Enough is Enough: Learning to Stop in Generative Systems

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Gene regulatory networks (GRNs) have been used to drive
artificial generative systems. These systems must begin and then stop
generation, or growth, akin to their biological counterpart. In nature,
this process is controlled automatically as an organism reaches its mature
form; in evolved generative systems, this is more typically controlled
by hardcoded limits, which can be difficult to determine. Removing parameters
from the evolutionary process and allowing stopping to occur
naturally within an evolved system would allow for more natural and
regulated growth. This paper illustrates that, within the appropriate
context, the introduction of memory components into GRNs allows a
stopping criterion to emerge. A Long Short-Term Memory style network
was implemented as a GRN for an Evo-Devo generative system and
was tested on one simple (single point target) and two more complex
(point clouds) problems with and without structure. The novel LSTMGRN
performed well in simple tasks to optimise stopping conditions,
but struggled to manage more complex environments. This early work
in self-regulating growth will allow for further research in more complex
systems to allow the removal of hyperparameters and allowing the evolutionary
system to stop dynamically and prevent organisms overshooting
the optimal.
Original languageEnglish
Title of host publication13th International Conference on Artificial Intelligence in Music, Sound, Art and Design
Subtitle of host publicationEvoMUSART
Number of pages16
Publication statusAccepted/In press - 10 Jan 2024
EventInternational Conference on Artificial Intelligence in Music, Sound, Art and Design - Aberystwyth, United Kingdom
Duration: 3 Apr 20245 Apr 2024
Conference number: 13

Publication series

NameLecture Notes in Computer Science


ConferenceInternational Conference on Artificial Intelligence in Music, Sound, Art and Design
Abbreviated titleEvoMUSART
Country/TerritoryUnited Kingdom

Bibliographical note

This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.


  • Generative Design
  • Self-regulation
  • EvoDevo

Cite this