arXiv subjects

Sachin Kasture

Publications and source records attributed to Sachin Kasture.

16 recordsLinked to original sources

arXiv2025

Multiparticle quantum walks for distinguishing hard graphs

Quantum random walks have been shown to be powerful quantum algorithms for certain tasks on graphs like database searching, quantum simulations etc. In this work we focus on its applications for the graph isomorphism problem. In particular we look at how we can compare multi-particle quantum walks and well known classical WL tests and how quantum walks can be used to distinguish hard graphs like CFI graphs which k-WL tests cannot distinguish. We provide theoretical proofs and empirical results to show that a k-QW with input superposition states distinguishes k-CFI graphs. In addition we also prove that a k-1 QW with localized input states distinguishes k-CFI graphs. We also prove some additional results about strongly regular graphs (SRGs).

quant-ph↗

arXiv2024

Quantum Positional Encodings for Graph Neural Networks

In this work, we propose novel families of positional encodings tailored to graph neural networks obtained with quantum computers. These encodings leverage the long-range correlations inherent in quantum systems that arise from mapping the topology of a graph onto interactions between qubits in a quantum computer. Our inspiration stems from the recent advancements in quantum processing units, which offer computational capabilities beyond the reach of classical hardware. We prove that some of these quantum features are theoretically more expressive for certain graphs than the commonly used relative random walk probabilities. Empirically, we show that the performance of state-of-the-art models can be improved on standard benchmarks and large-scale datasets by computing tractable versions of quantum features. Our findings highlight the potential of leveraging quantum computing capabilities to enhance the performance of transformers in handling graph data.

quant-ph↗

arXiv2023

Enhancing Graph Neural Networks with Quantum Computed Encodings

Transformers are increasingly employed for graph data, demonstrating competitive performance in diverse tasks. To incorporate graph information into these models, it is essential to enhance node and edge features with positional encodings. In this work, we propose novel families of positional encodings tailored for graph transformers. These encodings leverage the long-range correlations inherent in quantum systems, which arise from mapping the topology of a graph onto interactions between qubits in a quantum computer. Our inspiration stems from the recent advancements in quantum processing units, which offer computational capabilities beyond the reach of classical hardware. We prove that some of these quantum features are theoretically more expressive for certain graphs than the commonly used relative random walk probabilities. Empirically, we show that the performance of state-of-the-art models can be improved on standard benchmarks and large-scale datasets by computing tractable versions of quantum features. Our findings highlight the potential of leveraging quantum computing capabilities to potentially enhance the performance of transformers in handling graph data.

quant-ph↗

arXiv2023

Protocols for classically training quantum generative models on probability distributions

Quantum Generative Modelling (QGM) relies on preparing quantum states and generating samples from these states as hidden - or known - probability distributions. As distributions from some classes of quantum states (circuits) are inherently hard to sample classically, QGM represents an excellent testbed for quantum supremacy experiments. Furthermore, generative tasks are increasingly relevant for industrial machine learning applications, and thus QGM is a strong candidate for demonstrating a practical quantum advantage. However, this requires that quantum circuits are trained to represent industrially relevant distributions, and the corresponding training stage has an extensive training cost for current quantum hardware in practice. In this work, we propose protocols for classical training of QGMs based on circuits of the specific type that admit an efficient gradient computation, while remaining hard to sample. In particular, we consider Instantaneous Quantum Polynomial (IQP) circuits and their extensions. Showing their classical simulability in terms of the time complexity, sparsity and anti-concentration properties, we develop a classically tractable way of simulating their output probability distributions, allowing classical training to a target probability distribution. The corresponding quantum sampling from IQPs can be performed efficiently, unlike when using classical sampling. We numerically demonstrate the end-to-end training of IQP circuits using probability distributions for up to 30 qubits on a regular desktop computer. When applied to industrially relevant distributions this combination of classical training with quantum sampling represents an avenue for reaching advantage in the NISQ era.

quant-ph↗

arXiv2021

Discovering dependencies in complex physical systems using Neural Networks

In todays age of data, discovering relationships between different variables is an interesting and a challenging problem. This problem becomes even more critical with regards to complex dynamical systems like weather forecasting and econometric models, which can show highly non-linear behavior. A method based on mutual information and deep neural networks is proposed as a versatile framework for discovering non-linear relationships ranging from functional dependencies to causality. We demonstrate the application of this method to actual multivariable non-linear dynamical systems. We also show that this method can find relationships even for datasets with small number of datapoints, as is often the case with empirical data.

physics.data-an↗

arXiv2018

Integrated photonic platform for quantum information with continuous variables

Integrated quantum photonics provides a scalable platform for the generation, manipulation, and detection of optical quantum states by confining light inside miniaturized waveguide circuits. Here we show the generation, manipulation, and interferometric stage of homodyne detection of non-classical light on a single device, a key step towards a fully integrated approach to quantum information with continuous variables. We use a dynamically reconfigurable lithium niobate waveguide network to generate and characterize squeezed vacuum and two-mode entangled states, key resources for several quantum communication and computing protocols. We measure a squeezing level of -1.38+-0.04 dB and demonstrate entanglement by verifying an inseparability criterion I=0.77+-0.02<1. Our platform can implement all the processes required for optical quantum technology and its high nonlinearity and fast reconfigurability makes it ideal for the realization of quantum computation with time encoded continuous variable cluster states.

quant-ph↗

arXiv2017

Direct characterization of a nonlinear photonic circuit's wave function with laser light

Integrated photonics is a leading platform for quantum technologies including nonclassical state generation \cite{Vergyris:2016-35975:SRP, Solntsev:2014-31007:PRX, Silverstone:2014-104:NPHOT, Solntsev:2016:RPH}, demonstration of quantum computational complexity \cite{Lamitral_NJP2016} and secure quantum communications \cite{Zhang:2014-130501:PRL}. As photonic circuits grow in complexity, full quantum tomography becomes impractical, and therefore an efficient method for their characterization \cite{Lobino:2008-563:SCI, Rahimi-Keshari:2011-13006:NJP} is essential. Here we propose and demonstrate a fast, reliable method for reconstructing the two-photon state produced by an arbitrary quadratically nonlinear optical circuit. By establishing a rigorous correspondence between the generated quantum state and classical sum-frequency generation measurements from laser light, we overcome the limitations of previous approaches for lossy multimode devices \cite{Liscidini:2013-193602:PRL, Helt:2015-1460:OL}. We applied this protocol to a multi-channel nonlinear waveguide network, and measured a 99.28$\pm$0.31\% fidelity between classical and quantum characterization. This technique enables fast and precise evaluation of nonlinear quantum photonic networks, a crucial step towards complex, large-scale, device production.

quant-ph↗

arXiv2016

Frequency conversion between UV and telecom wavelengths in a lithium niobate waveguide for quantum communication with Yb+ trapped ions

We study and demonstrate the frequency conversion of UV radiation, resonant with 369.5 nm transition in Yb+ ions to the C-band wavelength 1580.3 nm and vice-versa using a reverse proton-exchanged waveguide in periodically poled lithium niobate. Our integrated device can interface trapped Yb+ ions with telecom infrastructure for the realization of an Yb+ based quantum repeater protocol and to efficiently distribute entanglement over long distances. We analyse the single photon frequency conversion efficiency from the 369.525 nm to the telecom wavelength and its dependence on pump power, device length and temperature. The single-photon noise generated by spontaneous Raman scattering of the pump is also measured. From this analysis we estimate a single photon conversion efficiency of 9% is achievable with our technology with almost complete suppression of the Raman noise.

physics.optics↗

arXiv2016

Nine Channel Mid-Power Bipolar Pulse Generator Based on a Field Programmable Gate Array

Many channel arbitrary pulse sequence generation is required for the electro-optic reconfiguration of optical waveguide networks in Lithium Niobate. Here we describe a scalable solution to the requirement for mid-power bipolar parallel outputs, based on pulse patterns generated by an externally clocked field programmable gate array (FPGA). Positive and negative pulses can be generated at repetition rates from up to 80~MHz with pulse width adjustable in increments of 1.6~ns across nine independent outputs. Each channel can provide 1.5W of RF power and it can be synchronised with the operation of other components in an optical network such as light sources and detectors through an external clock with adjustable delay.

physics.ins-det↗

arXiv2015

Anisotropic model for the fabrication of Annealed and Reverse Proton Exchanged waveguides in congruent Lithium Niobate

An anisotropic model for the fabrication of annealed and reverse proton exchange waveguides in lithium niobate is presented. We characterized the anisotropic diffusion properties of proton exchange, annealing and reverse proton exchange in Z-cut and X-cut substrates using planar waveguides. Using this model we fabricated high quality channel waveguides with propagation losses as low as 0.086 dB/cm and a coupling efficiency with optical fiber of 90% at 1550 nm. The splitting ratio of a set of directional couplers is predicted with an accuracy of +- 0.06.

physics.optics↗

arXiv2013

Plasmonic quasicrystals for designable ultra broadband transmission enhancement and second harmonic generation

Quasi-crystals are intriguing as they exhibit rotational symmetry and long range ordering but lack translational symmetry. 2-dimensional metal-dielectric patterns are interesting to make use of surface plasmon polariton (SPP) mediated local field enhancement and for near dispersionless SPP modes. In plasmonic crystals, the orientation and periodicity of the pattern dictate the polarization response and the discrete plasmon resonances while the interfaces define the plasmon dispersion. However, unique properties of plasmonic quasicrystals lead to polarization independence, designable k-space and broadband transmission enhancement due to SPP mediation. These are useful in many applications like energy harvesting, nonlinear optics and quantum plasmonics. We demonstrate design and fabrication of large area quasicrystal air hole patterns of pi/5 symmetry in metal film in which broadband, launch angle and polarization independent transmission enhancement as well as broadband second harmonic generation are observed. Designable transmission response, other symmetries and tilings are possible.

physics.optics↗

arXiv2013

Superluminal propagation and broadband omnidirectional antireflection in optical reflectionless potentials

Reflectionless potentials (RPs) represent a class of potentials that offer total transmission in the context of one dimensional scattering. Optical realization of RPs in stratified medium can exhibit broadband omnidirectional antireflection property. In addition to the antireflection property, RPs are also expected to demonstrate negative delay. We designed refractive index profiles conforming to RPs and realize them in stratified media consisting of Al2O3 and TiO2 heterolayers. In these structures we observed < 1% reflection over the broad wavelength range of 350 nm to 2500 nm for angles of incidence 0 - 50 degrees. The observed reflection and transmission response of RPs are polarization independent. A negative delay of about 31 fsec with discernible pulse narrowing was observed in passage through two RPs. These RPs can be interesting for optical instrumentation as broadband, omni-directional antireflection coatings as well as in pulse control and transmission applications like delay lines.

physics.optics↗

arXiv2012

Strong coupling of in-plane plasmon modes and their control

We show anti-crossings due to strong in-plane coupling of plasmon modes in dielectric-metal-dielectric structure with top 2D dielectric pattern. Experimentally measured anti-crossing widths are compared with those calculated by coupled mode theory. It is shown that the coupling strength of the plasmon modes can be controlled by the orientation of the sample.

cond-mat.other↗

arXiv2012

Near dispersion-less surface plasmon polariton resonances at a metal-dielectric interface

Omni-directional light coupling to surface plasmon polariton (SPP) modes to make use of plasmon mediated near-field enhancement is challenging. We report possibility of near dispersion-less modes in structures with unpatterned metal-dielectric interfaces having 2-D dielectric patterns on top. We show that the position and dispersion of the excited modes can be controlled by the excitation geometry and the 2-D pattern. The anti-crossings resulting from the in-plane coupling of different SPP modes are also shown.

cond-mat.mtrl-sci↗

arXiv2012

Modulation of a surface plasmon-polariton resonance by sub-terahertz diffracted coherent phonons

Coherent sub-THz phonons incident on a gold grating that is deposited on a dielectric substrate undergo diffraction and thereby induce an alteration of the surface plasmon-polariton resonance. This results in efficient high-frequency modulation (up to 110 GHz) of the structure's reflectivity for visible light in the vicinity of the plasmon-polariton resonance. High modulation efficiency is achieved by designing a periodic nanostructure which provides both plasmon-polariton and phonon resonances. Our theoretical analysis shows that the dynamical alteration of the plasmon-polariton resonance is governed by modulation of the slit widths within the grating at the frequencies of higher-order phonon resonances.

cond-mat.mtrl-sci↗

arXiv2012

Interfacing a quantum dot spin with a photonic circuit

A scalable optical quantum information processor is likely to be a waveguide circuit with integrated sources, detectors, and either deterministic quantum-logic or quantum memory elements. With microsecond coherence times, ultrafast coherent control, and lifetime-limited transitions, semiconductor quantum-dot spins are a natural choice for the static qubits. However their integration with flying photonic qubits requires an on-chip spin-photon interface, which presents a fundamental problem: the spin-state is measured and controlled via circularly-polarised photons, but waveguides support only linear polarisation. We demonstrate here a solution based on two orthogonal photonic nanowires, in which the spin-state is mapped to a path-encoded photon, thus providing a blue-print for a scalable spin-photon network. Furthermore, for some devices we observe that the circular polarisation state is directly mapped to orthogonal nanowires. This result, which is physically surprising for a non-chiral structure, is shown to be related to the nano-positioning of the quantum-dot with respect to the photonic circuit.

cond-mat.mes-hall↗