Talks

In this talk, I will describe how ideas from geometry and combinatorics can help us understand the mathematical and physical properties of cosmological integrals. From efficiently deriving canonical differential equations to a systematic method for finding a minimal basis for the physical subspace. Time permitting, I will also comment on how geometry and combinatorics control the zeros of cosmological integrands.

I will introduce how to incorporate loop contributions into positivity bounds in a massless scalar theory with shift symmetry. The allowed region of Wilson coefficients is called the EFT-hedron. With loop contributions, the n=4 null constraint, which plays an important role in determining the allowed region for g3–g4 (the Wilson coefficients of dimension-10 and dimension-12 operators, respectively), is modified. This modification of the null constraint reveals new features of the EFT-hedron. We obtained a much stronger bound for g2 (the Wilson coefficient of dimension-8 operators) without using the full unitarity condition. Furthermore, we found a negative g4, which is not possible under the tree-level bound. It would be interesting to generalize our method to more complicated theory and test our results.

All known human societies use gender to divide labor.  Extant game theoretic models in economics explain this division via appeal to rational predictions of what skill specializations will yield success in "marriage markets".  In this talk I argue that these high rationality assumptions are neither necessary nor appropriate in explaining gendered division of labor.  I use cultural evolutionary models to show how natural processes of learning and symmetry breaking predict and explain gendered division of labor.

Here we extend the Bala-Goyal framework to include differential levels of trust.  We will discuss how this modification might be used to model groups of scientists making judgments about the reliability of one another's work, and show how introducing trust dynamics can both slow learning and, in some cases, lead to stably polarized outcomes.  We will also discuss how agents learning over multiple domains can come to form epistemic factions, where unrelated beliefs become correlated.

In this general talk, we will look at the decisions that ants need to make in their day-to-day activities, by following their behaviours. Decisions in the context of stealing, learning, traffic jams and colony relocation using the model organisms Diacamma indicum a commonly found black ant in the Indian subcontinent. We will glimpse into both individual decisions and collective decisions made by these organisms. With a combination of control lab-based experiments and observations in the nature habitat we develop an understanding of the implications these decisions have on their survival and fitness.

This lecture examines the long-term behaviour of MEGs, focusing on fixation probabilities, fixation times, and stochastic slowdowns. It explores key questions such as if, when, and how a strategy persists in the long run. The transition from static equilibrium analysis to dynamic evolution is discussed, incorporating concepts like mutation-selection equilibrium, the 1/3 rule, risk dominance, and their generalisations to multiplayer settings. Additionally, the role of multiplayer games in mutualism is highlighted, showing how cooperative interactions persist over time in an ecological framework.

Continuous variable (CV) systems play an important role in quantum foundations and quantum information processing, especially because of their quantum optical realization. Gaussian states of CV systems, on the one hand have non-trivial quantum properties, and on the other hand, can be realized in the laboratory.  There are non-Gaussian states which can be obtained from Gaussian states via physically realizable operations and these too can be useful in enhancing  non-classicality and improving the performance of quantum information processing protocols.  In the talk, I will  introduce CV systems and their Gaussian and non-Gaussian states. The violation of Bell-type inequalities using CV systems will be discussed. Quantum key distribution protocols and quantum teleportation schemes using Gaussian and non-Gaussian states will also be taken up. 

This talk will introduce the Bala-Goyal model of epistemic networks, where agents on a network learn to solve a decision problem by performing actions and sharing the results of those actions with their neighbors.  We will discuss the conditions under which agents on these networks successfully learn to perform optimal outcomes, and how network structure can influence time to convergence and accuracy.

Many theorists have employed game theory to model the emergence of stable social norms, or natural “social contracts.” One branch of this literature uses bargaining games to show why many societies have norms and rules for fairness. In cultural evolutionary models, fair bargaining emerges endogenously because it is an efficient way to divide resources. But these models miss an important element of real human societies – divisions into groups or social categories. Once such groups are added to cultural evolutionary models, fairness is no longer the expected outcome.  Instead “discriminatory norms” often emerge where one group systematically gets more when dividing resources. I show why the addition of categories to bargaining models leads to unfairness, and discuss the role of power us in this process. I also address how categories might emerge to support inequity, and the possibility of modeling social change. Altogether this work emphasizes that if one wishes to understand the naturalistic emergence of social contracts, one must account for the presence of categorical divisions, and unfairness, as well as for norms of fairness.