The UvA Faculty of Science researchers who receive a grant from this round are Petra Bleeker (UvA-SILS), Sara Jabbari-Farouji (UvA-IoP) and Fernando Pascoal dos Santos  (UvA-IvI).

Self-defence of tomatoes

Petra Bleeker receives the ENW-M grant for the project ‘Trichome truckers: how self-defence metabolites are transported and stored in tomato gland cells’. Plants produce a large variety of specialised metabolites involved in stress response and interaction with the environment. Wild ancestors use specialised metabolites to protect themselves against pest and pathogens. In tomato self-defence metabolites are produced and stored in special glands, glandular trichomes. It is not yet known how they are transported from the cells that make them into the storage cavity for accumulation until needed. Bleeker and her team will elucidate this defence-metabolite transport. Understanding how wild tomato ancestors do this, might enable breeders to rearm cultivars with their own defence metabolites reducing their reliance on chemicals.

Smart filament-like materials

Sara Jabbari-Farouji receives the grant for the project ‘Understanding the Collective Patterns of Active Polymer-like Matter’. She and her team are ging to try and model the spectacular patterns of collective motion in birds and fishes. While simple statistical physics models successfully explain collective phenomena in some autonomous systems, real life is more complex. For example, biopolymers in the cell and blobs of entangled worms are not rigid, as assumed in models. Jabbari-Farouji will develop new theoretical and computational methods to understand patterns of collective behavior of semiflexible active filaments. This will help her team understand the biological function of collective self-organization and to uncover principles for designing smart filament-like materials with dynamic capabilities like shape-shifting or self-healing.

Cooperating artificial agents

Finally, Fernando Pascoal dos Santos is awarded the grant for the project ‘Reputation as a new route to cooperation in multi-agent reinforcement learning’. Artificial agents are likely to face the same dilemmas of cooperation that humans evolved to solve. To design efficient systems where collective gains are maximised, artificial agents must learn to forego their self-interest and spend effort to help others. How to design adaptive agents that autonomously learn to cooperate? In this project Pascoal dos Santos and his team study how to use reputation systems to sustain cooperation among agents adapting through trial-and-error (that is, through reinforcement learning). They will explore how reputations should be assigned to sustain cooperation in the long-run, and how to design reputation systems in increasingly complex environments.