DOI
10.34229/KCA2522-9664.26.2.3
UDC 004.032.26
M.M. Glybovets
National University of "Kyiv-Mohyla Academy," Kyiv, Ukraine,
glib@ukma.edu.ua
S.O. Medvid
National University of "Kyiv-Mohyla Academy," Kyiv, Ukraine,
s.medvid@ukma.edu.ua
MorphoNAS: EMBRYOGENIC NEURAL ARCHITECTURE SEARCH
THROUGH MORPHOGEN-GUIDED DEVELOPMENT
Abstract. A novel MorphoNAS approach is proposed for deterministic neural network growth through morphogen-guided self-organization based on the Free Energy Principle, reaction-diffusion systems, and gene regulatory networks. The developmental model is investigated, where compact genomes encode only morphogen dynamics and threshold-based cellular rules enabling single progenitor cell transformation into complex neural architectures. Full success (100%) is achieved in evolutionary search for genomes generating predefined graph configurations with 8-31 nodes. Minimal functional controllers (6–7 neurons) for the CartPole task are obtained under network size minimization pressure with 94% population success rate. The results demonstrate that biologically plausible developmental rules can serve as an effective mechanism for automated neural architecture search.
Keywords: MorphoNAS, morphogenesis, neural architecture search, reaction-diffusion systems, gene regulatory networks, evolutionary algorithms.
full text
REFERENCES
- 1. Zador A.M. A critique of pure learning and what artificial neural networks can learn from animal brains. Nature Communication. 2019. Vol. 10. Article number 3770. https://doi.org/10.1038/S41467-019-11786-6.
- 2. Friston K.J. Learning and inference in the brain. Neural Networks. 2003. Vol. 16, Iss. 9. P. 1325–1352. https://doi.org/10.1016/J.NEUNET.2003.06.005.
- 3. Turing A.M. The chemical basis of morphogenesis. Bulletin of Mathematical Biology. 1990. Vol. 52, Iss. 1–2. P. 153–197. https://doi.org/10.1016/S0092-8240(05)80008-4.
- 4. Thomas R. Boolean formalization of genetic control circuits. Journal of Theoretical Biology. 1973. Vol. 42, Iss. 3. P. 563–585. https://doi.org/10.1016/0022-5193(73)90247-6.
- 5. Polynikis A., Hogan S.J., di Bernardo M. Comparing different ODE modelling approaches for gene regulatory networks. Journal of Theoretical Biology. 2009. Vol 261, Iss. 4. P. 511–530. https://doi.org/10.1016/J.JTBI.2009.07.040.
- 6. Friston K.J., Levin M., Sengupta B., Pezzulo G. Knowing one’s place: A free-energy approach to pattern regulation. Journal of The Royal Society Interface. 2015. Vol. 12, Iss. 105. Article number 20141383. https://doi.org/10.1098/RSIF.2014.1383.
- 7. Friston K., Kilner J., Harrison L. A free energy principle for the brain. Journal of Physiology — Paris. 2006. Vol. 100, Iss. 1–3. P. 70–87. https://doi.org/10.1016/J.JPHYSPARIS.2006.10.001.
- 8. Babai L. Graph isomorphism in quasipolynomial time. Proc. Forty-Eighth Annual ACM Symposium on Theory of Computing (STOC’16) (19–21 June 2016, Cambridge MA, USA). Cambridge, USA. P.&nbso;684–697. https://doi.org/10.1145/2897518.2897542.
- 9. Edelman G.M., Gally J.A. Degeneracy and complexity in biological systems. Proceedings of the National Academy of Sciences. 2001. Vol. 98, N 24. P. 13763–13768. https://doi.org/10.1073/pnas.231499798.
- 10. Masel J., Trotter M.V. Robustness and evolvability. Trends Genet. 2010. Vol. 26, Iss. 9. P. 406–414. https://doi.org/10.1016/j.tig.2010.06.002.
- 11. De Jong K.A. An analysis of the behavior of a class of genetic adaptive systems: Ph.D. dissertation. Ann Arbor, Michigan, USA, 1975. 512 p. URL: https://search.proquest.com/openview/5bbdd6eb198f9e7d02353e994b3591c2/1?pq-origsite=gscholar&cbl=18750&diss=y.
- 12. Goldberg D.E., Deb K. A comparative analysis of selection schemes used in genetic algorithms. Foundations of Genetic Algorithms. 1991. Vol. 1. P. 69–93. https://doi.org/10.1016/B978-0-08-050684-5.50008-2.
- 13. ErdP., RБnyi A. On random graphs I. Publicationes mathematicae (Debrecen). 1959. Vol. 6. P. 290–297. https://doi.org/10.5486/PMD.1959.6.3-4.12.
- 14. Najarro E., Sudhakaran S., Risi S. Towards self-assembling artificial neural networks through neural developmental programs. arXiv:2307.08197v1 [cs.NE] 17 Jul 2023. https://doi.org/10.48550/arXiv.2307.08197.
- 15. Towers M., Kwiatkowski A., Terry J. et al. Gymnasium: A standard interface for reinforcement learning environments. arXiv:2407.17032v4 [cs.LG] 2 Nov 2025. https://doi.org/10.48550/arXiv.2407.17032.
- 16. Oller D., Glasmachers T., Cuccu G. Analyzing reinforcement learning benchmarks with random weight guessing. 19th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2020) (9–13 May 2020, Auckland, New Zealand, online event). Auckland, 2020. P. 975–982.
