The School of Computer Science and Statistics is proud to announce its involvement in the RELAX-DN (Relaxed Semantics Across the Data Analytics Stack) initiative as part of the European MSCA network. The project aims to develop scalable and efficient data-intensive software systems by fostering research and collaboration between 14 academic and industrial partners across 8 countries.

Two newly recruited PhD researchers, Hue Dang and Moule Lin, have joined the RELAX network as part of its mission to address challenges in the design of modern data analytics systems. They began their work in April and are co-supervised by SCSS’s Dr. Andrea Patane and Dr. Goetz Botterweck.

The RELAX European Doctoral Network aims to train a cohort of highly mobile and adaptable researchers to become experts in the design of scalable and efficient data-intensive software systems. These experts will master the specific skill of navigating the semantics or correctness conditions of applications, with the goal of enhancing scalability, response times, and availability. Working across the disciplinary specialisms of data science, data management, distributed computing and computing systems, the Fellows will develop knowledge of the broad issues underpinning data analytics systems.

Hue and Moule’s work is part of this larger vision, with each focusing on distinct but complementary areas of deep learning:

Hue Dang: Numeric Accuracy and Reproducibility in Deep Learning Training and Interface

Hue’s research focuses on the Different versions of machine-learning hardware and software typically yield slightly different answers due to differences in floating point order of evaluation. The result is often poorer accuracy, or the same overall accuracy but different classifications between the two implementations, with unpredictable results. The goal of this work is to develop methods for trained models with sharper distinctions between classifications so that the model is more resilient to minor changes.

Moule Lin: Arithmetic and Number Systems for Deep Learning

Moule’s work is centred on Developing numeric types that match value distributions and operations of training better than existing default types. Identify number systems that make better use of limited encodings for both inference and training. Investigate domain-specific and application-specific number systems and encodings for improved compactness and customize the level of precision of data to the movement of the data within the parallel/distributed computing system.

For more details on RELAX visit: RELAX Project