kysgattu/evaluating-cross-domain-adaptability-of-text-summarizer-news-article-summarization

Creating Text Summarizer using BART Model on a BBC News Dataset and Evaluating Cross Domain Adaptability

Hi there! You happened to stumble upon the profile of.....

KAMAL YESHODHAR SHASTRY GATTU

A Machine Learning and Data Science Enthusiast: Exploring the Field of Data Analysis, Deep Learning and Machine Learning Tools

📖 Master of Science in Computer Science
🎓 Bachelor of Technology in Computer Science & Engineering

📧 E-mail : [email protected]
🔎 LinkedIn : https://www.linkedin.com/in/kysgattu
🔗 Portfolio

I have done these!!!

📌 Evaluating Cross-domain Adaptability Of Text Summarizer: News Article Summarization

• Engineered advanced text summarizers, seamlessly integrating Extractive (TextRank) and Abstractive (BART) techniques for optimizing news article summarization.
• Enhanced BART Model performance significantly, achieving a 20% improvement over previous implementations.
• Streamlined a thorough evaluation of cross-domain adaptability, consistently outperforming benchmark ROUGE scores in contrast to the model's original implementation and a fine-tuned BBC News model.
• Analyzed the abstractive summarizer's adaptability to different domains, affirming its versatility and effectiveness.

📌 Pedestrian Detection System using YOLO

• Developed a Pedestrian Detection System for the University of Massachusetts Lowell's Campus Planning Department.
• Utilized advanced Machine Learning Techniques and the formidable YOLO Deep Learning Algorithm to accurately count individuals on specific campus pathways.
• Introduced a user-friendly interface to enable users to effortlessly upload videos and define detection regions.
• The system seamlessly executes real-time detection, tracking, and tallying of pedestrians, elevating the Campus Planning Department's efficiency in managing pedestrian flow and reduced 90% manual time.

📌 Chest X-Ray Classification to detect COVID-19 using Deep Neural Networks

• Developed a reliable method for detecting COVID-19 in patients by analyzing chest X-rays using Deep Learning systems.
• Four CNN architectures are trained on a Chest Radiography dataset.
• Analyzed the efficiency of standard LeNet, ResNet, and VGG networks on Chest X-ray image classification.
• GRADCAM is used to detect COVID-19-affected areas in the lungs.

📌 Climate Change Sentiment Analysis

• Analyzed the opinion of people by performing sentiment analysis on tweets by Twitter users on the topic of Climate Change.
• Performed exploratory data analysis and trained models using various traditional machine learning algorithms and a Deep Learning Model using the Recurrent Neural Network approach using Text in Tweets and the frequency of some keywords as features.

📌 Citizens Income Prediction - Comparision Of ML Algorithms

• Used Adult Income Dataset from UCI Repository to predict the income of citizens and classify people into two categories based on various dependent properties of a person collected during a Census.
• Developed models using different Traditional Machine Learning Algorithms and used the results obtained are used to compare these algorithms using various evaluation metrics.

📌 Face Mask Detection

• Developed a system to detect whether the person in a video/picture is wearing a mask or not using image recognition techniques

📌 Detection of Recolored Images Using Deep Discriminative Model

• Developed a novel deep learning approach to detect fraudulency in images based on colors in images using various attributes like Illumination Consistency, Inter-Channel Correlation, etc.,

📌 Aadhaar Based Online Voting System

• Developed an alternative voting channel to increase voter participation, and reduce election costs while upholding the highest security, verifiability, and integrity standards which enables voters to exercise their vote from anywhere using the internet.

📌 Health Care – Diabetes

• Analyzing historical data of a group of patients and training system to detect the possibility of a patient having diabetes based on previous health conditions and finding effects of each health condition on having diabetes.

📌 User Knowledge Modelling

• Developing a Machine Learning model for predicting the students’ knowledge status about the subject of Electrical DC Machines based on certain constraints.

Talk to me about these.....

• 💻 Programming Languages: Python, Java, C, C+
• 💻 Python Libraries: NumPy, Pandas, Scikit-learn, Keras, PyTorch OpenCV, Matplotlib, Seaborn, NLTK
• 💻 Database: MySQL, PostgreSQL, Amazon Redshift
• 💻 ETL Tools: Informatica PowerCenter, Informatica Intelligent Cloud Services (IICS)
• 💻 Web Designing: HTML, CSS, JavaScript, JSP
• 💻 Operating Systems: Unix, Windows
• 💻 IDE: Eclipse, NetBeans, PyCharm, Jupyter
• 💻 Others: GitHub, MS Office, REST API, Microsoft PowerApps

I worked with them.....

• 💼 University of Massachusetts Lowell Facilities Informations Systems Assistant/Intern (Python/REST API Developer; Machine Learning Tools expert)
• 💼 Cognizant Technology Solutions Pvt Ltd, Hyderabad Programmer Analyst Trainee (Data Integration - ETL Developer)
• 💼 Electronics Corporation of India Ltd., Hyderabad Java Trainee & Intern
• 💼 SmartBridge Educational Services Pvt. Ltd. Machine Learning Trainee & Intern

Know about me more?

🇮🇳-🇺🇸 Born in India - Now living in the USA
🏏 Follow and play Cricket
📚 Love to read Fantasy, Fiction Novels
🪄🐺 A Potterhead and Westerosi!

📷 Know about me more on Instagram