arXiv subjects

Sean D. Huver

Publications and source records attributed to Sean D. Huver.

10 recordsLinked to original sources

arXiv2026

NVIDIA Cosmos-H-Dreams: Real-Time Generative Physics Simulation for Surgical Robotics

Generative simulation for surgical robotics still lacks real-time interaction. Physical-robot experiments, often involving animal or cadaver labs, are time-consuming, costly, and difficult to reproduce, while classical simulators struggle to capture photorealistic appearance and deformable-tissue dynamics. We address this gap with Cosmos-H-Dreams, an integrated real-time surgical world-model system combining an action-conditioned generative model, a teacher-to-student distillation recipe, and a deployment stack built on the NVIDIA FlashDreams streaming-inference library. Starting from Cosmos-H-Surgical-Simulator, a multi-embodiment action-conditioned surgical video world model fine-tuned on the large-scale Open-H-Embodiment corpus, we post-train this checkpoint on embodiment- and procedure-specific data. By distilling the resulting bidirectional teacher into a causal, few-step student with Self Forcing, we turn a passive video generator into a controllable surgical simulator that streams at $\sim$160 inference FPS on a single NVIDIA RTX PRO 6000 Blackwell workstation GPU. Crucially, Cosmos-H-Dreams is controller-agnostic: any interface that emits a stream of robot kinematics can drive it. We demonstrate live control through a browser keyboard over WebRTC, a Meta Quest headset over WebXR, a commercial surgical robot console such as CMR Surgical's Versius, and learned policies operating in closed loop. To our knowledge, this is the first interactive surgical world model supporting live human and policy control. Human operators and policies alike can act inside the synthesized world and observe the consequences in real time. We release Cosmos-H-Dreams as an open surgical simulation system, providing a common foundation for surgical education, scalable synthetic data generation, and future intraoperative decision support.

cs.RO↗

arXiv2026

SurgVIL: Scaling Surgical Robot Imitation Learning with Open-source Surgical Videos

Learning-based surgical robot autonomy requires large-scale demonstrations with synchronized videos and robot actions, but such data are exceedingly rare in clinical or realistic tissue settings because robot kinematics are typically inaccessible outside controlled research systems. In contrast, phantom data collected on research platforms provide accurate action labels but lack the visual diversity of real tissue. We propose SurgVIL, a framework for scaling surgical robot imitation learning using open-source surgical videos. SurgVIL combines kinematically labeled phantom robot demonstrations with surgical videos from open-source datasets and online sources for policy learning. Since these videos lack robot motion labels, we estimate approximate kinematics as weak supervision. We evaluate SurgVIL on two da Vinci robot tasks: needle pick-up and cholecystectomy cutting. Across ACT, $π_0$, and GR00T-H backbones, adding surgical videos substantially improves generalization to real-tissue and out-of-distribution settings, suggesting a scalable path from phantom training toward generalizable surgical robot policies.

cs.RO↗

arXiv2026

A Ramsey Ion Gradiometer for Single-Molecule State Detection

The characterization of ligand--receptor interactions is a cornerstone of modern pharmacology; however, current methods are hampered by limitations such as ensemble averaging and invasive labeling. We propose a theoretical quantum sensing solution, the Quantum Ligand-Binding Interrogator (QLI), designed to overcome these challenges. The QLI is a differential sensor, or gradiometer, that uses a pair of co-trapped atomic ions to perform label-free detection of the electric field gradient produced by a single ligand binding to its receptor in vitrified samples. This gradiometric approach provides robust common-mode rejection of background electric field noise. To bridge the gap between the cryogenic, ultra-high-vacuum environment required for the sensor and the biological sample, we propose an architecture based on a vitrified sample mounted on a scanning probe. This enables the detection of the electrostatic signature of a single molecule in a specific conformational state (e.g., bound vs.\ unbound). This paper details the conceptual framework of the QLI, the experimental architecture, the measurement protocol using entangled two-ion spin states, and an analysis of key engineering risks. Anchoring to state-of-the-art single-ion low-frequency sensitivities (sub-mV\,m$^{-1}$/\,$\sqrt{\mathrm{Hz}}$), we project SNR\,=\,1 in tens of seconds at a 10\,\textmu m ion--sample separation for $Δp \sim 20$\,D, with feasibility dominated by the (as yet unmeasured) electrostatic stability of vitrified samples. If realized, QLI would provide direct single-molecule measurements of binding-induced electric field changes, offering a new path for experimental validation of computational models of drug--receptor interactions.

quant-ph↗

arXiv2026

G-SHARP: Gaussian Surgical Hardware Accelerated Real-time Pipeline

We propose G-SHARP, a commercially compatible, real-time surgical scene reconstruction framework designed for minimally invasive procedures that require fast and accurate 3D modeling of deformable tissue. While recent Gaussian splatting approaches have advanced real-time endoscopic reconstruction, existing implementations often depend on non-commercial derivatives, limiting deployability. G-SHARP overcomes these constraints by being the first surgical pipeline built natively on the GSplat (Apache-2.0) differentiable Gaussian rasterizer, enabling principled deformation modeling, robust occlusion handling, and high-fidelity reconstructions on the EndoNeRF pulling benchmark. Our results demonstrate state-of-the-art reconstruction quality with strong speed-accuracy trade-offs suitable for intra-operative use. Finally, we provide a Holoscan SDK application that deploys G-SHARP on NVIDIA IGX Orin and Thor edge hardware, enabling real-time surgical visualization in practical operating-room settings.

cs.CV↗

arXiv2018

On the use of deep neural networks in optical communications

Information transfer rates in optical communications may be dramatically increased by making use of spatially non-Gaussian states of light. Here we demonstrate the ability of deep neural networks to classify numerically-generated, noisy Laguerre-Gauss modes of up to 100 quanta of orbital angular momentum with near-unity fidelity. The scheme relies only on the intensity profile of the detected modes, allowing for considerable simplification of current measurement schemes required to sort the states containing increasing degrees of orbital angular momentum. We also present results that show the strength of deep neural networks in the classification of experimental superpositions of Laguerre-Gauss modes when the networks are trained solely using simulated images. It is anticipated that these results will allow for an enhancement of current optical communications technologies.

physics.app-ph↗

arXiv2009

Quantum Metrology with Two-Mode Squeezed Vacuum: Parity Detection Beats the Heisenberg Limit

We study the sensitivity and resolution of phase measurement in a Mach-Zehnder interferometer with two-mode squeezed vacuum ( photons on average). We show that super-resolution and sub-Heisenberg sensitivity is obtained with parity detection. In particular, in our setup, dependence of the signal on the phase evolves times faster than in traditional schemes, and uncertainty in the phase estimation is better than 1/ .

quant-ph↗

arXiv2009

Optimization of quantum interferometric metrological sensors in the presence of photon loss

We optimize two-mode, entangled, number states of light in the presence of loss in order to maximize the extraction of the available phase information in an interferometer. Our approach optimizes over the entire available input Hilbert space with no constraints, other than fixed total initial photon number. We optimize to maximize the Fisher information, which is equivalent to minimizing the phase uncertainty. We find that in the limit of zero loss the optimal state is the so-called N00N state, for small loss, the optimal state gradually deviates from the N00N state, and in the limit of large loss the optimal state converges to a generalized two-mode coherent state, with a finite total number of photons. The results provide a general protocol for optimizing the performance of a quantum optical interferometer in the presence of photon loss, with applications to quantum imaging, metrology, sensing, and information processing.

quant-ph↗

arXiv2008

Maximal Success Probabilities of Linear-Optical Quantum Gates

Numerical optimization is used to design linear-optical devices that implement a desired quantum gate with perfect fidelity, while maximizing the success rate. For the 2-qubit CS (or CNOT) gate, we provide numerical evidence that the maximum success rate is $S=2/27$ using two unentangled ancilla resources; interestingly, additional ancilla resources do not increase the success rate. For the 3-qubit Toffoli gate, we show that perfect fidelity is obtained with only three unentangled ancilla photons -- less than in any existing scheme -- with a maximum $S=0.00340$. This compares well with $S=(2/27)^2/2 \approx 0.00274$, obtainable by combining two CNOT gates and a passive quantum filter [PRA 68, 064303 (2003)]. The general optimization approach can easily be applied to other areas of interest, such as quantum error correction, cryptography, and metrology [arXiv:0807.4906, PRL 99 070801 (2007)].

quant-ph↗

arXiv2008

Entangled Fock states for Robust Quantum Optical Metrology, Imaging, and Sensing

We propose a class of path-entangled photon Fock states for robust quantum optical metrology, imaging, and sensing in the presence of loss. We model propagation loss with beam-splitters and derive a reduced density matrix formalism from which we examine how photon loss affects coherence. It is shown that particular entangled number states, which contain a special superposition of photons in both arms of a Mach-Zehnder interferometer, are resilient to environmental decoherence. We demonstrate an order of magnitude greater visibility with loss, than possible with N00N states. We also show that the effectiveness of a detection scheme is related to super-resolution visibility.

quant-ph↗

arXiv2007

Optimizing Optical Quantum Logic Gates using Genetic Algorithms

We introduce the method of using an annealing genetic algorithm to the numerically complex problem of looking for quantum logic gates which simultaneously have highest fidelity and highest success probability. We first use the linear optical quantum nonlinear sign (NS) gate as an example to illustrate the efficiency of this method. We show that by appropriately choosing the annealing parameters, we can reach the theoretical maximum success probability (1/4 for NS) for each attempt. We then examine the controlled-z (CZ) gate as the first new problem to be solved. We show results that agree with the highest known maximum success probability for a CZ gate (2/27) while maintaining a fidelity of 0.9997. Since the purpose of our algorithm is to optimize a unitary matrix for quantum transformations, it could easily be applied to other areas of interest such as quantum optics and quantum sensors.

quant-ph↗