Author:Joydeep Chandra

Publications
Adversarial Robustness in Optimized LLMs: Defending Against Attacks Joydeep Chandra and Prabal Manhas EasyChair Preprint 15857
Efficient and Scalable Self-Healing Databases Using Meta-Learning and Dependency-Driven Recovery Joydeep Chandra, Prabal Manhas, Aastha Sharma and Rashi Sahay EasyChair Preprint 15856
Mimicking Humanity: a Synthetic Data-Based Approach to Voice Cloning in Text-to-Speech Systems Michael Wang, Joydeep Chandra and Aleksandr Algazinov EasyChair Preprint 15780

Keyphrases

adversarial attacks, adversarial robustness, adversarial training, anomaly detection, Cascading Failure Prediction, computational efficiency, data scarcity, Database Dependency Modeling, Database Management Systems(DBMS), Dynamic Workload Adaptation, ethical concerns, Explainable AI (XAI), federated meta-learning, Graph Neural Networks (GNNs), Input Perturbations, Large Language Models (LLMs), linguistic diversity, LLM Security, meta-learning, Model-Agnostic Meta-Learning (MAML), Model Compression, model optimization, Model resilience, multi-objective optimization, Natural Language Processing (NLP), Privacy, Proactive Anomaly Prevention, Pruning, Quantization, Real-Time Adaptability, Recovery Optimization, Reinforcement Learning (RL), RL-Based Recovery, Scalable Database Systems, Self-Healing Databases, speaker embeddings, Speaker Generalization, speech synthesis, synthetic data, Task Generalization, Text-to-Speech (TTS), Voice Cloning.