Aims and Objectives
This group explores various issues related to dynamics of human societies with the help of Machine Learning and Data Sciences. The grand aim of the group is to develop a model of human societies (with a special focus on India), through which we can simulate the impact of various policies, natural disasters and climate change on the dynamics of the societies. We believe in multi-scale simulations, down to the level of individuals through agent-based modeling. The key challenge here lies in accurately representing the individual and group behavior/interaction characteristics of millions of such individuals, and their interactions with the "environment", in a computationally scalable way. We aim to use various concepts, algorithms and models from Machine Learning for this purpose.
Some of the vertical topics of research, i.e. specific aspects of human societies that we are currently focussing on (or have focussed on earlier) include
V1. Epidemiology (specifically how Covid-19 spread in Indian societies and how it can be controlled),
V2. Social Choice and representation (specifically, how political election outcomes are shaped by opinions, communities and geography)
V3. Mobility/transportation (especially how road safety can be improved through hyper-local interventions)
Some of the Machine Learning methodologies that we use for these research questions include:
M1. Agent-based Modeling
M2. Spatio-temporal Neural Networks (eg. ConvLSTM)
M3. Simulation-based Inference and Approximate Bayesian Computation
M4. Graph Neural Networks
Research Fellows
1) Ramanuj Bhattacharjee (PhD student since 2024) - multi-agent systems
2) Asmita Basu (project staff) - graph neural networks for transport networks
3) Pratiksha Jagadale (project staff) - Indian road accidents analysis
Ongoing and Past Collaborations
1) Prof. Palash Dey (IIT Kharagpur) - Election Modeling
2) Prof. Partha Pratim Chakraborty (IIT Kharagpur) - Epidemiological Modeling
3) Prof. Arkopal Kishore Goswami (IIT Kharagpur) - Mobility and Road Safety
4) Prof. Abhijnan Chakraborty (IIT Kharagpur) - Multi-agent systems
5) Dr. Rahul Verma (Centre for Policy Research) - Indian Elections
6) Ms. Aishwarya Raman (OMI Foundation) - Mobility and Road Safety
Publications
[C] Adway Mitra, Palash Dey (2024), International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), Evaluating District-based Election Surveys with Synthetic Dirichlet Likelihood (Supplementary Material)
[C] Adway Mitra (2023), International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), Agent-based Simulation of District-based Elections with Heterogeneous Populations (Extended Abstract)
[C] Varun Madhavan, Adway Mitra, Partha Pratim Chakraborti (2023), International Conference on Autonomous Agents and Multi-Agent Systems (AAMAS), Deep Learning-based Spatially Explicit Emulation of an Agent-Based Simulator for Pandemic in a City (Extended Abstract)
[C] Gaurav Suryawanshi, Varun Madhavan, Adway Mitra, Partha Pratim Chakrabarti (2021), Winter Simulation Conference, Phoenix, USA. City-scale Simulation of Covid-19 Pandemic and Intervention Policies using Agent-based Modelling
[C] Adway Mitra (2021), Winter Simulation Conference, Phoenix, USA. Electoral David vs Goliath: How does the Spatial Concentration of Electors affect District-based Elections?
Blog Posts
Simulating Indian General Election 2024
Products
We developed "Campus Modeler", a tool which allows us to simulate the activities in an university campus and visualize how a communicable disease like Covid-19 can spread through it. It was jointly developed with i-Hub Drishti Foundation of IIT Jodhpur and Prof. Rajesh Sundaresan's group from IISc Bangalore, funded by Department of Science and Technology (under the RAKSHAK scheme)