Building a Question Answering (QA) system using BERT

[gokublacko7]

Title

A quantum-inspired sentiment representation model

Team Name

Noah

Email

202311016@daiict.ac.in

Team Member 1 Name

Harsh Vyas

Team Member 1 Id

202311015

Team Member 2 Name

Birva Oza

Team Member 2 Id

202311050

Team Member 3 Name

Parshwa Dand

Team Member 3 Id

202311016

Team Member 4 Name

Darshit Kalariya

Team Member 4 Id

202311035

Category

reproductivity

Problem Statement

Preprocessing: Text cleaning, spelling correction, removal of stop words, and tokenization using HMM-based part-of-speech tagging. Embedding Generation: The model uses word embeddings to represent words in a vector space. These embeddings are trained using the Gensim API with a dimension of 100. Classifier Training: Three machine-learning classifiers were used: Naive Bayes (NB), Support Vector Machine (SVM), and Random Forest (RF). Additionally, deep learning models like CNN, LSTM, and fully connected deep neural networks (FCDNN) were employed​

Evaluation Strategy

To evaluate the overall classification performance of each methods accuracy, precision, recall, and F1 score will be used

Dataset

Obama-McCain Debate (OMD) dataset and the Sentiment140 Twitter dataset

Resources

https://link.springer.com/article/10.1007/s10489-019-01441-4

[parth126]

Deliverables are not clear:

• Expected outcome is implementing this from scratch without use of existing github repos and huggingface.
• Evaluation will be based on understanding of the code and how well the trained model performs

[gokublacko7]

A quantum-inspired sentiment representation model for twitter sentiment analysis

[parth126]

The team will use QSR model for classification task on several datasets. It will be compared to other baseline models (e.g. VM, NB RF)