FedSparQ: Adaptive Sparse Quantization with Error Feedback for Robust & Efficient Federated Learning