Projects:
De-identification of Indian clinical discharge summaries using transfer learning with Large Language Models (LLMs).
Generation of medical text using in-context learning with LLMs.
Development of a medical domain chatbot for Ayurvedic, Homeopathic, and Allopathic knowledge, leveraging Retrieval-Augmented Generation (RAG) and transfer learning.
Language Translation from English to Seven Indian Languages and vice versa using different Deep Learning Techniques like seq to seq LSTM, Bidirectional LSTM & GRU, and Transformers.
Worked on finding semantic relatedness between pairs of sentences. (Sem Eval Task 2024)
PICO Classification using Domain-specific Features.
Presented a Conference paper on PICO Classification using domain-specific features. Instead of Using pre-trained word embedding like tf-idf, word2vec, etc, domain-specific features were used as input to the classifiers.
Text mining for PICO identification in Biomedical Domain.
Text mining in medical literature using different machine learning and deep learning approaches. Different domain-related Entity extractors were used while performing text mining task.
Sentiment analysis on COVID-related tweets.
Performed clustering on the National Anthem of different Countries and tried to find similarities between the countries based on the National Anthem.
Developed Shudoku Solver using Deep learning.
Publication
Sanjeet Singh and Aditi Sharan “PICO Classification using Domain-specific Features” 3rd International Conference on Emergent Converging Technologies and Biomedical Systems(ETBS 2023). DOI: 10.1007/978-981-99-8646-0_35
Co-author in the book titled "Evaluation and Advancement in Biomedical Text Mining," contributing to the chapters on "Fundamentals of Vector Text Representation and Word Embeddings" and "Transformer-based Models for Text Representation." (Under Review)