CoPhe LAB

Join us!
Open positions available. Click to know more.

New papers!
Randomness in choice leads to cohesion in moving groups
Disentangling intrinsic and collective effects from data to perform better 'learning'

RESEARCH INTERESTS IN COPHE LAB

Welcome to the Collective Phenomena Lab (CoPhe LAB, in short; pronounced 'coffee').

Our research interests broadly encompass understanding the nonlinear interactions and group level dynamics in collective systems (made up of interacting agents). We look at problems across disciplines that include, droplets in a microchannel, cell migration, organismal motion, traffic and pedestrian dynamics. We use a combination of simple agent based models, optimisation techniques and data-driven approaches to study a variety of problems.

​

Emergence of collective phenomena.

We study how local interactions between agents in a collective lead to the emergence of group level phenomena that include robustness to perturbations, functionality that aids in survival of the group, etc. We use a combination of agent based models and tools from network science to discover the origins of emergence in a given system.

​

Engineering a collective.

We identify optimal operating and design conditions to manipulate collectives in desired ways. By posing them as optimisation problems, we successfully handle the large degrees of freedom that arise from the multi-agent nature of a collective system and the large combinatorial design space.

​

Collective phenomena meets data-science.

Often we are tasked with identifying the intrinsic properties of the agents, or their interactions with each other and the environment, or the equations governing the dynamics, from the data collected from the collective. We investigate how the inference problem depends on the physics of the collective. We develop hybrid physics cum data-based approaches to infer the hidden information underlying the dynamics of a collective.​