arXiv subjects

Qianxu Wang

Publications and source records attributed to Qianxu Wang.

8 recordsLinked to original sources

arXiv2026

Probing the Dynamics of Two-Level System Defect Ensembles via Broadband Cryogenic Transient Dielectric Spectroscopy

Two-level system (TLS) defects in dielectrics are a major source of decoherence in superconducting circuits, yet their microscopic origin and distribution remain poorly understood. Existing circuit-QED probes access limited frequency ranges and mode volumes, restricting studies of isolated materials and interfaces. Here, we present Broadband Cryogenic Transient Dielectric Spectroscopy (BCTDS), a technique for probing TLS-hosting materials over a broad frequency range at cryogenic temperatures. Under strong finite-duration microwave excitation, the transient homodyne I-Q response exhibits coherent phase dynamics after the drive is turned off. Fourier analysis of the transient phase reveals characteristic V-shaped structures that move between cooldowns, consistent with thermocycling-induced changes in the local TLS defect environment that shift defect resonance frequencies. The transient response of BCTDS further enables estimation of susceptibility and two-time correlation functions of the TLS defect ensemble. The observed phase dynamics are qualitatively captured by a driven standard tunneling model containing only a few representative TLS defects. Despite its simplicity relative to the full experimental ensemble, the model reproduces the essential Floquet-dressed dynamics during the drive and generates post-pulse V-shaped structures and interference fringes consistent with the experimental data. The observed BCTDS response may reflect a crossover from localized TLS defect dynamics to a delocalized regime under strong driving, before being quenched into a transient regime that reflects the TLS defect resonance frequencies. Overall, BCTDS represents a potentially useful broadband, time-resolved wafer-level approach for probing TLS defects relevant to quantum technologies.

quant-ph↗

arXiv2026

SmoothAgent: Efficient Long-Horizon LLM-Based Agent Serving with Lookahead Context Engineering

LLM-based agents execute multi-turn workflows with continuously growing contexts, where LLM calls are interleaved with tool invocations and environment feedback. To maintain model quality, modern agent frameworks rely on context engineering strategies such as offloading, reduction, and isolation to control the context length. However, these strategies introduce significant context transformation overhead: each transformation invalidates existing KV caches and triggers re-prefill, leading to increased time-to-first-token (TTFT). In this paper, we identify that context transformations are segment-decomposable, where the transformation of a prefix is independent of future tokens. This property enables transformations to be executed ahead of time. Based on this insight, we propose a lookahead programming model that allows agent frameworks to express context transformations as asynchronous operations without modifying their execution logic. The runtime proactively executes these transformations and prepares transformed KV caches in advance, enabling direct context replacement without blocking. We further design a lookahead-aware scheduler in LLM serving systems to support these asynchronous requests alongside latency-critical workloads with controlled interference. We implement our approach to support representative context engineering strategies and integrate it into existing agent frameworks and LLM serving systems. Experiments show that our approach effectively eliminates transformation overhead and reduces TTFT by up to 11.9x.

cs.DC↗

arXiv2026

KITE: Decoupling Kinematics and Interaction for Zero-Shot Cross-Embodiment Manipulation

Generalizing manipulation policies across robot embodiments remains difficult because standard policies entangle task reasoning with embodiment-specific motor control. We study zero-shot cross-embodiment manipulation, where a policy trained on source embodiments must be deployed on a structurally different target embodiment without additional task demonstrations. We introduce Kinematic Interaction Transfer across Embodiments (KITE), which decouples manipulation into embodiment-agnostic task reasoning and embodiment-specific motor control, connected through a learned latent representation of interaction intent based on contact patterns. Task reasoning is performed by a shared policy that predicts latent intents from source demonstrations, while motor control is performed by an intent-conditioned action decoder learned from each embodiment's kinematic model. With KITE, adaptation to a new embodiment requires only training a new action decoder using its kinematic model, without recollecting demonstration data. We evaluate KITE on three manipulation tasks spanning transfer between parallel grippers, dexterous hands, and composite embodiments. KITE consistently achieves zero-shot transfer to structurally different target embodiments, outperforming state-of-the-art baselines in transfer success and task-embodiment scope.

cs.RO↗

arXiv2025

Spectroscopy and Coherent Control of Two-Level System Defect Ensembles Using a Broadband 3D Waveguide

Defects in solid-state materials play a central role in determining coherence, stability, and performance in quantum technologies. Although narrowband techniques can probe specific resonances with high precision, a broadband spectroscopic approach captures the full spectrum of defect properties and dynamics. Two-level system (TLS) defects in amorphous dielectrics are a particularly important example because they are major sources of decoherence and energy loss in superconducting quantum devices. However, accessing and characterizing their collective dynamics remains far more challenging than probing individual TLS defects. Building on our previously developed Broadband Cryogenic Transient Dielectric Spectroscopy (BCTDS) technique, we study the coherent control and time-resolved dynamics of TLS defect ensembles over a wide frequency range of 3-5 GHz without requiring full device fabrication, revealing quantum interference effects, memory-dependent dynamics, and dressed-state evolution within the TLS defect bath. The spectral response reveals distinct V-shaped structures corresponding to the bare eigenmode frequencies. Using these features, we extract a TLS defect spectral density of 84 GHz^-1 for a silicon sample, across a 4.1-4.6 GHz span. Furthermore, we systematically investigate amplitude- and phase-controlled interference fringes for multiple temperatures and inter-pulse delays, providing direct evidence of coherent dynamics and control. A driven minimal spin model with dipole-dipole interactions that qualitatively capture the observed behavior is presented. Our results establish BCTDS as a versatile platform for broadband defect spectroscopy, offering new capabilities for diagnosing and mitigating sources of decoherence, engineering many-body dynamics, and exploring non-equilibrium phenomena in disordered quantum systems.

quant-ph↗

arXiv2024

Neural Attention Field: Emerging Point Relevance in 3D Scenes for One-Shot Dexterous Grasping

One-shot transfer of dexterous grasps to novel scenes with object and context variations has been a challenging problem. While distilled feature fields from large vision models have enabled semantic correspondences across 3D scenes, their features are point-based and restricted to object surfaces, limiting their capability of modeling complex semantic feature distributions for hand-object interactions. In this work, we propose the \textit{neural attention field} for representing semantic-aware dense feature fields in the 3D space by modeling inter-point relevance instead of individual point features. Core to it is a transformer decoder that computes the cross-attention between any 3D query point with all the scene points, and provides the query point feature with an attention-based aggregation. We further propose a self-supervised framework for training the transformer decoder from only a few 3D pointclouds without hand demonstrations. Post-training, the attention field can be applied to novel scenes for semantics-aware dexterous grasping from one-shot demonstration. Experiments show that our method provides better optimization landscapes by encouraging the end-effector to focus on task-relevant scene regions, resulting in significant improvements in success rates on real robots compared with the feature-field-based methods.

cs.RO↗

arXiv2024

SparseDFF: Sparse-View Feature Distillation for One-Shot Dexterous Manipulation

Humans demonstrate remarkable skill in transferring manipulation abilities across objects of varying shapes, poses, and appearances, a capability rooted in their understanding of semantic correspondences between different instances. To equip robots with a similar high-level comprehension, we present SparseDFF, a novel DFF for 3D scenes utilizing large 2D vision models to extract semantic features from sparse RGBD images, a domain where research is limited despite its relevance to many tasks with fixed-camera setups. SparseDFF generates view-consistent 3D DFFs, enabling efficient one-shot learning of dexterous manipulations by mapping image features to a 3D point cloud. Central to SparseDFF is a feature refinement network, optimized with a contrastive loss between views and a point-pruning mechanism for feature continuity. This facilitates the minimization of feature discrepancies w.r.t. end-effector parameters, bridging demonstrations and target manipulations. Validated in real-world scenarios with a dexterous hand, SparseDFF proves effective in manipulating both rigid and deformable objects, demonstrating significant generalization capabilities across object and scene variations.

cs.RO↗

arXiv2023

ImageManip: Image-based Robotic Manipulation with Affordance-guided Next View Selection

In the realm of future home-assistant robots, 3D articulated object manipulation is essential for enabling robots to interact with their environment. Many existing studies make use of 3D point clouds as the primary input for manipulation policies. However, this approach encounters challenges due to data sparsity and the significant cost associated with acquiring point cloud data, which can limit its practicality. In contrast, RGB images offer high-resolution observations using cost effective devices but lack spatial 3D geometric information. To overcome these limitations, we present a novel image-based robotic manipulation framework. This framework is designed to capture multiple perspectives of the target object and infer depth information to complement its geometry. Initially, the system employs an eye-on-hand RGB camera to capture an overall view of the target object. It predicts the initial depth map and a coarse affordance map. The affordance map indicates actionable areas on the object and serves as a constraint for selecting subsequent viewpoints. Based on the global visual prior, we adaptively identify the optimal next viewpoint for a detailed observation of the potential manipulation success area. We leverage geometric consistency to fuse the views, resulting in a refined depth map and a more precise affordance map for robot manipulation decisions. By comparing with prior works that adopt point clouds or RGB images as inputs, we demonstrate the effectiveness and practicality of our method. In the project webpage (https://sites.google.com/view/imagemanip), real world experiments further highlight the potential of our method for practical deployment.

cs.RO↗

arXiv2023

Learning a Universal Human Prior for Dexterous Manipulation from Human Preference

Generating human-like behavior on robots is a great challenge especially in dexterous manipulation tasks with robotic hands. Scripting policies from scratch is intractable due to the high-dimensional control space, and training policies with reinforcement learning (RL) and manual reward engineering can also be hard and lead to unnatural motions. Leveraging the recent progress on RL from Human Feedback, we propose a framework that learns a universal human prior using direct human preference feedback over videos, for efficiently tuning the RL policies on 20 dual-hand robot manipulation tasks in simulation, without a single human demonstration. A task-agnostic reward model is trained through iteratively generating diverse polices and collecting human preference over the trajectories; it is then applied for regularizing the behavior of polices in the fine-tuning stage. Our method empirically demonstrates more human-like behaviors on robot hands in diverse tasks including even unseen tasks, indicating its generalization capability.

cs.RO↗