Research
I'm interested in natural language processing, deep learning, and generative AI. Most of my research is about language models, representation learning and applications of NLP.
|
|
SYNFAC-EDIT: Synthetic Imitation Edit Feedback for Factual Alignment in Clinical Summarization
Prakamya Mishra*,
Zonghai Yao*,
Parth Vashisht,
Feiyun Ouyang,
Beining Wang,
Vidhi Dhaval Mody,
Hong Yu
arXiv, 2024
arXiv
This study leverages synthetic edit feedback to improve factual accuracy in clinical summarization using DPO and SALT techniques. Our approach demonstrates the effectiveness of GPT-generated edits in enhancing the reliability of clinical NLP applications.
|
|
Clustering-based sampling for few-shot cross-domain keyphrase extraction
Prakamya Mishra*,
Lincy Pattanaik*,
Arunima Sundar*,
Nishant Yadav,
Mayank Kulkarni
Findings at, EACL 2024
Paper
We propose a novel clustering-based few-shot
sampling approach that leverages intrinsically
available sub-domain information as topics
from the dataset to extract few-shot samples
to be labelled from the target domains and be
used for fine-tuning.
|
|
Synthetic Imitation Edit Feedback for Factual Alignment in Clinical Summarization
Prakamya Mishra*,
Zonghai Yao*,
Shuwei Chen,
Beining Wang,
Rohan Mittal,
Hong Yu
SyntheticData4ML workshop at, NeurIPS 2023
Paper
In this work, we propose a new pipeline using ChatGPT instead of human experts to generate high-quality feedback data for improving factual consistency in the clinical note summarization task.
|
|
STEPs-RL: Speech-Text Entanglement for Phonetically Sound Representation Learning
Prakamya Mishra
Long paper at, PAKDD 2021   (Oral Presentation)
Paper
In this work, we present a novel multi-modal deep neural network architecture that uses speech and text entanglement for learning phonetically sound spoken-word representations.
|
|
NeuralNERE: Neural Named Entity Relationship Extraction for End-to-End Climate Change Knowledge Graph Construction
Prakamya Mishra,
Rohan Mittal
Tackling Climage Change using Machine Learning workshop at, ICML 2021   (Spotligh Presentation)
Paper
/
Presentation
We propose NeuralNERE, an end-to-end Neural Named Entity Relationship Extraction model
for constructing climate change knowledge graphs directly from the raw text of relevant
news articles. Additionally, we introduce a new climate change news dataset (called SciDCC dataset)
containing over 11k news articles scraped from the Science Daily website.
|
|
Bi-ISCA: Bidirectional Inter-Sentence Contextual Attention Mechanism for Detecting Sarcasm in User Generated Noisy Short Text
Prakamya Mishra,
Saroj Kaushik,
Kuntal Dey
MRC-HCCS wrokshop at, IJCAI 2021
Paper
Developed novel Bi-directional Inter-Sentence Contextual Attention mechanism (Bi-ISCA) to capture inter-sentence dependencies for detecting sarcasm. Explained model behaviors and predictions by analyzing the attention maps and identifying words responsible for invoking sarcasm.
|
|