Featured Papers
Natural Language Processing Advancements By Deep Learning: A Survey
A comprehensive survey covering the evolution of deep learning techniques in NLP, from word embeddings and recurrent architectures to transformer-based models. Reviews key advances in machine translation, text classification, question answering, and language generation.
Differentially Private Synthetic Medical Data Generation using Convolutional GANs
Proposes a framework for generating synthetic medical data with formal differential privacy guarantees using convolutional GANs. Enables sharing of realistic medical datasets while protecting patient privacy, addressing a critical bottleneck in healthcare AI research.
3D Convolutional Neural Networks for Cross Audio-Visual Matching Recognition
Introduces a 3D CNN architecture for joint audio-visual recognition, leveraging spatiotemporal features from synchronized audio and video streams. Demonstrates effectiveness on speaker verification and lip-reading tasks with end-to-end training.