Text/HTML

ARTIFICIAL INTELLIGENCE & DEEP LEARNING 2.0 (AI & DL 2.0)

Programme Lead: Dr Vijay CHANDRASEKHAR

The AI hype cycle has been driven by rapid advancements in deep learning in recent years.  In an initiative called Deep Learning 2.0, we are building the next generation deep learning toolbox, by tackling 10x problems in deep learning. Challenging research problems being tackled include: building 100x larger networks, reducing network size by 100x while maintaining performance, reducing video classification error rate by 10x, learning with 10x fewer labelled samples and building interpretable neural network models. More broadly, projects in speech, language and vision aim to bring these technologies one step closer to truly understanding people and the world around us by incorporating knowledge and reasoning. Join us in the next step towards general AI and building a world full of intelligent machines!

Contact Us to learn more!

Projects
Compact Descriptors for Video Analysis
Deep Learned Descriptors for Small Object Search in Videos
Deep Learning for Computational Semantics
Deep learning for intelligent search over big heterogeneous data
Deep Learning for Machine Translation, Information Retrieval, and Beyond
Deep learning, game theory, and dynamical systems
Deep Machine Learning Frameworks for Knowledge Extraction, Interpretation, and Summarization in Natural Language Processing and Understanding
Deep Reinforcement Learning on Embedded Platforms
Deep Representation Learning for 3D Surface Patches
Extreme Deep learning
Integration of Bayesian Methodologies with Deep Learning
Learning and mapping neural networks on neuromorphic hardware
Learning with fewer samples: towards closing the gap between supervised and unsupervised learning
Manufacturing as a Service (MaaS) - Towards the Design of a Deep Learning based Integrated Cloud-based Service Architecture for Digital Manufacturing
Massive Distributed Neural Networks
Multimodal Deep Learning
Online Deep Learning
Online Learning in Deep Spiking Neural Networks
Toward a deep learning approach for Knowledge Graph reasoning
White-Box Deep Learning for Biomedical Imaging Applications
White-Box Deep Learning for Decision Support in Healthcare