Harshal Dharpure Email & Phone Number

India

India

India

Title - Prompt-Persuasive: Prompting for Multimodal Persuasive Meme Detection TaskMemes, being highly influential in online communication, can shape individuals' attitudes, behaviors, and beliefs, both positively and negatively. The identification of such persuasive memes is an essential task in the automatic classification of memes. A significant challenge in this realm is the semantic gap often found in multimodal models due to the inadequate alignment of textual and visual elements in memes. Recent approaches have involved fine-tuning pre-trained vision-language models (PVLMs) for meme detection.

Utilising Excel and SQL, I worked on exploratory data analysis (EDA) on a variety of data sets.built an interactive dashboard using Tableau Public and Excel.worked on E2E in SQL, including creating databases, tables, importing data, and using joins, cohort views, etc.The most likely default clients were predicted using information such as IMDB movie ratings, Instagram analytics, car attributes influencing pricing, and bank loan case study.

AWS Academy Graduate - AWS Academy Data Analytics Program offered By eduskills Foundation AICTE.

India

AWS Academy Graduate - AWS Academy Cloud Foundations Program

Project Title - Diabetics Disease Prediction using PySpark - Logistics RegressionData Loading, Data Cleaning and using the Logistics Regression using Machine Learning Technique, got 90 per cent Accuracy.

Ahmedabad, Gujarat, India

Research Title - How does Multilingual BERT Encode the Hidden Biases in the Code- Mixed Sentences?Quantify Biases in the m-BERT (Multilingual BERT) using Word Analogy Test and WEAT (Word Embedding Association Test) by these methods it helps to find the bias in the Word Embedding. For this, we have used the m-BERT Transformer Model.1) Created the Visualization of the Projection of the Professions (neutral and specific by the Orthography) on the gender Direction and Calculated the Score.2) Did Analysis on the gender Category and Profession and calculated the scores of Different Professions before debiasing and after debiasing3) Perform the WEAT test and find the results of the similarity score between different attributes and target words.4) Created the Hugging Face pipeline to perform the Biases finding.

Bachlor'S Degree, Computer Science, 80.03%

Artificial Intelligence and Data Science

Cbse Aissce, Mathematics And Computer Science

Cbse Aisse, Science, 9.2 Cgpa

Activities and Societies: Captain of Sapphire House