Ian Mertz’s ERC Project Aims to Improve Computer Memory

In his research, Ian William Mertz pushes boundaries of how mankind understands computer memory, one of the most fundamental resources in computation. The supported project studies a counter-intuitive question: can computers use memory which is already full of data to perform other tasks? Ordinary computer programs typically use empty memory for computing. The field of catalytic computing, on the other hand, studies if these programs can instead temporarily borrow full memory to perform their computations, with the condition that they restore the original data after finishing their work. Last year, Mertz's catalytic work was the key piece of a major breakthrough on the relationship between computational time and space.

The goal of STRAPS is to link the field of catalytic computing with the traditional study of memory-bounded algorithms. „This recent breakthrough has reinforced our belief that studying full memory is key to unlocking the most fundamental problems in the field. With this project we hope to attack questions which have been open since the earliest days of computing, using decades of past research on memory to inform this new idea of catalytic computation and vice versa. We also aim to understand when these algorithms can be utilized in practical settings where a lack of memory is obstructing progress, such AI, cloud computing, and modern quantum devices,“ says Ian Mertz.

The ERC Starting Grant is designed to support the independent careers of outstanding young scientists at a stage when they are establishing their own research team or programme. Formally speaking, applicants may be researchers of any age or nationality who obtained their PhD between 2 and 7 years prior to the fixed date of 1 January of the year to which the current ERC work programme relates. The principal investigator must commit to devoting at least 50 per cent of their total working time to the ERC project and to spending at least 50 per cent of that time in the EU Member State or country associated with the Framework Programme where the project is hosted. The amount of funding may be up to EUR 1.5 million over a period of five years.

Ian William Mertz works as a researcher at the Computer Science Institute, Charles University. He studied at Rutgers University, obtained his PhD in computer science from the University of Toronto in Canada, and previously held a postdoctoral position at the University of Warwick in the UK. He specialises in computational theory, specifically computational complexity and the mathematical foundations of memory usage. He is a recognised expert in catalytic computation; his work is regularly presented at the field’s most prestigious conferences, such as STOC and FOCS, including the highly acclaimed paper “Tree Evaluation is in Space O(log n · log log n)”, on which he collaborated with James Cook.