Seminars

In this talk, we will discuss tree-level scattering of two massless and two massive particles with arbitrary integer spin (Compton Scattering). We discuss the three point functions, contributions of exchange diagrams and the role of contact terms. At the end, we discuss about the high energy behaviour and constraints on the coupling constants.

A long time ago, Newman and Janis showed that a complex transformation of the Schwarzschild solution leads to the Kerr solution. The Newman-Janis (NJ) algorithm on the space of classical solutions in GR and electromagnetism can be used in scattering amplitudes to map an amplitude with external scalar states to the one associated with the scattering of “infinite spin particles”. The minimal coupling of these particles to the gravitational or electromagnetic field corresponds to the classical coupling of the Kerr black hole with linearized gravity or the so-called $\sqrt{Kerr}$ charged state with the electromagnetic field. In this talk, I will discuss the idea of the NJ algorithm on the space of scalar QED amplitudes to compute classical observables such as the angular momentum impulse and the radiative field in the electromagnetic scattering of $\sqrt{Kerr}$ objects (analog of Kerr black holes in electromagnetism) at leading order in the coupling, via the Kosower, Maybee, O’Connell (KMOC) formalism. I will also discuss the relevance of the infinite hierarchy of the soft factorization theorems for gravitational tree-level amplitudes in the context of such classical (gravitational) scattering processes in four spacetime dimensions.

Randomized algorithms depend on accurate sampling from probability distributions, as their correctness and performance hinge on the quality of the generated samples. However, even for common distributions like Binomial, exact sampling is computationally challenging, leading standard library implementations to rely on heuristics. These heuristics, while efficient, suffer from approximation and system representation errors, causing deviations from the ideal distribution. Although seemingly minor, such deviations can accumulate in downstream applica- tions requiring large-scale sampling, potentially undermining algorithmic guarantees. We will take a look at this often overlooked issue related to correctness of randomized algorithms and have a discussion on what we can do to address this issue? This is a joint work with Kuldeep Meel and Uddalok Sarkar. This work will appear in CAV 2025.

"Industrial opportunities for PhD students"

"Industrial opportunities for PhD students"

Scattering amplitudes are the cornerstones of our current understanding of quantum field theories. On-shell methods offer efficient ways to calculate amplitudes in some theories and uncover beautiful structures usually obscured by the Feynman diagrammatics. The planar maximal supersymmetric (N=4) gauge theory is well suited for the use of on-shell methods, and the amplitudes have a positive geometry description in terms of the amplituhedron. We discuss the spontaneously symmetry broken N=4 super Yang-Mills theory and pure gauge theory in this talk. We study the so-called `on-shell functions' in the massive N=4 SYM, and realize the massive BCFW shifts as on-shell diagrams. We stumble upon mass deforming BCFW shifts. We discuss the maximal cut of simple loop diagrams and, using the generalized unitarity, find the loop amplitudes in massive N=4 SYM. In the second part of the talk, we discuss the pure gauge theory and realize its amplitudes from the positive geometry and combinatorics. The associahedron is a combinatorial object capturing the combinatorics of triangulations of an n-gon, and hence planar trivalent graphs. We make use of the Corolla polynomials to spin up the canonical form of the associahedron, yielding the gluon amplitudes. The similar Corolla polynomials for one loop lead us to the loop integrand of the n-gluon scattering. We use another representation of the Corolla polynomial to spin up the recently discovered curve integral formula.

Glassy dynamics in polymeric systems remain an area of intense interest, particularly in topologically constrained ring polymers, which exhibit unique relaxation mechanisms due to the absence of free ends and the presence of topological interactions. In this talk, I will present results from large-scale molecular dynamics simulations investigating the aging behavior of dense ring polymer systems with varying backbone stiffness. We identify the glass transition temperature (Tg) as a function of polymer stiffness and demonstrate that increasing stiffness enhances configurational constraints, leading to a higher Tg. Analysis of intermediate scattering function reveals that the relaxation time grows sublinearly with waiting time, highlighting the slow relaxation dynamics characteristic of these systems. A key feature of our system is the presence of threading interactions, which become increasingly persistent as the system ages. Our findings provide a comprehensive picture of aging in glassy ring polymer systems, demonstrating how topological constraints and polymer stiffness govern slow relaxation and non-equilibrium behavior. Reference: "Aging of ring polymeric topological glass formers via thermal quench", Arabinda Behera, Projesh Roy, Pinaki Chaudhuri, Satyavani Vemparala, arXiv:2504.02557, 2025.

-

We discuss results on the study of homogeneous spaces under connected linear algebraic groups beyond the classical case of number fields. https://www.imsc.res.in/~anupdixit/IMSc-CMI-NT-seminar.html

The generalised BMS group arose from the attempts to establish the equivalence of the soft graviton theorems with the conservation laws associated with asymptotic symmetries at null infinity. The canonical realization of this group at null infinity, however, requires treating the celestial metric as a variable in the gravitational phase space. This motivates an extension of the standard phase space defined in the early 80s by Ashtekar and Struebel. In this talk, we will introduce this extended phase space and present the Poisson brackets derived thereof through a detailed constraint analysis.

As computing permeates our lives, the reliability of algorithms has become indispensable. Particularly challenging is the task of ensuring the reliability of concurrent data structures, which underlie the design of fast parallel algorithms for the modern multicores in our phones, laptops, and data center servers. In this talk, I will describe some uses of concurrent data structures, motivate the need for their machine-verification, and describe an elegant technique for machine-verification of concurrent data structures, known as meta-configuration tracking [Jayanti, Jayanti, Yavuz, Hernandez: ACM POPL 2024]. This technique is simple to use, requiring only forward reasoning, yet it is universal and complete, i.e., powerful enough to prove the correctness of even intricate data structures with future and far-future dependent linearization structures. We have used meta-configuration tracking to machine-verify impactful data structures, including: the Jayanti-Tarjan concurrent union-find data structure which enables "parallel clustering algorithms which scale to graphs with tens of billions of edges" at Google [received Google Healthys Platinum Award], the ParlayHash data structure which is "the fastest concurrent hash table at Google" [received Google Healthys Gold Award], and MemSnap which is a fast far-future dependent snapshot object [received ACM SPAA 2025 Distinguished Paper Award].

Forecasting chaotic dynamical systems is a central challenge across science and engineering. In this talk, we will explore how random feature maps can be adapted to deliver remarkably strong performance on this task. A key ingredient for successful forecasting is ensuring that the features produced by the model lie in the nonlinear region of the activation function. We will see how this can be achieved through careful selection of the internal weights in a data-driven way using a hit-and-run algorithm. With a few additional modifications, such as increasing the depth of the model and introducing localization, we can achieve state-of-the-art forecasting results on a variety of high-dimensional chaotic systems, reaching up to 512 dimensions. Our method produces accurate short-term trajectory predictions, as well as reliable estimates of long-term statistical behavior in the test cases.