Novel approaches for big data analytics
Analytics is the most visible component of big data applications, creating value from the data by extracting knowledge and insights that are valuable for users or customers. Modern algorithms for data analytics, data mining and multi-dimensional analytical queries combine advanced statistical methods and machine learning algorithms such as deep learning. The challenge for such algorithms is to analyse huge datasets reliably, accurately and in a reasonable amount of time.
In spite of the processing power provided by distributed computing and dedicated processors, it may still take days for algorithms to learn from training data. Speeding up this step requires either more efficient algorithms or clever ways to prune or compress the input datasets, providing good results with less training data.
This project made several theoretical advances in the field of language models, in particular for conversational agents or dialogue systems used to answer queries.
This project devised novel ways to analyse performance in cloud data centres, an important task in managing computing resources efficiently while minimising energy consumption.
The project focused on developing machine learning theory and methodology for learning systems.
In this NRP 75 project, novel algorithms for the efficient analysis of large data sets have been developed.
The goal was to build a petabyte-scale analytics system that enables non-computer scientists to analyse high-performance data streams.
This NRP 75 project focused on a new approach to create algorithms that deliver both maximum power and speed.