userelated posts. With technology, content can be automatically filtered out from To detect related posts, employ...

10.13016/m2nywr-a9qk article EN International Conference on Tools with Artificial Intelligence 2017-11-01

Arpita Roy, Youngja Park, Shimei Pan. Proceedings of the 2019 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). 2019.

10.18653/v1/n19-1293 article EN 2019-01-01

The prediction of skin cancer at the earlier stage is extremely essential for melanoma. There a need intellectual computer analysis lesions. segmentation lesion boundaries vital to accurately identify lesions from dermoscopic images where diagnosis complex various types. Thus, some pre-processing steps are required attain higher sensitive boundary and classification. Initially, done with median filter offer reputation preservation does not in-cooperate newer pixel values processed image....

10.1142/s0218539323500249 article EN International Journal of Reliability Quality and Safety Engineering 2023-08-20

Word embedding is a Natural Language Processing (NLP) technique that automatically maps words from vocabulary to vectors of real numbers in an space. It has been widely used recent years boost the performance variety NLP tasks such as named entity recognition, syntactic parsing and sentiment analysis. Classic word methods Word2Vec GloVe work well when they are given large text corpus. When input texts sparse many specialized domains (e.g., cybersecurity), these often fail produce...

10.1109/ictai.2019.00226 article EN 2019-11-01

A convolutional intermittent neural network (CRNN) for labeling music is presented in this paper. Convolutional networks (CNNs) are used by CRNNs to extract nearby elements, while recurrent (RNNs) briefly summarize the extracted highlights. We compare two CNN structures and CRNN that have been applied labeling, controlling number of parameters preparation time each test. Overall, our research revealed perform well terms parameter quantity time, proving usefulness their crossbreed structure...

10.1109/iccsai59793.2023.10421180 article EN 2023-11-23

Retailers/Businessmen search for quick benefits with fewer speculations. This paper focuses on structuring an application to yield more profits retailers by utilizing Machine Learning. By considering the properties such as spot of retail, season impact product(s), and many produce a where product(s) can gain retailers/business people. knowing proper item right spot, purchase required through application. Learning it helps discover "Pace Recommendation (exactness)." Through that precision,...

10.1109/iccsea49143.2020.9132903 article EN 2020 International Conference on Computer Science, Engineering and Applications (ICCSEA) 2020-03-01
Coming Soon ...