Publications
Selected
Comment on "Spin-1/2 Kagome Heisenberg Antiferromagnet: Machine Learning Discovery of the Spinon Pair-Density-Wave Ground State"
Comment on a recent machine-learning study of the kagome Heisenberg antiferromagnet. We show that the reported low variational energies arise from broken ergodicity in Metropolis–Hastings sampling thus calling the original claims into question.
Optimizing the dynamical preparation of quantum spin lakes on the ruby lattice
Physical Review Letters 137, 093402
Using neural networks we demonstrate that exotic phases can be dynamically prepared in Rydberg quantum simulators at specific length scales, contrary to the equilibrium predictions.
* equal contribution
Earlier work
Alternative quantisation condition for wavepacket dynamics in a hyperbolic double well
Journal of Physics A: Mathematical and Theoretical 54, 035304
Proposed a new analytical way of finding allowed energies in the class of hyperbolic-double well potentials by connecting it to a problem of finding roots of some polynomial. Applied this approach to understanding the role of non-adiabatic effects during enhanced ionization.
Online Learning and matching for resource allocation problems
SIAM SIURO vol. 13
Devised, provided performance guarantees, and implemented algorithms integrating dual problems in convex optimization with a subclass of reinforcement learning techniques. Applied these algorithms to the traffic-shaping problem.
Analytical modelling of temperature effects on AMPA-type synapse
Journal of Computational Neuroscience 44, 379-391
Used ODE-based modelling for understanding temperature effects on AMPA-type synapses in brain. Simplified the ODEs using some physically-motivated assumptions and shown how the obtained analytical solution faithfully reproduces the results of biological experiments.
Refereeing
- Physical Review Letters (multiple times)
- ICML — AI4Science Workshop
- Physical Review A
- Quantum Information Processing (QIP)
- Quantum Computing Theory in Practice (QCTIP)