arXiv subjects

Kai Xu

Publications and source records attributed to Kai Xu.

At least 109 records · Page 6Linked to original sources

arXiv2025

Deliberate Planning of 3D Bin Packing on Packing Configuration Trees

Online 3D Bin Packing Problem (3D-BPP) has widespread applications in industrial automation. Existing methods usually solve the problem with limited resolution of spatial discretization, and/or cannot deal with complex practical constraints well. We propose to enhance the practical applicability of online 3D-BPP via learning on a novel hierarchical representation, packing configuration tree (PCT). PCT is a full-fledged description of the state and action space of bin packing which can support packing policy learning based on deep reinforcement learning (DRL). The size of the packing action space is proportional to the number of leaf nodes, making the DRL model easy to train and well-performing even with continuous solution space. We further discover the potential of PCT as tree-based planners in deliberately solving packing problems of industrial significance, including large-scale packing and different variations of BPP setting. A recursive packing method is proposed to decompose large-scale packing into smaller sub-trees while a spatial ensemble mechanism integrates local solutions into global. For different BPP variations with additional decision variables, such as lookahead, buffering, and offline packing, we propose a unified planning framework enabling out-of-the-box problem solving. Extensive evaluations demonstrate that our method outperforms existing online BPP baselines and is versatile in incorporating various practical constraints. The planning process excels across large-scale problems and diverse problem variations. We develop a real-world packing robot for industrial warehousing, with careful designs accounting for constrained placement and transportation stability. Our packing robot operates reliably and efficiently on unprotected pallets at 10 seconds per box. It achieves averagely 19 boxes per pallet with 57.4% space utilization for relatively large-size boxes.

cs.RO↗

arXiv2025

Demonstration of Maxwell Demon-assistant Einstein-Podolsky-Rosen Steering via Superconducting Quantum Processor

The concept of Maxwell demon plays an essential role in connecting thermodynamics and information theory, while entanglement and non-locality are fundamental features of quantum theory. Given the rapid advancements in the field of quantum information science, there is a growing interest and significance in investigating the connection between Maxwell demon and quantum correlation. The majority of research endeavors thus far have been directed towards the extraction of work from quantum correlation through the utilization of Maxwell demon. Recently, a novel concept called Maxwell demon-assistant Einstein-Podolsky-Rosen (EPR) steering has been proposed, which suggests that it is possible to simulate quantum correlation by doing work. This seemingly counterintuitive conclusion is attributed to the fact that Alice and Bob need classical communication during EPR steering task, a requirement that does not apply in the Bell test. In this study, we demonstrate Maxwell demon-assistant EPR steering with superconducting quantum circuits. By compiling and optimizing a quantum circuit to be implemented on a 2D superconducting chip, we were able to achieve a steering parameter of $S_{2} = 0.770 \pm 0.005$ in the case of two measurement settings, which surpasses the classical bound of $1/\sqrt{2}$ by 12.6 standard deviations. In addition, experimental observations have revealed a linear correlation between the non-locality demonstrated in EPR steering and the work done by the demon. Considering the errors in practical operation, the experimental results are highly consistent with theoretical predictions. Our findings not only suggest the presence of a Maxwell demon loophole in the EPR steering, but also contribute to a deeper comprehension of the interplay between quantum correlation, information theory, and thermodynamics.

quant-ph↗

arXiv2025

CogNav: Cognitive Process Modeling for Object Goal Navigation with LLMs

Object goal navigation (ObjectNav) is a fundamental task in embodied AI, requiring an agent to locate a target object in previously unseen environments. This task is particularly challenging because it requires both perceptual and cognitive processes, including object recognition and decision-making. While substantial advancements in perception have been driven by the rapid development of visual foundation models, progress on the cognitive aspect remains constrained, primarily limited to either implicit learning through simulator rollouts or explicit reliance on predefined heuristic rules. Inspired by neuroscientific findings demonstrating that humans maintain and dynamically update fine-grained cognitive states during object search tasks in novel environments, we propose CogNav, a framework designed to mimic this cognitive process using large language models. Specifically, we model the cognitive process using a finite state machine comprising fine-grained cognitive states, ranging from exploration to identification. Transitions between states are determined by a large language model based on a dynamically constructed heterogeneous cognitive map, which contains spatial and semantic information about the scene being explored. Extensive evaluations on the HM3D, MP3D, and RoboTHOR benchmarks demonstrate that our cognitive process modeling significantly improves the success rate of ObjectNav at least by relative 14% over the state-of-the-arts.

cs.CV↗

arXiv2025

Observation of Inelastic Meson Scattering in a Floquet System using a Digital Quantum Simulator

Lattice gauge theories provide a non-perturbative framework for understanding confinement and hadronic physics, but their real-time dynamics remain challenging for classical computations. However, quantum simulators offer a promising alternative for exploring such dynamics beyond classical capabilities. Here, we experimentally investigate meson scattering using a superconducting quantum processor. Employing a digital protocol, we realize a Floquet spin chain equivalent to a one-dimensional Floquet $\mathbb{Z}_2$ lattice gauge theory. We observe Bloch oscillations of single kinks and strong binding between adjacent kinks, signaling confinement and the formation of stable mesons in this Floquet system. Using full-system joint readout, we resolve meson populations by string length, enabling identification of meson scattering channels. Our results reveal the fragmentation of a long-string meson into multiple short-string mesons, which is also an experimental signature of string breaking. Moreover, we directly observe inelastic meson scattering, where two short-string mesons can merge into a longer one. Our results pave the way for studying interacting gauge particles and composite excitations on digital quantum simulators.

quant-ph↗

arXiv2025

Room-temperature anisotropic in-plane spin dynamics in graphene induced by PdSe$_2$ proximity

Van der Waals heterostructures offer a versatile platform for tailoring electrical, magnetic, optical, and spin transport properties of materials through proximity effects. Notably, hexagonal transition metal dichalcogenides have been shown to induce valley-Zeeman spin-orbit coupling (SOC) in graphene, resulting in significant spin lifetime anisotropy between in-plane and out-of-plane spin orientations. However, in-plane lifetimes remain isotropic due to the inherent threefold symmetry of the heterostructure. Here, we demonstrate that pentagonal PdSe$_2$, characterised by unique in-plane anisotropy, induces an unprecedented gate-tunable SOC in graphene. Our measurements reveal a remarkable 10-fold modulation of the spin lifetime for spins oriented within the graphene plane at room temperature. Moreover, the directional dependence of the spin lifetimes, along the three spatial directions, suggests the existence of a persistent in-plane spin texture component that dominates the spin dynamics. These findings deepen our understanding of spin dynamics in van der Waals heterostructures and open avenues for designing and engineering novel topological phases in graphene-based heterostructures within the strong SOC regime.

cond-mat.mes-hall↗

arXiv2025

BoxFusion: Reconstruction-Free Open-Vocabulary 3D Object Detection via Real-Time Multi-View Box Fusion

Open-vocabulary 3D object detection has gained significant interest due to its critical applications in autonomous driving and embodied AI. Existing detection methods, whether offline or online, typically rely on dense point cloud reconstruction, which imposes substantial computational overhead and memory constraints, hindering real-time deployment in downstream tasks. To address this, we propose a novel reconstruction-free online framework tailored for memory-efficient and real-time 3D detection. Specifically, given streaming posed RGB-D video input, we leverage Cubify Anything as a pre-trained visual foundation model (VFM) for single-view 3D object detection by bounding boxes, coupled with CLIP to capture open-vocabulary semantics of detected objects. To fuse all detected bounding boxes across different views into a unified one, we employ an association module for correspondences of multi-views and an optimization module to fuse the 3D bounding boxes of the same instance predicted in multi-views. The association module utilizes 3D Non-Maximum Suppression (NMS) and a box correspondence matching module, while the optimization module uses an IoU-guided efficient random optimization technique based on particle filtering to enforce multi-view consistency of the 3D bounding boxes while minimizing computational complexity. Extensive experiments on ScanNetV2 and CA-1M datasets demonstrate that our method achieves state-of-the-art performance among online methods. Benefiting from this novel reconstruction-free paradigm for 3D object detection, our method exhibits great generalization abilities in various scenarios, enabling real-time perception even in environments exceeding 1000 square meters.

cs.CV↗

arXiv2025

Non-Equilibrium Criticality-Enhanced Quantum Sensing with Superconducting Qubits

Exploiting quantum features allows for estimating external parameters with precisions well beyond the capacity of classical sensors, a phenomenon known as quantum-enhanced precision. Quantum criticality has been identified as a resource for achieving such enhancements with respect to the probe size. However, they demand complex probe preparation and measurement and the achievable enhancement is ultimately restricted to narrow parameter regimes. On the other hand, non-equilibrium probes harness dynamics, enabling quantum-enhanced precision with respect to time over a wide range of parameters through simple probe initialization. Here, we unify these approaches through a Stark-Wannier localization platform, where competition between a linear gradient field and particle tunneling enables quantum-enhanced sensitivity across an extended parameter regime. The probe is implemented on a 9-qubit superconducting quantum device, in both single- and double-excitation subspaces, where we explore its performance in the extended phase, the critical point and the localized phase. Despite employing only computational-basis measurements we have been able to achieve near-Heisenberg-limited precision by combining outcomes at distinct evolution times. In addition, we demonstrate that the performance of the probe in the entire extended phase is significantly outperforming the performance in the localized regime. Our results highlight Stark-Wannier systems as versatile platforms for quantum sensing, where the combination of criticality and non-equilibrium dynamics enhances precision over a wide range of parameters without stringent measurement requirements.

quant-ph↗

arXiv2025

Rollout Roulette: A Probabilistic Inference Approach to Inference-Time Scaling of LLMs using Particle-Based Monte Carlo Methods

Large language models (LLMs) have achieved significant performance gains via scaling up model sizes and/or data. However, recent evidence suggests diminishing returns from such approaches, motivating scaling the computation spent at inference time. Existing inference-time scaling methods, usually with reward models, cast the task as a search problem, which tends to be vulnerable to reward hacking as a consequence of approximation errors in reward models. In this paper, we instead cast inference-time scaling as a probabilistic inference task and leverage sampling-based techniques to explore the typical set of the state distribution of a state-space model with an approximate likelihood, rather than optimize for its mode directly. We propose a novel inference-time scaling approach by adapting particle-based Monte Carlo methods to this task. Our empirical evaluation demonstrates that our methods have a 4-16x better scaling rate over our deterministic search counterparts on various challenging mathematical reasoning tasks. Using our approach, we show that Qwen2.5-Math-1.5B-Instruct can surpass GPT-4o accuracy in only 4 rollouts, while Qwen2.5-Math-7B-Instruct scales to o1 level accuracy in only 32 rollouts. Our work not only presents an effective method to inference-time scaling, but also connects the rich literature in probabilistic inference with inference-time scaling of LLMs to develop more robust algorithms in future work. Code, videos, and further information available at https://probabilistic-inference-scaling.github.io.

cs.LG↗

arXiv2025

Microwave Engineering of Tunable Spin Interactions with Superconducting Qubits

Quantum simulation has emerged as a powerful framework for investigating complex many - body phenomena. A key requirement for emulating these dynamics is the realization of fully controllable quantum systems enabling various spin interactions. Yet, quantum simulators remain constrained in the types of attainable interactions. Here we demonstrate experimental realization of multiple microwave - engineered spin interactions in superconducting quantum circuits. By precisely controlling the native XY interaction and microwave drives, we achieve tunable spin Hamiltonians including: (i) XYZ spin models with continuously adjustable parameters, (ii) transverse - field Ising systems, and (iii) Dzyaloshinskii - Moriya interacting systems. Our work expands the toolbox for analogue - digital quantum simulation, enabling exploration of a wide range of exotic quantum spin models.

quant-ph↗

arXiv2025

Tomography-assisted noisy quantum circuit simulator using matrix product density operators

In recent years, efficient quantum circuit simulations incorporating ideal noise assumptions have relied on tensor network simulators, particularly leveraging the matrix product density operator (MPDO) framework. However, experiments on real noisy intermediate-scale quantum (NISQ) devices often involve complex noise profiles, encompassing uncontrollable elements and instrument-specific effects such as crosstalk. To address these challenges, we employ quantum process tomography (QPT) techniques to directly capture the operational characteristics of the experimental setup and integrate them into numerical simulations using MPDOs. Our QPT-assisted MPDO simulator is then applied to explore a variational approach for generating noisy entangled states, comparing the results with standard noise numerical simulations and demonstrations conducted on the Quafu cloud quantum computation platform. Additionally, we investigate noisy MaxCut problems, as well as the effects of crosstalk and noise truncation. Our results provide valuable insights into the impact of noise on NISQ devices and lay the foundation for enhanced design and assessment of quantum algorithms in complex noise environments.

quant-ph↗

arXiv2025

Direct Implementation of High-Fidelity Three-Qubit Gates for Superconducting Processor with Tunable Couplers

Three-qubit gates can be constructed using combinations of single-qubit and two-qubit gates, making their independent realization unnecessary. However, direct implementation of three-qubit gates reduces the depth of quantum circuits, streamlines quantum programming, and facilitates efficient circuit optimization, thereby enhancing overall performance in quantum computation. In this work, we propose and experimentally demonstrate a high-fidelity scheme for implementing a three-qubit controlled-controlled-Z (CCZ) gate in a flip-chip superconducting quantum processor with tunable couplers. This direct CCZ gate is implemented by simultaneously leveraging two tunable couplers interspersed between three qubits to enable three-qubit interactions, achieving an average final state fidelity of $97.94\%$ and a process fidelity of $93.54\%$. This high fidelity cannot be achieved through a simple combination of single- and two-qubit gate sequences from processors with similar performance levels. Our experiments also verify that multilayer direct implementation of the CCZ gate exhibits lower leakage compared to decomposed gate approaches. As a showcase, we utilize the CCZ gate as an oracle to implement the Grover search algorithm on three qubits, demonstrating high performance with the target probability amplitude significantly enhanced after two iterations. These results highlight the advantage of our approach, and facilitate the implementation of complex quantum circuits.

quant-ph↗

arXiv2025

SQuat: Subspace-orthogonal KV Cache Quantization

The key-value (KV) cache accelerates LLMs decoding by storing KV tensors from previously generated tokens. It reduces redundant computation at the cost of increased memory usage. To mitigate this overhead, existing approaches compress KV tensors into lower-bit representations; however, quantization errors can accumulate as more tokens are generated, potentially resulting in undesired outputs. In this paper, we introduce SQuat (Subspace-orthogonal KV cache quantization). It first constructs a subspace spanned by query tensors to capture the most critical task-related information. During key tensor quantization, it enforces that the difference between the (de)quantized and original keys remains orthogonal to this subspace, minimizing the impact of quantization errors on the attention mechanism's outputs. SQuat requires no model fine-tuning, no additional calibration dataset for offline learning, and is grounded in a theoretical framework we develop. Through numerical experiments, we show that our method reduces peak memory by 2.17 to 2.82, improves throughput by 2.45 to 3.60, and achieves more favorable benchmark scores than existing KV cache quantization algorithms.

cs.LG↗

arXiv2025

Beyond the Broadcast: Enhancing VR Tennis Broadcasting through Embedded Visualizations and Camera Techniques

Virtual Reality (VR) broadcasting has emerged as a promising medium for providing immersive viewing experiences of major sports events such as tennis. However, current VR broadcast systems often lack an effective camera language and do not adequately incorporate dynamic, in-game visualizations, limiting viewer engagement and narrative clarity. To address these limitations, we analyze 400 out-of-play segments from eight major tennis broadcasts to develop a tennis-specific design framework that effectively combines cinematic camera movements with embedded visualizations. We further refine our framework by examining 25 cinematic VR animations, comparing their camera techniques with traditional tennis broadcasts to identify key differences and inform adaptations for VR. Based on data extracted from the broadcast videos, we reconstruct a simulated game that captures the players' and ball's motion and trajectories. Leveraging this design framework and processing pipeline, we develope Beyond the Broadcast, a VR tennis viewing system that integrates embedded visualizations with adaptive camera motions to construct a comprehensive and engaging narrative. Our system dynamically overlays tactical information and key match events onto the simulated environment, enhancing viewer comprehension and narrative engagement while ensuring perceptual immersion and viewing comfort. A user study involving tennis viewers demonstrate that our approach outperforms traditional VR broadcasting methods in delivering an immersive, informative viewing experience.

cs.HC↗

arXiv2025

Curve-Aware Gaussian Splatting for 3D Parametric Curve Reconstruction

This paper presents an end-to-end framework for reconstructing 3D parametric curves directly from multi-view edge maps. Contrasting with existing two-stage methods that follow a sequential ``edge point cloud reconstruction and parametric curve fitting'' pipeline, our one-stage approach optimizes 3D parametric curves directly from 2D edge maps, eliminating error accumulation caused by the inherent optimization gap between disconnected stages. However, parametric curves inherently lack suitability for rendering-based multi-view optimization, necessitating a complementary representation that preserves their geometric properties while enabling differentiable rendering. We propose a novel bi-directional coupling mechanism between parametric curves and edge-oriented Gaussian components. This tight correspondence formulates a curve-aware Gaussian representation, \textbf{CurveGaussian}, that enables differentiable rendering of 3D curves, allowing direct optimization guided by multi-view evidence. Furthermore, we introduce a dynamically adaptive topology optimization framework during training to refine curve structures through linearization, merging, splitting, and pruning operations. Comprehensive evaluations on the ABC dataset and real-world benchmarks demonstrate our one-stage method's superiority over two-stage alternatives, particularly in producing cleaner and more robust reconstructions. Additionally, by directly optimizing parametric curves, our method significantly reduces the parameter count during training, achieving both higher efficiency and superior performance compared to existing approaches.

cs.CV↗

arXiv2025

Many-body delocalization with a two-dimensional 70-qubit superconducting quantum simulator

Quantum many-body systems with sufficiently strong disorder can exhibit a non-equilibrium phenomenon, known as the many-body localization (MBL), which is distinct from conventional thermalization. While the MBL regime has been extensively studied in one dimension, its existence in higher dimensions remains elusive, challenged by the avalanche instability. Here, using a 70-qubit two-dimensional (2D) superconducting quantum simulator, we experimentally explore the robustness of the MBL regime in controlled finite-size 2D systems. We observe that the decay of imbalance becomes more pronounced with increasing system sizes, scaling up from 21, 42 to 70 qubits, with a relatively large disorder strength, and for the first time, provide an evidence for the many-body delocalization in 2D disordered systems. Our experimental results are consistent with the avalanche theory that predicts the instability of MBL regime beyond one spatial dimension. This work establishes a scalable platform for probing high-dimensional non-equilibrium phases of matter and their finite-size effects using superconducting quantum circuits.

quant-ph↗

arXiv2025

Self-supervised Learning of Hybrid Part-aware 3D Representations of 2D Gaussians and Superquadrics

Low-level 3D representations, such as point clouds, meshes, NeRFs and 3D Gaussians, are commonly used for modeling 3D objects and scenes. However, cognitive studies indicate that human perception operates at higher levels and interprets 3D environments by decomposing them into meaningful structural parts, rather than low-level elements like points or voxels. Structured geometric decomposition enhances scene interpretability and facilitates downstream tasks requiring component-level manipulation. In this work, we introduce PartGS, a self-supervised part-aware reconstruction framework that integrates 2D Gaussians and superquadrics to parse objects and scenes into an interpretable decomposition, leveraging multi-view image inputs to uncover 3D structural information. Our method jointly optimizes superquadric meshes and Gaussians by coupling their parameters within a hybrid representation. On one hand, superquadrics enable the representation of a wide range of shape primitives, facilitating flexible and meaningful decompositions. On the other hand, 2D Gaussians capture detailed texture and geometric details, ensuring high-fidelity appearance and geometry reconstruction. Operating in a self-supervised manner, our approach demonstrates superior performance compared to state-of-the-art methods across extensive experiments on the DTU, ShapeNet, and real-world datasets.

cs.CV↗

arXiv2025

Flexible Readout and Unconditional Reset for Superconducting Multi-Qubit Processors with Tunable Purcell Filters

Achieving high-fidelity qubit readout and reset while preserving qubit coherence is essential for quantum error correction and other advanced quantum algorithms. Here, we design and experimentally demonstrate a scalable architecture employing frequency-tunable nonlinear Purcell filters, enabling flexible readout and fast unconditional reset of multiple superconducting qubits. Our readout protocol dynamically adjusts the effective linewidth of the readout resonator through a tunable Purcell filter, optimizing the signal-to-noise ratio during measurement while suppressing photon noise during idle periods. We achieve a readout fidelity of $99.3\%$ without any quantum-limited amplifier, even with a small dispersive shift. Moreover, by leveraging a reset channel formed via the adjacent coupling between the filter and the coupler, we realize unconditional qubit reset of both leakage-induced $|2\rangle$ and $|1\rangle$ states within 200 ns and reset of the $|1\rangle$ state alone within 75 ns, with error rates $\leq 1\%$. The filter also mitigates both photon-induced dephasing and the Purcell effect, thereby preserving qubit coherence. This scalable Purcell filter architecture shows exceptional performance in qubit readout, reset, and protection, marking it as a promising hardware component for advancing fault-tolerant quantum computing systems.

quant-ph↗

arXiv2025

Noisy Probabilistic Error Cancellation and Generalized Physical Implementability

Decoherence severely limits the performance of quantum processors, posing challenges to reliable quantum computation. Probabilistic error cancellation, a quantum error mitigation method, counteracts noise by quasiprobabilistically simulating (non-physical) inverse noise operations. However, existing formulations of physical implementability, quantifying the minimal cost of simulating non-physical operations using physical channels, do not fully account for the experimental constraints, since noise also affects the cancellation process and not all physical channels are experimentally accessible. Here, we generalize the physical implementability to encompass arbitrary convex sets of experimentally available quantum states and operations. Within this generalized framework, we demonstrate noiseless error cancellation with noisy Pauli operations and analyze the bias of noisy cancellation. Furthermore, we establish connections between generalized physical implementability and quantum information measures, e.g., diamond norm, logarithmic negativity, and purity. These findings enhance the practical applicability of probabilistic error cancellation and open new avenues for robust quantum information processing and quantum computing.

quant-ph↗