arXiv subjects

Saumya Biswas

Publications and source records attributed to Saumya Biswas.

6 recordsLinked to original sources

arXiv2026

An LLM-Orchestrated Agent for Directional-Coupler Design with Self-Consistent Eigenmode and FDTD Validation

We present a design agent which is a Large Language Model (LLM) that orchestrates, but does not perform, the numerical simulations to design a silicon-on-insulator (SOI) $2\times2$ directional coupler. We choose a symmetric phase-matched coupler where a lot of analytical results are available that help the design strategy. The LLM proposes candidate gap values (a geometrical dimension size) and judges convergence, while all physics is owned by deterministic solvers: a frequency-domain eigenmode solver estimates the coupling coefficient~$κ$ for the current design, and an independent Finite-Difference Time-Domain (FDTD) stage validates it. Both solvers operate on a common slab-projected two-dimensional (2D) effective-index reduction of the silicon film, so the design~$κ$ and the FDTD response are consistent by problem design; the residual between them is shown to be a single constant phase offset~$ϕ$, attributable to a fixed excess coupling length $L_{\mathrm{extra}}=\SI{2.837(11)}{\micro\meter}$ that we find invariant across a factor-of-two range in~$κ$. Folding this offset into a closed-loop length correction, the agent delivers a $50/50$ splitter whose FDTD-measured cross fraction is $0.498$ (target $0.500$), a residual of $0.0017$. Results are made self-consistent within the 2D effective-index model; and the LLM succeeds in delivering a suitable design over a number of attempts.

physics.optics↗

arXiv2025

PERM EQ x GRAPH EQ: Equivariant Neural Networks for Quantum Molecular Learning

In hierarchal order of molecular geometry, we compare the performances of Geometric Quantum Machine Learning models. Two molecular datasets are considered: the simplistic linear shaped LiH-molecule and the trigonal pyramidal molecule NH3. Both accuracy and generalizability metrics are considered. A classical equivariant model is used as a baseline for the performance comparison. The comparative performance of Quantum Machine Learning models with no symmetry equivariance, rotational and permutational equivariance, and graph embedded permutational equivariance is investigated. The performance differentials and the molecular geometry in question reveals the criteria for choice of models for generalizability. Graph embedding of features is shown to be an effective pathway to greater trainability for geometric datasets. Permutational symmetric embedding is found to be the most generalizable quantum Machine Learning model for geometric learning.

cs.LG↗

arXiv2025

HARLI CQUINN: Higher Adjusted Randomness with Linear In Complexity QUantum INspired Networks for K-Means

We contrast a minimalistic implementation of quantum k-means algorithm to classical k-means algorithm. With classical simulation results, we demonstrate a quantum performance, on and above par, with the classical k-means algorithm. We present benchmarks of its accuracy for test cases of both well-known and experimental datasets. Despite extensive research into quantum k-means algorithms, our approach reveals previously unexplored methodological improvements. The encoding step can be minimalistic with classical data imported into quantum states more directly than existing approaches. The proposed quantum-inspired algorithm performs better in terms of accuracy and Adjusted Rand Index (ARI) with respect to the bare classical k-means algorithm. By investigating multiple encoding strategies, we provide nuanced insights into quantum computational clustering techniques.

quant-ph↗

arXiv2024

Topological argument for robustness of coherent states in quantum optics

Coherent states, being the closest analog to classical states of wave systems, are well known to possess special properties that set them apart from most other quantum optical states. For example, they are robust against photon loss and do not easily get entangled upon interaction with a beamsplitter, and hence are called ``pointer states'', which is often attributed to them being eigenstates of the annihilation operator. Here we provide insights into a topological argument for their robustness using two separate but exact mappings of a prototypical quantum optics model - the driven Jaynes-Cummings model. The first mapping is based on bosonization and refermionization of the Jaynes-Cummings model into the fermionic Su-Schrieffer-Heeger model hosting zero-energy topologically protected edge states. The second mapping is based on the algebra of deformed f-oscillators. We choose these mappings to explicitly preserve the translational symmetry of the model along a Fock-state ladder basis, which is important for maintaining the symmetry-protected topology of such 1D lattices. In addition, we show that the edge state form is preserved even when certain chiral symmetry is broken, corresponding to a single-photon drive for the quantum optics model that preserves the coherent state; however, the addition of two-photon drive immediately disturbs the edge state form, as confirmed by numerical simulations of the mapped SSH model; this is expected since two-photon drive strongly perturbs the coherent state into a squeezed state. Our theory sheds light on a fundamental reason for the robustness of coherent states, both in existence and entanglement -- an underlying connection to topology.

quant-ph↗

arXiv2021

Detecting two photons with one molecule

We apply input-output theory with quantum pulses [AH Kiilerich, K Mølmer, Phys. Rev. Lett. {\bf 123}, 123604 (2019)] to a model of a new type of two-photon detector consisting of one molecule that can detect two photons arriving sequentially in time. The underlying process is distinct from the usual two-photon absorption process where two photons arriving simultaneously and with frequencies adding up to the resonance frequency are absorbed by a single molecule in one quantum jump. Our detector model includes a Hamiltonian description of the amplification process necessary to convert the microscopic change in the single molecule to a macroscopic signal.

quant-ph↗

arXiv2020

The Heisenberg picture of photodetection

We construct a class of Hamiltonians that describe the photodetection process from beginning to end. Our Hamiltonians describe the creation of a photon, how the photon travels to an absorber (such as a molecule), how the molecule absorbs the photon, and how the molecule after irreversibly changing its configuration triggers an amplification process---at a wavelength that may be very different from the photon's wavelength---thus producing a macroscopic signal. We use a simple prototype Hamiltonian to describe the single-photon detection process analytically in the Heisenberg picture, which neatly separates desirable from undesirable effects. Extensions to more complicated Hamiltonians are pointed out.

quant-ph↗