Research

Dicty - Social Amoeba Dictyostelium discoideum as an Inspiration for Higher-Order Emergence in Collective Adaptive Systems and Swarm Robotics

This project is supported by the Swiss National Science Foundation (SNSF) - grant number: 205321 179023

 

Contact person: Mohammad Parhizkar 

4pictures.001.jpeg

Introduction

Understanding collective behavior in nature and its potential links to the engineering of collective artificial behavior attracts many researchers from biology, computer science, and swarm robotics. It impacts different scientific and industrial topics such as cell-biology, cancer study, environment cleaning, swarms of drones, unmanned robots, and more generally collective adaptive systems based on IoT or massive ICT deployment. For instance, cancer cells exhibit collective behaviors, biomedicine researchers look for different examples from nature to design anti-cancer drugs to shrink tumors in human bodies. An interesting form of collective system is demonstrated by Dictyostelium discoideum and its multicellular development process. D. discoideum is a social amoeba able to change its behavior to survive in response to nutrient starvation. Most of its life, the organism lives in the soil as a single amoeba and feeds on bacteria. Individual cells move around on their own when there is plenty of food. Then, when food is scarce, the cells start a multicellular developmental process. Up to a million amoeboid cells artfully self-aggregate via pattern formation (first-order emergent behavior) to build a coherent and cohesive super-organism, similar to a motile slug structure. This complex super-organism has several properties that none of the cells has on its own (e.g. sensitivity to light and heat). The slug moves as a whole (second-order emergent behavior) looking for a suitable place to transform into a fruiting body in which about 20% of the cells die to lift the remaining cells up to a better place for sporulation and dispersal on the surface of the soil. Interestingly, at this point, the cells resume their individual behavior.

D. discoideum life cycle is an excellent example of the emergent phenomenon. These characteristics inspire us to investigate the relationship between first-order and higher-order collective behaviors in terms of emergence. Second- order emergent behavior arises from the interactions of individuals, which are themselves the result of first-order emergent societies. According to Sawyer’s book, second-order emergence, refers to systems in which agents recognize the existence of groups that emerged from their own collective behaviors. In the case of D. discoideum, higher-order emergent behavior refers to collective behavior at the level of slugs (themselves the result of collective behavior at the level of cells). Additionally, this social, relatively simple but yet powerful, the behavior is particularly appealing to inspire the engineering of collective adaptive systems, where a large number of simple homogeneous agents coordinate, self-organize and adapt themselves to environmental changes. The Dicty project, therefore, involves the combination of different disciplines - cell biology; self-organizing systems and swarm intelligence into one activity.

On the biology field, although D. discoideum provides a promising research path, not all phases are currently understood and modeled at the micro-level. From the computer science and artificial systems perspectives, first- order emergence is well studied, but higher-order levels of emergent behavior have not received much attention yet. Finally, from the collective adaptive systems perspective, there is no attempt at applying higher-order emergent behavior to this type of systems. The main objectives of this project are to:

  • Provide agent-based models of the different phases of D. discoideum life cycle.
  • Extract pertinent mechanisms for higher-order emergent behavior and provide them as design patterns for artificial systems.

 

Dicty project tackles the following research questions:

  • What are the social relations and configurations of D. discoideum behaviors at the different phases of its life cycle and how to model them? 

  • What are the mechanisms favoring higher-order emergence in swarms and artificial collective behavior? 

  • How to translate and implement those mechanisms into collective adaptive systems? 
Dicty will substantially advance the state of the art by providing:
    • Fine-grained understanding of D. discoideum individual cells behaviors at all phases of its life cycle and provision of corresponding agent-based models validated with actual biological experiments; 

    • Novel self-organizing mechanisms for higher-order emergent behaviors expressed and defined as design patterns for artificial systems. 


 

Phases of work:

  • Understanding and modeling D. discoideum first-order emergent behavior
.
  • Translating D. discoideum first-order collective behavior to swarms of robots.

  • Understanding and modeling D. discoideum higher-order emergent behavior.


        Additional step: Translating D. discoideum higher-order collective behavior into the swarm of Kilobots (50 units).

 

Partners

  • University of Lyon,  Prof. Salima Hassas – Professor of computer sciences - Leader of the Multi-Agents System (SMA) team at “LIRIS-CNRS” laboratory - Prof. Hassas works currently on the higher-order collective behavior of multi-agent systems. Prof. Hassas research interests relate to the area of Artificial Intelligence and Multi-Agents Systems and more specifically to Constructivist Approaches and Self-* (Self-Organization, Self-Adaptation, etc.) properties. From the computational perspective, Prof. Hassas is interested in the development of Intelligent Systems, with decentralized control and self-* properties, based on situated multi-agents systems. At a more fundamental level, Prof. Hassas co-supervised Lana de Carvalho PhD thesis and is interested in embodied and situated cognition, inactive approaches for cognition and developmental learning. 
  • University of Geneva, Prof. Thierry Soldati – Professor of biochemistry- Specialist of D. discoideum behavior. Former president of the society Life Sciences Switzerland (LS2) - Head of the Doctoral Program Biochemistry and Medicine for the attribution of PhD in Biochemistry- The major goal of Prof. Soldati’s group is to understand the integration, the cooperation of signaling, cytoskeleton and membrane trafficking in phagocytosis and its relevance to host-pathogen interactions. To this end, Prof. Soldati’s group uses the social amoeba Dictyostelium as a model organism as it is a professional phagocyte very similar to mammalian phagocytes of the innate immune system in morphology and behavior, but which is genetically and biochemically tractable. 
  • University of Graz, Prof. Thomas Schmickl – Professor of swarm robotics - Department of zoology - He is the founder of the Artificial Life Lab Graz, which is an interdisciplinary research lab hosting biologists, computer scientists, and simulation engineers. The laboratory researches swarm intelligence, biological self-organization, swarm robotics, modular robotics, and bio-inspired algorithms. He is an expert, project coordinator and Principal Investigator of numerous swarm robotics projects, among others: EU-funded Horizon 2020 project “subCULTRon”, “ASSISI-bf” 2013-2018, “CoCoRo” 2011-2014, FWF-funded project “REBODIMENT” 2012-2014, EU-FP7 project “Symbrion” 2008-2013 - Symbiotic Evolutionary Robot Organisms and “REPLICA- TOR” 2008-2013. In these projects, Prof. Schmickl’s team designed several real-world robot systems, developed several distributed control algorithms, communication approaches as well as a simulation platform (e.g. LaRoSim).

 

Publications

  • Conferences
  • Journal
    • Mohammad Parhizkar and Giovanna Di Marzo Serugendo, Agent-based models for first- and second-order emergent collective behaviors of social amoeba Dictyostelium discoideum aggregation and migration phases, Artificial Life and Robotics Journal, Volume 23, Issue 4, pp. 498–507, 2018 (link).

 

  • Annual Reports 
    •  Annual report 2015 (link) 

    • Annual report 2016 (link)
    • Annual report 2017 (link)
    • Annual report 2018 (link)

 

 

(Link to the simulation and validation videos in youtube)