Motivating Interactive Self-Organisation

LIFELIKE @ ALIFE 2020

Sebastian von Mammen

Self-Organisation

Müller-Schloer, Schmeck, and Ungerer (2011)

Consider a system \(S\) consisting of \(n\) components \(C\)
Each component \(C\) may interact with other components in the system
\(C\) exhibits a behaviour induced by a control mechanism \(M\)

Strong Self-Organisation

\(S\) is called strongly self-organising, if there are \(n\) control mechanisms \(M\).

Swarms

Local vs. global
Homogeneous vs. heterogeneous
Flat vs. hierarchical

Spatial
Observable
Interactive

Interactive Self-Organisation

Interactive self-organisation describes the effort to making large, self-organising technical systems transparent, malleable and controllable by human designers, decision makers and users.

von Mammen (2016)

Outline

Development Life Cycle
Individualised, Agent-Based & Interactive
Wetlab Application Scenario
Challenges

Outline

Development Life Cycle
Individualised, Agent-Based & Interactive
Wetlab Application Scenario
Challenges

Development Life Cycle

Discovery Phase

retrace the system’s and agents’ states and their interrelations

Development Phase

cast a model
pre-select, simplify, discretise
make assumptions

Exploration Phase

new relationships may emerge

Interactivity enriches the development at each step. But it also challenges the developer.

Outline

Development Life Cycle
Individualised, Agent-Based & Interactive
Wetlab Application Scenario
Challenges

Individualising Simulation

From consumer to creator.

Provision and (semi-)automated matching of template situations to fit the user’s challenges. (Mammen et al. 2019)
A natural interface for parametric adjustments, introducing additional programming code where necessary, navigating a simulation, as well as harnessing optimisation methods to improve one’s basis for decision making.

Modelling, Bottom-Up, Agent-Based

State-of-the-Art

Mathematical modelling tools (Matlab, Maple, Mathematica,…)
Requires professional expertise
Requires a lot of time

Bottom-Up

Elements, particles, agents,…
States and interactions \((Sit, Act, Dat, f_{Ag})\) (Denzinger and Kordt 2000)
Hierarchical organisation (Parry and Bithell 2012)

Interactive Simulation & Self-Organisation

Simulation Loop \(\rightarrow\) Human-in-the-Loop

(Rothrock and Narayanan 2011)

Interactive simulations synchronise with our minds.

Advantages

Understand systems, communicate systems, detect inconsistencies, learn about unaccounted aspects, immerse into emergency situations

… Corresponding Activities

Control, decision making, learning/training, analysis/debugging, refinement/engineering, empirical studies

Outline

Development Life Cycle
Individualised, Agent-Based & Interactive
Wetlab Application Scenario
Challenges

Wetlab Application Scenario

Davison, Samavati, and Jacob (2019)

Overview

Quickly prototype a model
Comprised of hundreds of thousands of agents
System analysis (pattern detection, convergence analysis,…)
Sensitivity analytical visualisations
Integration with larger, hierarchically organised system
Rich visualisation & multi-modal interaction
Bridge between virtual and actual reality by means of augmentation

Wetlab Application Scenario

Knote et al. (2019)

Use-Case Developmental Biology

Definition of intercellular interactions, biochemical and biophysical interactions
Observation of the emergence of morphological processes/convergence to metabolic equilibria

Add model details at different levels of scale, e.g.
- … genetic pathways
- … structural properties of the cells
Superimposition of high-level properties such as a cell’s surface tension
- … adjustment of lower-level parameters
Provide in-silico assays in analogy to in-vitro assays

Outline

Development Life Cycle
Individualised, Agent-Based & Interactive
Wetlab Application Scenario
Challenges
- Standardising Representations
- Innovating User Interfaces
- Reaching Application-Relevant Scales

Challenge: Standardising Representations

Arbitrary algorithmic descriptions

Require programming expertise
Defy comparability
Results in limited dissemination and lack of research continuity

Solution

Standardised model representations
Standardised primitives
Standardised numeric representation and integration

Formal protocolls rather involved
Algorithmic support of formal descriptions

Challenge: Innovating User Interfaces

\(D^2S\) Properties

Definition and selection of subpopulations
Creation and tracing of relationships (chains, cycles,…)
Recognise, fuel or disrupt phase transitions/emergent effects

Solution

Appripriate, helpful, efficient visualisation
Natural, efficient interaction routines
Support methods, e.g. deploying optimisers, to bridge levels of abstraction in behaviour, goal definitions, or regulating the agents’ degrees of autonomy

Merging modelling and simulation environments reduces indirections and prepares for real-world applications

Challenge: Reaching Application-Relevant Scales

\(D^2S\) Properties

Costly computations

Solution

Optimise and parallelise
Identify (partial) states and (partial) processes that allow for simplification of
- … the model
- … its computation
Dynamically put these simplifications in effect (build and revoke model hierarchies)

References

Davison, Timothy, Faramarz Samavati, and Christian Jacob. 2019. “Lifebrush: Painting, Simulating, and Visualizing Dense Biomolecular Environments.” Computers & Graphics 82. Elsevier: 232–42.

Denzinger, Jörg, and Michael Kordt. 2000. “Evolutionary on-Line Learning of Cooperative Behavior with Situation-Action-Pairs.” In Proceedings of the 4th International Conference on Multi-Agent Systems (Icmas 2000), 103–10. Boston, MA, USA.

Knote, A., S. C. Fischer, S. Cussat-Blanc, F. Niebling, D. Bernard, F. Cogoni, and S. von Mammen. 2019. “Immersive Analysis of 3D Multi-Cellular in-Vitro and in-Silico Cell Cultures.” In 2019 Ieee International Conference on Artificial Intelligence and Virtual Reality (Aivr), 82–827. doi:10.1109/AIVR46125.2019.00021.

Mammen, Sebastian von, Andreas Müller, Marc Erich Latoschik, Mario Botsch, Kirsten Brukamp, Carsten Schröder, and Michel Wacker. 2019. “VIA Vr: A Technology Platform for Virtual Adventures for Healthcare and Well-Being.” In 2019 11th International Conference on Virtual Worlds and Games for Serious Applications (Vs-Games), 1–2. IEEE.

Müller-Schloer, Christian, Hartmut Schmeck, and Theo Ungerer, eds. 2011. Organic Computing - a Paradigm Shift for Complex Systems. Autonomic Systems. Birkhäuser Verlag.

Parry, Hazel R, and Mike Bithell. 2012. “Large Scale Agent-Based Modelling: A Review and Guidelines for Model Scaling.” In Agent-Based Models of Geographical Systems, 271–308. Springer.

Rothrock, Ling, and S Narayanan. 2011. Human-in-the-Loop Simulations: Methods and Practice. Springer.

von Mammen, Sebastian. 2016. “Interactive Self-Organisation.” Habilitation Thesis, University of Augsburg.