arXiv subjects

Archisman Ghosh

Publications and source records attributed to Archisman Ghosh.

At least 37 records · Page 2Linked to original sources

arXiv2025

The Art of Optimizing T-Depth for Quantum Error Correction in Large-Scale Quantum Computing

Quantum Error Correction (QEC), combined with magic state distillation, ensures fault tolerance in large-scale quantum computation. To apply QEC, a circuit must first be transformed into a non-Clifford (or T) gate set. T-depth, the number of sequential T-gate layers, determines the magic state cost, impacting both spatial and temporal overhead. Minimizing T-depth is crucial for optimizing resource efficiency in fault-tolerant quantum computing. While QEC scalability has been widely studied, T-depth reduction remains an overlooked challenge. We establish that T-depth reduction is an NP-hard problem and systematically evaluate multiple approximation techniques: greedy, divide-and-conquer, Lookahead-based brute force, and graph-based. The Lookahead-based brute-force algorithm (partition size 4) performs best, optimizing 90\% of reducible cases (i.e., circuits where at least one algorithm achieved optimization) with an average T-depth reduction of around 51\%. Additionally, we introduce an expansion factor-based identity gate insertion strategy, leveraging controlled redundancy to achieve deeper reductions in circuits initially classified as non-reducible. With this approach, we successfully convert up to 25\% of non-reducible circuits into reducible ones, while achieving an additional average reduction of up to 11.8\%. Furthermore, we analyze the impact of different expansion factor values and explore how varying the partition size in the Lookahead-based brute-force algorithm influences the quality of T-depth reduction.

quant-ph↗

arXiv2025

Rudraksh: A compact and lightweight post-quantum key-encapsulation mechanism

Resource-constrained devices such as wireless sensors and Internet of Things (IoT) devices have become ubiquitous in our digital ecosystem. These devices generate and handle a major part of our digital data. However, due to the impending threat of quantum computers on our existing public-key cryptographic schemes and the limited resources available on IoT devices, it is important to design lightweight post-quantum cryptographic (PQC) schemes suitable for these devices. In this work, we explored the design space of learning with error-based PQC schemes to design a lightweight key-encapsulation mechanism (KEM) suitable for resource-constrained devices. We have done a scrupulous and extensive analysis and evaluation of different design elements, such as polynomial size, field modulus structure, reduction algorithm, and secret and error distribution of an LWE-based KEM. Our explorations led to the proposal of a lightweight PQC-KEM, Rudraksh, without compromising security. Our scheme provides security against chosen ciphertext attacks (CCA) with more than 100 bits of Core-SVP post-quantum security and belongs to the NIST-level-I security category (provide security at least as much as AES-128). We have also shown how ASCON can be used for lightweight pseudo-random number generation and hash function in the lattice-based KEMs instead of the widely used Keccak for lightweight design. Our FPGA results show that Rudraksh currently requires the least area among the PQC KEMs of similar security. Our implementation of Rudraksh provides a $\sim3\times$ improvement in terms of the area requirement compared to the state-of-the-art area-optimized implementation of Kyber, can operate at $63\%$-$76\%$ higher frequency with respect to high-throughput Kyber, and improves time-area-product $\sim2\times$ compared to the state-of-the-art compact implementation of Kyber published in HPEC 2022.

cs.CR↗

arXiv2024

Forensics of Transpiled Quantum Circuits

Many third-party cloud providers set up quantum hardware as a service that includes a wide range of qubit technologies and architectures to maximize performance at minimal cost. However, there is little visibility to where the execution of the circuit is taking place. This situation is similar to the classical cloud. The difference in the quantum scenario is that the success of the user program is highly reliant on the backend used. Besides, the third-party provider may be untrustworthy and execute the quantum circuits on less efficient and more error-prone hardware to maximize profit. Thus, gaining visibility on the backend from various aspects will be valuable. Effective forensics can have many applications including establishing trust in quantum cloud services. We introduce the problem of forensics in the domain of quantum computing. We trace the coupling map of the hardware where the transpilation of the circuit took place from the transpiled program. We perform experiments on various coupling topologies (linear, T-shaped, H-shaped, and loop) on IBM backends. We can derive the coupling map from the transpiled circuits with complete accuracy for almost every transpiled circuit we considered. We could correctly trace 97.33% of the programs to the correct backend.

cs.CR↗

arXiv2024

Optimizing Quantum Embedding using Genetic Algorithm for QML Applications

Quantum Embeddings (QE) are essential for loading classical data into quantum systems for Quantum Machine Learning (QML). The performance of QML algorithms depends on the type of QE and how features are mapped to qubits. Traditionally, the optimal embedding is found through optimization, but we propose framing it as a search problem instead. In this work, we use a Genetic Algorithm (GA) to search for the best feature-to-qubit mapping. Experiments on the MNIST and Tiny ImageNet datasets show that GA outperforms random feature-to-qubit mappings, achieving 0.33-3.33 (MNIST) and 0.5-3.36 (Tiny ImageNet) higher fitness scores, with up to 15% (MNIST) and 8.8% (Tiny ImageNet) reduced runtime. The GA approach is scalable with both dataset size and qubit count. Compared to existing methods like Quantum Embedding Kernel (QEK), QAOA-based embedding, and QRAC, GA shows improvements of 1.003X, 1.03X, and 1.06X, respectively.

quant-ph↗

arXiv2024

Quantum Prometheus: Defying Overhead with Recycled Ancillas in Quantum Error Correction

Quantum error correction (QEC) is crucial for ensuring the reliability of quantum computers. However, implementing QEC often requires a significant number of qubits, leading to substantial overhead. One of the major challenges in quantum computing is reducing this overhead, especially since QEC codes depend heavily on ancilla qubits for stabilizer measurements. In this work, we propose reducing the number of ancilla qubits by reusing the same ancilla qubits for both X- and Z-type stabilizers. This is achieved by alternating between X and Z stabilizer measurements during each half-round, cutting the number of required ancilla qubits in half. This technique can be applied broadly across various QEC codes, we focus on rotated surface codes only and achieve nearly $25\%$ reduction in total qubit overhead. We also present a few use cases where the proposed idea enables the usage of higher-distance surface codes at a relatively lesser qubit count. Our analysis shows that the modified approach enables users to achieve similar or better error correction with fewer qubits, especially for higher distances ($d \geq 13$). Additionally, we identify conditions where the modified code allows for extended distances ($d + k$) while using the same or fewer resources as the original, offering a scalable and practical solution for quantum error correction. These findings emphasize the modified surface code's potential to optimize qubit usage in resource-constrained quantum systems.

quant-ph↗

arXiv2024

A 0.03${mm}^2$ 100-250MHz Charge-Pump or Amplifier-Less Integrating Sub-Sampling PLL for Ultra-low Power Communication and Computing

Clock generation is an essential part of wireless or wireline communication modules. To facilitate recent advancements in wireline-like communication and in-sensor computing modules at relatively lower data rates, ultra-low power, and accurate clock generation are of the utmost importance. This paper presents a unique implementation of integrating sub-sampling phase locked loop, which alleviates the usage of additional gain elements in the PLL and reduces the noise injection in the system. In this design, the ring oscillator-based PLL can operate a wide frequency range of 100-250MHz while consuming 0.03mm2 of area and 131.8$μW$ of power at 250MHz. The area-normalized figure of merit (FOM) of the integrating SSPLL is found to be -236, while showing a reference spur of -43.2dB.

eess.SP↗

arXiv2024

AI-driven Reverse Engineering of QML Models

Quantum machine learning (QML) is a rapidly emerging area of research, driven by the capabilities of Noisy Intermediate-Scale Quantum (NISQ) devices. With the progress in the research of QML models, there is a rise in third-party quantum cloud services to cater to the increasing demand for resources. New security concerns surface, specifically regarding the protection of intellectual property (IP) from untrustworthy service providers. One of the most pressing risks is the potential for reverse engineering (RE) by malicious actors who may steal proprietary quantum IPs such as trained parameters and QML architecture, modify them to remove additional watermarks or signatures and re-transpile them for other quantum hardware. Prior work presents a brute force approach to RE the QML parameters which takes exponential time overhead. In this paper, we introduce an autoencoder-based approach to extract the parameters from transpiled QML models deployed on untrusted third-party vendors. We experiment on multi-qubit classifiers and note that they can be reverse-engineered under restricted conditions with a mean error of order 10^-1. The amount of time taken to prepare the dataset and train the model to reverse engineer the QML circuit being of the order 10^3 seconds (which is 10^2x better than the previously reported value for 4-layered 4-qubit classifiers) makes the threat of RE highly potent, underscoring the need for continued development of effective defenses.

quant-ph↗

arXiv2024

R-STELLAR: A Resilient Synthesizable Signature Attenuation SCA Protection on AES-256 with built-in Attack-on-Countermeasure Detection

Side channel attacks (SCAs) remain a significant threat to the security of cryptographic systems in modern embedded devices. Even mathematically secure cryptographic algorithms, when implemented in hardware, inadvertently leak information through physical side channel signatures such as power consumption, electromagnetic (EM) radiation, light emissions, and acoustic emanations. Exploiting these side channels significantly reduces the search space of the attacker. In recent years, physical countermeasures have significantly increased the minimum traces to disclosure (MTD) to 1 billion. Among them, signature attenuation is the first method to achieve this mark. Signature attenuation often relies on analog techniques, and digital signature attenuation reduces MTD to 20 million, requiring additional methods for high resilience. We focus on improving the digital signature attenuation by an order of magnitude (MTD 200M). Additionally, we explore possible attacks against signature attenuation countermeasure. We introduce a Voltage drop Linear region Biasing (VLB) attack technique that reduces the MTD to over 2000 times less than the previous threshold. This is the first known attack against a physical side-channel attack (SCA) countermeasure. We have implemented an attack detector with a response time of 0.8 milliseconds to detect such attacks, limiting SCA leakage window to sub-ms, which is insufficient for a successful attack.

cs.CR↗

arXiv2024

The Quantum Imitation Game: Reverse Engineering of Quantum Machine Learning Models

Quantum Machine Learning (QML) amalgamates quantum computing paradigms with machine learning models, providing significant prospects for solving complex problems. However, with the expansion of numerous third-party vendors in the Noisy Intermediate-Scale Quantum (NISQ) era of quantum computing, the security of QML models is of prime importance, particularly against reverse engineering, which could expose trained parameters and algorithms of the models. We assume the untrusted quantum cloud provider is an adversary having white-box access to the transpiled user-designed trained QML model during inference. Reverse engineering (RE) to extract the pre-transpiled QML circuit will enable re-transpilation and usage of the model for various hardware with completely different native gate sets and even different qubit technology. Such flexibility may not be obtained from the transpiled circuit which is tied to a particular hardware and qubit technology. The information about the number of parameters, and optimized values can allow further training of the QML model to alter the QML model, tamper with the watermark, and/or embed their own watermark or refine the model for other purposes. In this first effort to investigate the RE of QML circuits, we perform RE and compare the training accuracy of original and reverse-engineered Quantum Neural Networks (QNNs) of various sizes. We note that multi-qubit classifiers can be reverse-engineered under specific conditions with a mean error of order 1e-2 in a reasonable time. We also propose adding dummy fixed parametric gates in the QML models to increase the RE overhead for defense. For instance, adding 2 dummy qubits and 2 layers increases the overhead by ~1.76 times for a classifier with 2 qubits and 3 layers with a performance overhead of less than 9%. We note that RE is a very powerful attack model which warrants further efforts on defenses.

quant-ph↗

arXiv2024

Approximate DCT and Quantization Techniques for Energy-Constrained Image Sensors

Recent expansions in multimedia devices gather enormous amounts of real-time images for processing and inference. The images are first compressed using compression schemes, like JPEG, to reduce storage costs and power for transmitting the captured data. Due to inherent error resilience and imperceptibility in images, JPEG can be approximated to reduce the required computation power and area. This work demonstrates the first end-to-end approximation computing-based optimization of JPEG hardware using i) an approximate division realized using bit-shift operators to reduce the complexity of the quantization block, ii) loop perforation, and iii) precision scaling on top of a multiplier-less fast DCT architecture to achieve an extremely energy-efficient JPEG compression unit which will be a perfect fit for power/bandwidth-limited scenario. Furthermore, a gradient descent-based heuristic composed of two conventional approximation strategies, i.e., Precision Scaling and Loop Perforation, is implemented for tuning the degree of approximation to trade off energy consumption with the quality degradation of the decoded image. The entire RTL design is coded in Verilog HDL, synthesized, mapped to TSMC 65nm CMOS technology, and simulated using Cadence Spectre Simulator under 25$^{\circ}$\textbf{C}, TT corner. The approximate division approach achieved around $\textbf{28\%}$ reduction in the active design area. The heuristic-based approximation technique combined with accelerator optimization achieves a significant energy reduction of $\textbf{36\%}$ for a minimal image quality degradation of $\textbf{2\%}$ SAD. Simulation results also show that the proposed architecture consumes 15uW at the DCT and quantization stages to compress a colored 480p image at 6fps.

eess.IV↗

arXiv2024

Guardians of the Quantum GAN

Quantum Generative Adversarial Networks (qGANs) are at the forefront of image-generating quantum machine learning models. To accommodate the growing demand for Noisy Intermediate-Scale Quantum (NISQ) devices to train and infer quantum machine learning models, the number of third-party vendors offering quantum hardware as a service is expected to rise. This expansion introduces the risk of untrusted vendors potentially stealing proprietary information from the quantum machine learning models. To address this concern we propose a novel watermarking technique that exploits the noise signature embedded during the training phase of qGANs as a non-invasive watermark. The watermark is identifiable in the images generated by the qGAN allowing us to trace the specific quantum hardware used during training hence providing strong proof of ownership. To further enhance the security robustness, we propose the training of qGANs on a sequence of multiple quantum hardware, embedding a complex watermark comprising the noise signatures of all the training hardware that is difficult for adversaries to replicate. We also develop a machine learning classifier to extract this watermark robustly, thereby identifying the training hardware (or the suite of hardware) from the images generated by the qGAN validating the authenticity of the model. We note that the watermark signature is robust against inferencing on hardware different than the hardware that was used for training. We obtain watermark extraction accuracy of 100% and ~90% for training the qGAN on individual and multiple quantum hardware setups (and inferencing on different hardware), respectively. Since parameter evolution during training is strongly modulated by quantum noise, the proposed watermark can be extended to other quantum machine learning models as well.

quant-ph↗

arXiv2024

Application of Quantum Tensor Networks for Protein Classification

We show that protein sequences can be thought of as sentences in natural language processing and can be parsed using the existing Quantum Natural Language framework into parameterized quantum circuits of reasonable qubits, which can be trained to solve various protein-related machine-learning problems. We classify proteins based on their subcellular locations, a pivotal task in bioinformatics that is key to understanding biological processes and disease mechanisms. Leveraging the quantum-enhanced processing capabilities, we demonstrate that Quantum Tensor Networks (QTN) can effectively handle the complexity and diversity of protein sequences. We present a detailed methodology that adapts QTN architectures to the nuanced requirements of protein data, supported by comprehensive experimental results. We demonstrate two distinct QTNs, inspired by classical recurrent neural networks (RNN) and convolutional neural networks (CNN), to solve the binary classification task mentioned above. Our top-performing quantum model has achieved a 94% accuracy rate, which is comparable to the performance of a classical model that uses the ESM2 protein language model embeddings. It's noteworthy that the ESM2 model is extremely large, containing 8 million parameters in its smallest configuration, whereas our best quantum model requires only around 800 parameters. We demonstrate that these hybrid models exhibit promising performance, showcasing their potential to compete with classical models of similar complexity.

quant-ph↗

arXiv2023

A novel approach to infer population and cosmological properties with gravitational waves standard sirens and galaxy surveys

Gravitational wave (GW) sources at cosmological distances can be used to probe the expansion rate of the Universe. GWs directly provide a distance estimation of the source but no direct information on its redshift. The optimal scenario to obtain a redshift is through the direct identification of an electromagnetic (EM) counterpart and its host galaxy. With almost 100 GW sources detected without EM counterparts (dark sirens), it is becoming crucial to have statistical techniques able to perform cosmological studies in the absence of EM emission. Currently, only two techniques for dark sirens are used on GW observations: the spectral siren method, which is based on the source-frame mass distribution to estimate conjointly cosmology and the source's merger rate, and the galaxy survey method, which uses galaxy surveys to assign a probabilistic redshift to the source while fitting cosmology. It has been recognized, however, that these two methods are two sides of the same coin. In this paper, we present a novel approach to unify these two methods. We apply this approach to several observed GW events using the \textsc{glade+} galaxy catalog discussing limiting cases. We provide estimates of the Hubble constant, modified gravity propagation effects, and population properties for binary black holes. We also estimate the binary black hole merger rate per galaxy to be $10^{-6}-10^{-5} {\rm yr^{-1}}$ depending on the galaxy catalog hypotheses.

astro-ph.CO↗

arXiv2023

ICAROGW: A python package for inference of astrophysical population properties of noisy, heterogeneous and incomplete observations

We present icarogw 2.0, a pure CPU/GPU python code developed to infer astrophysical and cosmological population properties of noisy, heterogeneous, and incomplete observations. icarogw 2.0 is mainly developed for compact binary coalescence (CBC) population inference with gravitational wave (GW) observations. The code contains several models for masses, spins, and redshift of CBC distributions, and is able to infer population distributions as well as the cosmological parameters and possible general relativity deviations at cosmological scales. We present the theoretical and computational foundations of icarogw 2.0, and we describe how the code can be employed for population and cosmological inference using (i) only GWs, (ii) GWs and galaxy surveys and (iii) GWs with electromagnetic counterparts. We discuss the code performance on Graphical Processing Units (GPUs), finding a gain in computation time of about two orders of magnitudes when more than 100 GW events are involved for the analysis. We validate the code by re-analyzing GW population and cosmological studies, finding very good agreement with previous publications.

astro-ph.CO↗

arXiv2023

Joint cosmological and gravitational-wave population inference using dark sirens and galaxy catalogues

In the absence of numerous gravitational-wave detections with confirmed electromagnetic counterparts, the "dark siren" method has emerged as a leading technique of gravitational-wave cosmology. The method allows redshift information of such events to be inferred statistically from a catalogue of potential host galaxies. Due to selection effects, dark siren analyses necessarily depend on the mass distribution of compact objects and the evolution of their merger rate with redshift. Informative priors on these quantities will impact the inferred posterior constraints on the Hubble constant ($H_0$). It is thus crucial to vary these unknown distributions during an $H_0$ inference. This was not possible in earlier analyses due to the high computational cost, restricting them to either excluding galaxy catalogue information, or fixing the gravitational-wave population mass distribution and risking introducing bias to the $H_0$ measurement. This paper introduces a significantly enhanced version of the Python package GWCOSMO, which allows joint estimation of cosmological and compact binary population parameters. This thereby ensures the analysis is now robust to a major source of potential bias. The gravitational-wave events from the Third Gravitational-Wave Transient Catalogue are reanalysed with the GLADE+ galaxy catalogue, and an updated, more reliable measurement of $H_0=69^{+12}_{-7}$ km s$^{-1}$ Mpc$^{-1}$ is found (maximum a posteriori probability and 68% highest density interval). This improved method will enable cosmological analyses with future gravitational-wave detections to make full use of the information available (both from galaxy catalogues and the compact binary population itself), leading to promising new independent bounds on the Hubble constant.

astro-ph.CO↗

arXiv2023

Impact of modelling galaxy redshift uncertainties on the gravitational-wave dark standard siren measurement of the Hubble constant

Gravitational wave science is a new and rapidly expanding field of observational astronomy. Multimessenger observations of the binary neutron star merger GW170817 have provided some iconic results including the first gravitational-wave standard-siren measurement of the Hubble constant, opening up a new way to probe cosmology. The majority of the compact binary sources observed in gravitational waves are however without bright electromagnetic counterparts. In these cases, one can fall back on the ``dark standard siren'' approach to include information statistically from potential host galaxies. For such a measurement, we need to be cautious about all possible sources of systematic errors. In this paper, we begin to study the possible errors coming from the galaxy catalogue sector, and in particular, look into the effect of galaxy redshift uncertainties for the cases where these are photometry-based. We recalculate the dark standard siren Hubble constant using the latest GWTC-3 events and associated galaxy catalogues, with different galaxy redshift uncertainty models, namely, the standard Gaussian, a modified Lorentzian, and no uncertainty at all. We find that not using redshift uncertainties at all can lead to a potential bias comparable with other potential systematic effects previously considered for the GWTC-3 $H_0$ measurement (however still small compared to the overall statistical error in this measurement). The difference between different uncertainty models leads to small differences in the results for the current data; their impact is much smaller than the current statistical errors and other potential sources of systematic errors which have been considered in previous robustness studies.

gr-qc↗

arXiv2023

Science with the Einstein Telescope: a comparison of different designs

The Einstein Telescope (ET), the European project for a third-generation gravitational-wave detector, has a reference configuration based on a triangular shape consisting of three nested detectors with 10 km arms, where in each arm there is a `xylophone' configuration made of an interferometer tuned toward high frequencies, and an interferometer tuned toward low frequencies and working at cryogenic temperature. Here, we examine the scientific perspectives under possible variations of this reference design. We perform a detailed evaluation of the science case for a single triangular geometry observatory, and we compare it with the results obtained for a network of two L-shaped detectors (either parallel or misaligned) located in Europe, considering different choices of arm-length for both the triangle and the 2L geometries. We also study how the science output changes in the absence of the low-frequency instrument, both for the triangle and the 2L configurations. We examine a broad class of simple `metrics' that quantify the science output, related to compact binary coalescences, multi-messenger astronomy and stochastic backgrounds, and we then examine the impact of different detector designs on a more specific set of scientific objectives.

gr-qc↗

arXiv2023

A 334$μ$W 0.158mm$^2$ ASIC for Post-Quantum Key-Encapsulation Mechanism Saber with Low-latency Striding Toom-Cook Multiplication Authors Version

The hard mathematical problems that assure the security of our current public-key cryptography (RSA, ECC) are broken if and when a quantum computer appears rendering them ineffective for use in the quantum era. Lattice based cryptography is a novel approach to public key cryptography, of which the mathematical investigation (so far) resists attacks from quantum computers. By choosing a module learning with errors (MLWE) algorithm as the next standard, National Institute of Standard & Technology (NIST) follows this approach. The multiplication of polynomials is the central bottleneck in the computation of lattice based cryptography. Because public key cryptography is mostly used to establish common secret keys, focus is on compact area, power and energy budget and to a lesser extent on throughput or latency. While most other work focuses on optimizing number theoretic transform (NTT) based multiplications, in this paper we highly optimize a Toom-Cook based multiplier. We demonstrate that a memory-efficient striding Toom-Cook with lazy interpolation, results in a highly compact, low power implementation, which on top enables a very regular memory access scheme. To demonstrate the efficiency, we integrate this multiplier into a Saber post-quantum accelerator, one of the four NIST finalists. Algorithmic innovation to reduce active memory, timely clock gating and shift-add multiplier has helped to achieve 38% less power than state-of-the art PQC core, 4x less memory, 36.8% reduction in multiplier energy and 118x reduction in active power with respect to state-of-the-art Saber accelerator (not silicon verified). This accelerator consumes 0.158mm2 active area which is lowest reported till date despite process disadvantages of the state-of-the-art designs.

cs.CR↗