AI Systems That Can Self-Design & Self-Assemble to Surpass Human-Built Systems

Scientists are designing artificially intelligent computer systems with the ability to self-design and self-assemble to work more efficiently when compared to the best human-developed systems.

The newly launched research project has completely transformed the conventional software development process and will place computers in the leading role - allowing them to autonomously self-assemble their algorithms to most efficiently complete tasks - thereby saving energy consumption and operating costs.

Scientists from Lancaster University will develop an enormous toolkit of small code blocks that can be selected and arranged by the autonomous systems in the optimal way to realize tasks. In addition, the systems will be in a position to create brand new code blocks themselves as required, thus persistently figuring out optimal ways to perform their tasks when they operate.

Although the aim of this study is to realize automated writing and assembly of a wide range of software, its initial focus will be on the highly intricate ecosystem of modern data centers which have to constantly handle millions of diversified requests with the maximum possible efficiency.

In order to achieve this, the study will investigate the manner in which scores of distinctive interconnected self-assembling computer programs, operating over several machines in different locations, can work together to realize certain objectives—more rapidly processing requests, with less energy-consuming computational muscle, and in response to the ways in which popular content and services vary over time.

“We’re looking at entire swarms of computer programs all working together across many different computers, which are all individually self-assembling but also working together to achieve the programmer’s goal,” stated Dr. Barry Porter, a lecturer at Lancaster University’s School of Computing and Communications and lead researcher on the project.

“By fully automating the writing of the source code of each little block of behavior, the software continually creates its own new building blocks for systems without humans having to write them.”

“This unleashes systems from their programming, allowing them to continually produce more novel and innovative solutions to achieve their objectives.”

The eventual outcome could redefine the role of a computer programmer. It will assist in considerably reducing the human effort required to write software, reducing costs, and could also lead to the development of software with the potential to redesign itself to work best for its human users, by perceiving over the passage of time the way individuals like to work and use their technology.

This will help deliver a fundamental new paradigm of software development in which computer programmers will be freed from laboriously writing all the fine detail of every system, and will instead work at a much higher creativity level to guide the construction of complex software in collaboration with advanced machine learning.

It is a bit like a self-driving car of computer programming, in which programmers, or even end-users, define the destination and the machine figures out the best way to get there.

Dr. Barry Porter

While the initial focus of the scientists is on increasing the efficiency of data centers, this study could also help shape the future of artificial intelligence itself by way of innovative types of intelligent software with the ability to write and rewrite its own functionality, and with the potential to gain in-depth knowledge of the impact of its functionality on the world around it and the ability of the software to change that functionality for proving beneficial.

Using this, even non-programmers could be in a position to explain their requirements to their computer or smartphone and leave their device to come up with a solution that is far superior to anything it was ever programmed to perform.

The Leverhulme Trust has funded the three and a half-year project with £252,890.