arXiv subjects

Harsh Gupta

Publications and source records attributed to Harsh Gupta.

At least 19 recordsLinked to original sources

arXiv2026

Substrate-metal interface engineering enhances TaN/Ta thin film superconducting resonator performance

Tantalum has been demonstrated as a promising material for superconducting qubits. However, comparatively little attention has been given to its nitrides. Tantalum nitride exhibits a range of stoichiometries, resulting in a variety of material properties, including both superconducting and non-superconducting phases. Owing to this versatility, tantalum nitrides can serve multiple purposes in superconducting qubits: as seed layers for alpha-Ta growth, as a superconducting base material and as a non-superconducting barrier in the Josephson junction. In this study, we explore the performance of superconducting TaN and Ta thin film combinations on silicon substrates in terms of internal quality factor Qi. We find that standalone TaN films exhibit Qi values of about 1.5x10^5 at 100mK in the single-photon regime. Surprisingly, a resonator made from Ta grown on a few-nanometers-thick TaN seed layer yields largely the same performance. However, adding an additional, few-nanometers-thick Ta buffer layer between the Si substrate and this TaN seed layer enhances Qi significantly up to 5.9x10^5. Supporting transmission electron microscopy measurements reveal nitrogen accumulation and structural disorder at the TaN-Si interface, while this interfacial modification is suppressed when the Ta buffer layer is introduced. The observed improvement in resonator performance is consistent with a reduction of interface-related two-level system losses and strongly supports the hypothesis that controlling the substrate-metal interface is pivotal for the performance of superconducting qubit circuitry.

cond-mat.mtrl-sci↗

arXiv2026

Interfacial Strain and Structural Defects Govern the Performance of Tantalum Superconducting Waveguide Resonators

Tantalum (Ta) is a promising material for reaching long coherence times in superconducting qubits. A detailed understanding of the underlying structure-property relationship remains elusive though. In the present study, we sputter-deposited 200 nm thick Ta films on high-resistivity silicon (100) substrates at temperatures ranging from T = 20{\deg}C to 600{\deg}C, as well as on different seed layers (Nb, TiN and TaN). Alpha-Ta thin films were readily obtained at temperatures above 500{\deg}C and on all seed layers. The films were characterized in terms of surface morphology, residual-resistance ratio, crystal phase composition and superconducting transition temperature, as well as RF-performance using coplanar waveguide resonators. Internal quality factors of up to 1.5 million were measured at 100 mK in the single-photon regime. Despite similar bulk material properties, alpha-Ta films on different seed layers exhibit markedly different RF-performance, which we attribute to dissimilar strain and structural defects at the substrate-metal interfaces. Williamson-Hall analysis of XRD data reveals a clear correlation between decreasing microstrain and increasing quality factor. Cross-sectional HR-TEM further supports this interpretation by directly resolving interfacial disorder. Our results highlight the critical role of interface engineering in optimizing superconducting thin films for low-loss quantum computing circuitry.

cond-mat.mtrl-sci↗

arXiv2026

LUCID: Learning Embodiment-Agnostic Intent Models from Unstructured Human Videos for Scalable Dexterous Robot Skill Acquisition

The most widely-adopted robot learning pipelines today learn skills from robot demonstrations or structured human data, which are expensive to collect and tied to specific embodiments. In contrast, unstructured human videos provide a scalable alternative. They contain diverse manipulation demonstrations across objects, scenes, and strategies, but are not directly connected to robot action. We propose LUCID, a two-stage framework that learns task intent from unstructured human videos drawn from internet-scale datasets and learns robot control in massively-parallel simulation. The intent model predicts short-horizon intent (what should happen next in the scene) from the current observation in closed loop. An embodiment-specific sensorimotor policy converts this intent into robot actions. The intent interface is shared across controllers, so the same intent model can be applied to different embodiments, from our primary dexterous hand to a parallel-jaw gripper. We evaluate LUCID on five real-world manipulation tasks: stirring, wiping, and binning supervised by only internet video, with zero-shot transfer to novel scenes and object instances; and push-T and cable routing supervised by 1 hr each of self-collected smartphone video. Project page: https://lucid-robot.github.io/.

cs.RO↗

arXiv2026

Function-based Parametric Co-Design Optimization of Dexterous Hands

Despite advances in dexterous hand manipulation, robotic hand design is still largely decoupled from task-driven evaluation and control, limiting systematic optimization. Existing robotic hand co-design approaches are often limited in scope, optimizing a small subset of design parameters. We introduce a comprehensive parametric framework for robotic hand generation that unifies palm structure, finger kinematics, fingertip geometry, and fine-scale surface curvatures within a single design space. Fine geometric features are introduced through parametric surface deformation kernels that directly influence contact interactions. We validate the framework on design optimization in grasp stability tasks in simulation and real-world dynamic scenarios. Our framework produces simulation- and fabrication-ready hand models and will be released as open-source to enable rapid design iteration for dexterous hand co-design optimization frameworks and cross-embodiment policy training and control research.

cs.RO↗

arXiv2026

Enhanced Tantalum Superconducting Resonator Performance via All-Surface Organic Monolayer Passivation

Tantalum is a promising platform for superconducting quantum circuits, yet coherence times remain limited by dielectric losses from interfacial two-level systems (TLS), exacerbated by native oxide regrowth. Here, we implement molecular surface passivation using self-assembled organic monolayers on freshly etched tantalum and silicon in coplanar waveguide resonators. Surface characterization by contact angle, XPS, FTIR and TEM confirm the formation of ordered, nanometer-thick films that suppress oxide formation. Microwave measurements in the ~5-9 GHz range reveal internal quality factors up to 1.8x10^6 in the single-photon regime at 100 mK, representing a ~140% improvement over untreated devices with native oxide. Power and temperature dependent measurements attribute this enhancement to reduced TLS-induced losses. These results demonstrate that molecular passivation effectively engineers low-loss interfaces and provides a scalable route toward high-coherence superconducting quantum devices.

cond-mat.mtrl-sci↗

arXiv2026

Grasp to Act: Dexterous Grasping for Tool Use in Dynamic Settings

Achieving robust grasping with dexterous hands remains challenging, especially when manipulation involves dynamic forces such as impacts, torques, and continuous resistance--situations common in real-world tool use. Existing methods largely optimize grasps for static geometric stability and often fail once external forces arise during manipulation. We present Grasp-to-Act, a hybrid system that combines physics-based grasp optimization with reinforcement-learning-based grasp adaptation to maintain stable grasps throughout functional manipulation tasks. Our method synthesizes robust grasp configurations informed by human demonstrations and employs an adaptive controller that residually issues joint corrections to prevent in-hand slip while tracking the object trajectory. Grasp-to-Act enables robust zero-shot sim-to-real transfer across five dynamic tool-use tasks--hammering, sawing, cutting, stirring, and scooping--consistently outperforming baselines. Across simulation and real-world hardware trials with a 16-DoF dexterous hand, our method reduces translational and rotational in-hand slip and achieves the highest task completion rates, demonstrating stable functional grasps under dynamic, contact-rich conditions.

cs.RO↗

arXiv2025

UMI-on-Air: Embodiment-Aware Guidance for Embodiment-Agnostic Visuomotor Policies

We introduce UMI-on-Air, a framework for embodiment-aware deployment of embodiment-agnostic manipulation policies. Our approach leverages diverse, unconstrained human demonstrations collected with a handheld gripper (UMI) to train generalizable visuomotor policies. A central challenge in transferring these policies to constrained robotic embodiments-such as aerial manipulators-is the mismatch in control and robot dynamics, which often leads to out-of-distribution behaviors and poor execution. To address this, we propose Embodiment-Aware Diffusion Policy (EADP), which couples a high-level UMI policy with a low-level embodiment-specific controller at inference time. By integrating gradient feedback from the controller's tracking cost into the diffusion sampling process, our method steers trajectory generation towards dynamically feasible modes tailored to the deployment embodiment. This enables plug-and-play, embodiment-aware trajectory adaptation at test time. We validate our approach on multiple long-horizon and high-precision aerial manipulation tasks, showing improved success rates, efficiency, and robustness under disturbances compared to unguided diffusion baselines. Finally, we demonstrate deployment in previously unseen environments, using UMI demonstrations collected in the wild, highlighting a practical pathway for scaling generalizable manipulation skills across diverse-and even highly constrained-embodiments. All code, data, checkpoints, and result videos can be found at umi-on-air.github.io.

cs.RO↗

arXiv2025

High temporal stability of niobium superconducting resonators by surface passivation with organophosphonate self-assembled monolayers

One main limiting factor towards achieving high coherence times in superconducting circuits is two level system (TLS) losses. Mitigating such losses requires controlling the formation of native oxides at the metal-air interface. Here, we report the growth of alkyl-phosphonate self-assembled monolayers (SAMs) on Nb thin films following oxide removal. The impact of passivation was evaluated via the performance of coplanar waveguide resonators at 10mK, in terms of quality factor and resonant frequency, over six days of air exposure. Un-passivated resonators exhibited an ~80% increase in loss at single-photon power levels, whereas SAM-passivated resonators maintained excellent temporal stability, attributed to suppressed oxide regrowth. By employing a two-component TLS model, we discern distinct prominent loss channels for each resonator type and quantified the characteristic TLS loss of the SAMs to be ~5x10^-7. We anticipate our passivation methodology to offer a promising route toward industrial-scale qubit fabrication, particularly where long-term device stability is critical.

cond-mat.mtrl-sci↗

arXiv2025

Topologically Protected Polaritonic Bound State in the Continuum

Bound states in the continuum (BICs) have emerged as powerful tools to realize ultra-high-Q resonances in nanophotonics. While previous implementations have primarily relied on dielectric metasurfaces, their optical confinement remains fundamentally limited by diffraction. In this work, we theoretically and numerically demonstrate and experimentally validate the existence of topologically protected phonon-polaritonic BICs in periodic arrays of cylindrical nanoresonators composed of isotopically enriched hexagonal boron nitride (h11BN), which support two restrahlen bands (lower (type-I) and upper (type II)), with the present work focusing on the lower Reststrahlen band (RB-1). Owing to the uniaxial anisotropy of hBN and the rotational symmetry of the structure, these systems support topologically symmetry-protected BICs at the {\Gamma}-point, where radiative losses are suppressed. The total quality factor is ultimately bounded by the intrinsic phonon damping of h11BN, enabling high-Q polaritonic modes with minimal radiation leakage. When cylindrical symmetry is broken via angular tilting of incident light away from normal incidence, these BICs transition into quasi-BICs (q-BICs), with strong field confinement and tunable radiation leakage. Their topological features enable robust control over mode lifetimes and confinement, paving the way towards scalable polaritonic platforms for mid-infrared optoelectronics, sensing and quantum nanophotonics.

physics.optics↗

arXiv2025

Sensor-Invariant Tactile Representation

High-resolution tactile sensors have become critical for embodied perception and robotic manipulation. However, a key challenge in the field is the lack of transferability between sensors due to design and manufacturing variations, which result in significant differences in tactile signals. This limitation hinders the ability to transfer models or knowledge learned from one sensor to another. To address this, we introduce a novel method for extracting Sensor-Invariant Tactile Representations (SITR), enabling zero-shot transfer across optical tactile sensors. Our approach utilizes a transformer-based architecture trained on a diverse dataset of simulated sensor designs, allowing it to generalize to new sensors in the real world with minimal calibration. Experimental results demonstrate the method's effectiveness across various tactile sensing applications, facilitating data and model transferability for future advancements in the field.

cs.RO↗

arXiv2025

Fault-tolerant syndrome extraction in [[n,1,3]] non-CSS code family generated using measurements on graph states

The reliability of quantum computation critically depends on the performance of quantum error-correcting codes (QECCs). Performance of QECCs can be severely degraded by hook errors, which effectively reduce the code distance. In this work, we construct a family of $[[n,1,3]]$ non-CSS QECCs, which are fault-tolerant (FT) against noisy syndrome measurements. We employ the bare-ancilla method of Muyuan Li \emph{et al.} to demonstrate fault tolerance against hook errors during syndrome extraction. We present a systematic protocol for generating these QECCs using graph codes and propose a family of $[[n,1,3]]$ codes that preserve the fault-tolerant properties of the bare ancilla codes. We use a custom lookup-table decoder and simulate the code's performance under both anisotropic and circuit-level depolarizing noise. Our results reveal a trade-off in performance with respect to the code rate and identify optimized codes under these noise models. We benchmark our results against the flag-qubit method of Chao \emph{et al}. Notably, we report a new bare ancilla code with improved code rate while maintaining the same distance compared to the bare code used in the work of Muyuan Li \emph{et al.}

quant-ph↗

arXiv2024

Development of TiN/AlN-based superconducting qubit components

This paper presents the fabrication and characterization of superconducting qubit components from titanium nitride (TiN) and aluminum nitride (AlN) layers to create Josephson junctions and superconducting resonators in an all-nitride architecture. Our methodology comprises a complete process flow for the fabrication of TiN/AlN/TiN junctions, characterized by scanning electron microscopy (SEM), atomic force microscopy (AFM), ellipsometry and DC electrical measurements. We evaluated the sputtering rates of AlN under varied conditions, the critical temperatures of TiN thin films for different sputtering environments, and the internal quality factors of TiN resonators in the few-GHz regime, fabricated from these films. Overall, this offered insights into the material properties critical to qubit performance. Measurements of the dependence of the critical current of the TiN / AlN / TiN junctions yielded values ranging from 150 ${\mu}$A to 2 ${\mu}$A, for AlN barrier thicknesses up to ca. 5 nm, respectively. Our findings demonstrate advances in the fabrication of nitride-based superconducting qubit components, which may find applications in quantum computing technologies based on novel materials.

physics.app-ph↗

arXiv2024

Tantalum thin films sputtered on silicon and on different seed layers: material characterization and coplanar waveguide resonator performance

Superconducting qubits are a promising platform for large-scale quantum computing. Besides the Josephson junction, most parts of a superconducting qubit are made of planar, patterned superconducting thin films. In the past, most qubit architectures have relied on niobium (Nb) as the material of choice for the superconducting layer. However, there is also a variety of alternative materials with potentially less losses, which may thereby result in increased qubit performance. One such material is tantalum (Ta), for which high-performance qubit components have already been demonstrated. In this study, we report the sputter-deposition of Ta thin films directly on heated and unheated silicon (Si) substrates as well as onto different, nanometer-thin seed layers from tantalum nitride (TaN), titanium nitride (TiN) or aluminum nitride (AlN) that were deposited first. The thin films are characterized in terms of surface morphology, crystal structure, phase composition, critical temperature, residual resistance ratio (RRR) and RF-performance. We obtain thin films indicative of pure alpha-Ta for high temperature (600{\deg}C) sputtering directly on silicon and for Ta deposited on TaN or TiN seed layers. Coplanar waveguide (CPW) resonator measurements show that the Ta deposited directly on the heated silicon substrate performs best with internal quality factors $Q_i$ reaching 1 x $10^6$ in the single-photon regime, measured at $T=100 {\space \rm mK}$.

cond-mat.mtrl-sci↗

arXiv2024

Text2Place: Affordance-aware Text Guided Human Placement

For a given scene, humans can easily reason for the locations and pose to place objects. Designing a computational model to reason about these affordances poses a significant challenge, mirroring the intuitive reasoning abilities of humans. This work tackles the problem of realistic human insertion in a given background scene termed as \textbf{Semantic Human Placement}. This task is extremely challenging given the diverse backgrounds, scale, and pose of the generated person and, finally, the identity preservation of the person. We divide the problem into the following two stages \textbf{i)} learning \textit{semantic masks} using text guidance for localizing regions in the image to place humans and \textbf{ii)} subject-conditioned inpainting to place a given subject adhering to the scene affordance within the \textit{semantic masks}. For learning semantic masks, we leverage rich object-scene priors learned from the text-to-image generative models and optimize a novel parameterization of the semantic mask, eliminating the need for large-scale training. To the best of our knowledge, we are the first ones to provide an effective solution for realistic human placements in diverse real-world scenes. The proposed method can generate highly realistic scene compositions while preserving the background and subject identity. Further, we present results for several downstream tasks - scene hallucination from a single or multiple generated persons and text-based attribute editing. With extensive comparisons against strong baselines, we show the superiority of our method in realistic human placement.

cs.CV↗

arXiv2023

Bound states in the continuum and long-range coupling of polaritons in hexagonal boron nitride nanoresonators

Bound states in the continuum (BICs) garnered significant for their potential to create new types of nanophotonic devices. Most prior demonstrations were based on arrays of dielectric resonators, which cannot be miniaturized beyond the diffraction limit, reducing the applicability of BICs for advanced functions. Here, we demonstrate BICs and quasi-BICs based on high-quality factor phonon-polariton resonances in isotopically pure h11BN and how these states can be supported by periodic arrays of nanoresonators with sizes much smaller than the wavelength. We theoretically illustrate how BICs emerge from the band structure of the arrays and verify both numerically and experimentally the presence of these states and enhanced quality factor. Furthermore, we identify and characterize simultaneously quasi-BICs and bright states. Our method can be generalized to create a large number of optical states and to tune their coupling with the environment, paving the way to miniaturized nanophotonic devices with more advanced functions.

physics.optics↗

arXiv2022

Cross-Geography Generalization of Machine Learning Methods for Classification of Flooded Regions in Aerial Images

Identification of regions affected by floods is a crucial piece of information required for better planning and management of post-disaster relief and rescue efforts. Traditionally, remote sensing images are analysed to identify the extent of damage caused by flooding. The data acquired from sensors onboard earth observation satellites are analyzed to detect the flooded regions, which can be affected by low spatial and temporal resolution. However, in recent years, the images acquired from Unmanned Aerial Vehicles (UAVs) have also been utilized to assess post-disaster damage. Indeed, a UAV based platform can be rapidly deployed with a customized flight plan and minimum dependence on the ground infrastructure. This work proposes two approaches for identifying flooded regions in UAV aerial images. The first approach utilizes texture-based unsupervised segmentation to detect flooded areas, while the second uses an artificial neural network on the texture features to classify images as flooded and non-flooded. Unlike the existing works where the models are trained and tested on images of the same geographical regions, this work studies the performance of the proposed model in identifying flooded regions across geographical regions. An F1-score of 0.89 is obtained using the proposed segmentation-based approach which is higher than existing classifiers. The robustness of the proposed approach demonstrates that it can be utilized to identify flooded regions of any region with minimum or no user intervention.

cs.CV↗

arXiv2022

Construction of non-CSS quantum codes using measurements on cluster states

The Measurement-based quantum computation provides an alternate model for quantum computation compared to the well-known gate-based model. It uses qubits prepared in a specific entangled state followed by single-qubit measurements. The stabilizers of cluster states are well defined because of their graph structure. We exploit this graph structure extensively to design non-CSS codes using measurement in a specific basis on the cluster state. % We aim to construct $[[n,1]]$ non-CSS code from a $(n+1)$ qubit cluster state. The procedure is general and can be used specifically as an encoding technique to design any non-CSS codes with one logical qubit. We show there exists a $(n+1)$ qubit cluster state which upon measurement gives the desired $[[n,1]]$ code.

quant-ph↗

arXiv2020

Combining Reinforcement Learning with Model Predictive Control for On-Ramp Merging

We consider the problem of designing an algorithm to allow a car to autonomously merge on to a highway from an on-ramp. Two broad classes of techniques have been proposed to solve motion planning problems in autonomous driving: Model Predictive Control (MPC) and Reinforcement Learning (RL). In this paper, we first establish the strengths and weaknesses of state-of-the-art MPC and RL-based techniques through simulations. We show that the performance of the RL agent is worse than that of the MPC solution from the perspective of safety and robustness to out-of-distribution traffic patterns, i.e., traffic patterns which were not seen by the RL agent during training. On the other hand, the performance of the RL agent is better than that of the MPC solution when it comes to efficiency and passenger comfort. We subsequently present an algorithm which blends the model-free RL agent with the MPC solution and show that it provides better trade-offs between all metrics -- passenger comfort, efficiency, crash rate and robustness.

cs.RO↗