Question Answering System (QAS) for Online Services
- Improving the accuracy of the Text-Visual Question Answering System (Text-VQA) that can read text from the daily scene images and document images-to help the visually impaired people and for a better navigation system. Use the individual and multimodal attention mechanism for effectively joining all the modalities in a common space to provide the exact answer.
- Developing a Visual Dialogue System (Visual Chatbot) by identifying the Intent in the question, Image, and dialogue history.
- Developing a Visual Question Answering System (VQA) that can answer a natural language question by extracting the visual features from the given Image. Improving the accuracy of the Multi-lingual VQA system for the Indian languages.
- Developing an Ontology-based Question Answering System with the LCQUAD dataset and improving the problem of Entity/Relation resolution using the knowledge graph.
- Improving the accuracy of the FAQ-based Question Answering System, which recommends a list of similar questions to the user.
Fake News Detection in Social Media (FND)
- Cross-Lingual Fake News Detection using the Fact-checked given by the high-resource languages to be used in the Low-resource languages.
- Fake News Detection by leveraging the latest Few shot and Zero shot learning methods for the Multilingual using the Low-resource languages such as Indian Languages: Hindi, Bengali etc.
- Design a Multimodal and Multilingual framework for the Weibo dataset (Chinese) using Transfer Learning.
- Using Large Language Models to improve Machine Translation for Indic languages
Human Action Recognition
- Visual feature determination and system development for Alzheimer's.
- Visual feature determination and system development for dementia.
- Visual cues determination and system development Post Stroke.
Designing Optimal and Targeted Advertisement Campaign Strategy for Online E-Commerce Platforms using AI and Machine Learning
- Analyzing and predicting factors affecting the overall profitability of an advertisement campaign.
- Identifying and explaining the impact of changing input variables [Bid, Budget] on the output metrics [ %impression share, ROAS - return on adspend] across such campaigns.
Conversational AI to understand customer behavior :
- Developing a Conversational Chatbot for Omni-Channel Customer Experiences.
- Build a Conversational Model Based on Reinforcement Learning Which Generates Responses in Native-Language.
- Improving Conversational System of Unscripted, Unplanned Responses using Machine Learning to get Accurate Responses.
- Developing a Multi-Lingual Embodied Conversational Question- Answering agents.
- Implementing a Rasa-X: Simultaneous Conversational Chatbot for Multi-intent.
- Implementing a Neural Conversational Model for Multilingual Question-Answering System.
Please fill in your academic details by 25th October 2023 for the SOP/ LOP /Thesis
Please fill your interest https://forms.gle/wDaKXTF9nS8qpSTP8