Projects
LLM Evaluation Dashboard
This dashboard evaluates the performance of various large language models, including ๐๐ฎ๐ฐ๐ฒ๐ฏ๐ผ๐ผ๐ธ ๐๐๐ฅ๐ง, ๐ง๐ฑ, ๐๐ฃ๐ง-๐ฎ, ๐๐๐ฅ๐ง (๐ฏ๐ผ๐๐ต ๐๐ป๐ฐ๐ฎ๐๐ฒ๐ฑ ๐ฎ๐ป๐ฑ ๐ฐ๐ฎ๐๐ฒ๐ฑ). It provides insights into:
Average generation time,
Token length,
Proportionality between token and generation,
Runtime complexity of each model.
Source code: Here.ย
Probabilistic Impact Score Generation using Ktrain-BERT to Identify Hate Words from Twitter Discussions
This repository contains the source code for experimentation and the public dataset (train and test) of the Hate Speech and Offensive Content Identification in English and Indo-Aryan Languages from the 13th meeting of the Forum for Information Retrieval Evaluation (FIRE 2021).ย
Predicting the pandemic: sentiment evaluation and predictive analysis from large-scale tweets on Covid-19 by deep convolutional neural network
This repository contains the Twitter data (validation only) on Covid-19, initially from early 2020 (March-June). We develop a large tweet corpus exclusively based on the Coronavirus tweets. We split the data into train and test sets, and we perform polarity classification and trend analysis. The refined outcome from the trend analysis helps to train the data to provide an incremental learning curvature for our neural network, and we obtain an accuracy of 90.67%. Finally, we provide a statistical-based future prediction for Coronavirus case growth. Our model outperforms several previous state-of-the-art experiments in overall sentiment accuracy comparison for similar tasks, but it also maintains thorough performance stability among all the test cases when tested with several popular open-source text corpora.ย
Sentiment classification with GST tweet data on LSTM based on polarity-popularity model
A full-scale framework for determining the polarity-popularity occurrence order of words extracted from tweets, based on a large-scale economic reform. This research work is one of the most comprehensive approaches to large-scale Twitter data, i.e., based on the implementation of GST in India in July 2017.ย
Our papers were a few of the first papers ever published on this research topic.
My Talk on โNLP: From Inception to GPT and Beyondโ