PopularMajor Project
Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
₹4,900₹5,999
4.8/301 reviews
Artificial Intelligence · Major Project
Text Data Augmentation with LLMs
Paraphrases scarce training data to strengthen a small classifier
2-3 weeks build
Overview
Tests whether LLM paraphrases help when labelled data is scarce. A text classifier is trained on small slices of the public SST-2 sentiment dataset, then again with paraphrased copies generated by a language model and filtered for label drift and near-duplicates. Results are compared with classic augmentation such as synonym swaps and back-translation.
What makes it stand up
- 01Controlled low-data experiments at several training-set sizes
- 02LLM paraphrase generation with prompt templates
- 03Filters for duplicates and paraphrases that change the label
- 04Comparison with synonym replacement and back-translation
- 05Learning curves and per-class F1 reported for every setting
Module breakdown
- 01
Baseline
Classifier trained on small stratified data slices.
- 02
Generation
Paraphrase prompts and batch generation.
- 03
Filtering
Similarity and label-consistency checks on new samples.
- 04
Classic methods
Synonym swap and back-translation for comparison.
- 05
Analysis
Learning curves and significance across random seeds.
After this, you will be able to
- Explain when augmentation helps and when it adds noise
- Defend your experimental controls and repeated seeds
- Show the learning curves and interpret them for examiners
Technology stack
- Python
- Transformers
- Hugging Face
- scikit-learn
- Pandas
- Level
- Major Project
- Domain
- Artificial Intelligence
- Build time
- 2-3 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support