Загрузка видео...

Не удалось загрузить видео

На главную

Synthetic ecosystems that autonomously and continuously evolve in silico 👾 An Alifer dream we pursue with Flow-Lenia ! If you are interested in complex systems with (1) emergent creatures and (2) intrinsic evolutionary dynamics, go check our new paper ! A 🧵

206,715 просмотров • 1 год назад •via X (Twitter)

Комментарии: 11

Фото профиля erwan plantec
erwan plantec1 год назад

FL allows to embed the CA parameters within its dynamics, making them localised, just like genomes ! This makes FL a unique system to simulate multi-species environments where rules can compete for mass, spread and undergo mutations: all the ingredients for evolution to emerge !

Фото профиля erwan plantec
erwan plantec1 год назад

We go larger scales in this work by simulating larger environments for longer ! We then use evolutionary activity metrics (see a framework designed to analyse systems that evolve under intrinsic fitnesses, to evaluate the long-term evolutionary dynamics

Фото профиля erwan plantec
erwan plantec1 год назад

We use these metrics to evaluate the evolutionary dynamics emerging from Flow-Lenia local rules in different conditions and also visualise emerging evolutionary trees:

Фото профиля erwan plantec
erwan plantec1 год назад

What if we could then engineer synthetic ecosystems to test our theories on the emergence of life and the evolution of complexity ? That’s what we propose by designing 2 additional environments within FL and measure their effects on the long-term behaviour of the system:

Фото профиля erwan plantec
erwan plantec1 год назад

1. A dissipative environment with constant in- and out-flows of matter(=energy) 2. An environment where creatures need to consume, and compete for, external resources in order to counter their inevitable dissipation.

Фото профиля erwan plantec
erwan plantec1 год назад

While we observe evolutionary dynamics in the system they still tend to settle and produce bounded complexity. The quest for open-ended dynamics is then still on ! In your opinion what is lacking now for more complexity to evolve ?

Фото профиля erwan plantec
erwan plantec1 год назад

This work also recapitulates previous results with the Flow-Lenia system Like the discovery of complex emergent creatures form random search only:

Фото профиля erwan plantec
erwan plantec1 год назад

And the optimization of the simulation parameters towards desired objectives, e.g creatures sensing and climbing gradients :

Фото профиля erwan plantec
erwan plantec1 год назад

With @hamongautier, @mayalen_etc, @BertChakovsky, @pyoudeyer and @ClemMoulinFrier ! Go check our paper for more ! 📄 📄

Фото профиля Mobile Scanner
Mobile Scanner1 год назад

Scan any documents, convert images into text, PDF files, etc. 👍

Фото профиля Bio Protocol
Bio Protocol1 год назад

This is literally a synthetic biologist’s dream

Похожие видео

Introducing Digital Red Queen (DRQ): Adversarial Program Evolution in Core War with LLMs Blog: Core War is a programming game where self-replicating assembly programs, called warriors, compete for control of a virtual machine. In this dynamic environment, where there is no distinction between code and data, warriors must crash opponents while defending themselves to survive. In this work, we explore how LLMs can drive open-ended adversarial evolution of these programs within Core War. Our approach is inspired by the Red Queen Hypothesis from evolutionary biology: the principle that species must continually adapt and evolve simply to survive against ever-changing competitors. We found that running our DRQ algorithm for longer durations produces warriors that become more generally robust. Most notably, we observed an emergent pressure towards convergent evolution. Independent runs, starting from completely different initial conditions, evolved toward similar general-purpose behaviors—mirroring how distinct species in nature often evolve similar traits to solve the same problems. Simulating these adversarial dynamics in an isolated sandbox offers a glimpse into the future, where deployed LLM systems might eventually compete against one another for computational or physical resources in the real world. This project is a collaboration between MIT and Sakana AI led by Akarsh Kumar Full Paper (Website): Full Paper (arxiv): Code:

Sakana AI

143,892 просмотров • 7 месяцев назад