Token Systems are complex, hence it is hard for a human being to detect possible flaws in the system. Moreover, we can calculate token distribution or exchange rates but stakeholders’ behaviour is hard to predict. Machine learning can help us to tackle this challenge. This workshop session will give an introduction to agent-based models to simulate token system designs.
Piotr Grudzien, founder of Incentivai will provide a general overview on the potential of simulations and the specific approach of Incentivai. Here, behaviour is not an input to the simulation but rather an output that emerges based on the assumed agent objectives for the particular design. Agents are free to discover their own strategies and be modelled as deviating from rational behaviour, as needed. The simulation approach allows for testing large numbers of scenarios with varying design parameters and agent behaviour assumptions.
Incentivai was part of the Y Combinator Summer 2018 batch, has since published a number of case studies and worked with customers including Augur and NuCypher.
The session will focus on the following topics:
* Introduction to simulations, ML and agent-based models
* Motivation and inner workings of the Incentivai approach to agent-based modelling and simulation
* Interactive discussion and walkthrough of the Incentivai approach based on a bonding curve example
* Presentation of a case study (simulation of the Augur economy)
Learn more about Incentivai at http://incentivai.co and follow their work at https://medium.com/incentivai and on Twitter @incentivai or @GruPiotr.
Read more about the case studies Piotr will present:
Bonding curve case study blog post: https://medium.com/incentivai/bonding-curve-simulation-using-incentivai-2b2bfe0c6400
Simulation of the Augur economy blog post: https://medium.com/incentivai/simulation-of-the-augur-economy-682636d2840f