arXiv subjects

Peng Xue

Publications and source records attributed to Peng Xue.

At least 19 recordsLinked to original sources

arXiv2026

Ask on Demand: Targeted Peer Communication for Parallel Agentic Systems

Parallel LLM agents tackle complex tasks by exploring multiple trajectories, yet useful discoveries remain isolated in individual execution histories. Existing sharing methods often select or summarize information before other agents' needs become clear, potentially overlooking details that become important later. We propose Ask on Demand (AOD), a communication framework that enables agents to seek targeted peer assistance when specific information needs arise. Upon an approved request, AOD retrieves relevant evidence from peer execution histories, selects a suitable peer, and generates a response tailored to the request. To determine when such assistance is worthwhile, we train a lightweight gate using paired team continuations, with supervision derived from final team outcomes and inference cost. Experiments on AppWorld and WebShop demonstrate that AOD consistently improves task performance over independent parallel execution while achieving lower inference cost than existing information-sharing baselines. These findings highlight the benefits of aligning peer communication with agents' evolving information needs to achieve more effective and efficient collaboration.

cs.AI↗

arXiv2026

Scale-free localization in pseudo-Hermitian quantum walk

Scale-free localization is a distinctive phenomenon in non-Hermitian systems, where specific eigenstates, known as scale-free localized states (LSs), possess localization lengths that scale linearly with the system size. Here, we show that a single non-Hermitian defect in one-dimensional quantum walks (QWs) induces scale-free localization. Driven by non-Hermiticity, the QW system undergoes a pseudo-Hermitian (PH) phase transition and the PH symmetry breaking is accompanied by the emergence of scale-free localization. As non-Hermiticity is further increased, the system undergoes a second phase transition, marked by the emergence of two pairs of bound states. Crucially, we identify the critical points of both transitions using localization properties. Furthermore, we extend our investigation to two-dimensional QWs, where a single non-Hermitian defect also induces scale-free localization, thereby demonstrating the universality of this phenomenon. Our work provides a valuable perspective for further exploring defect physics and a type of novel LS in non-Hermitian quantum systems.

quant-ph↗

arXiv2026

Spontaneous Chern-Euler Duality Transitions

Topological phase transitions are typically characterized by abrupt changes in a quantized invariant. Here we report a contrasting paradigm in spontaneous parity-time symmetry-breaking transitions induced by non-Hermiticity. The topological invariant remains conserved, but its nature transitions between the Chern number, characteristic of chiral transport in complex bands, and the Euler number, which characterizes the number of nodal points in pairs of real bands. This conserved morphing of topological properties applies to all pairs of bands whose spectral gaps to other bands remain open during the transition. It features qualitative changes in the non-Abelian geometric phases, which quantize in different forms across the transition. Our findings establish a generic topological duality principle governing transitions across symmetry classes and reveal unique connections among topology, symmetry, and non-Abelian gauge structure in non-unitary processes.

cond-mat.mes-hall↗

arXiv2026

Parallel Time-Aligned Spiking Self-Attention for Consistent Integer-Valued Training and Spike-Driven Inference

Integer-valued leaky integrate-and-fire (I-LIF) neurons and spike firing approximation (SFA) reduce temporal training cost by representing spike trains as firing counts and normalized firing rates, respectively. However, applying spiking self-attention (SSA) directly to these compressed query, key, and value representations introduces cross-time interactions that are absent during spike-driven inference. We term this operator-level discrepancy Temporal Interaction Mismatch (TIM). We propose Parallel Time-Aligned Spiking Self-Attention (PT-SSA), which reconstructs consecutive virtual spike slices from either I-LIF counts or SFA firing rates, computes attention only between time-aligned slices in parallel, and sums the per-step outputs. To accommodate the reduced attention output scale under SFA, we further introduce Adaptive PT-SSA, which learns a positive per-block rescaling before the output SFA neuron to improve firing-level utilization. Experiments on CIFAR-10, CIFAR-100, and ImageNet-1K show that the proposed methods substantially reduce train--inference mismatch. On CIFAR-100 with I-LIF, PT-SSA reduces the mean Top-1 gap from 1.85 to 0.29 percentage points. On ImageNet-1K, Adaptive PT-SSA reduces the Top-1 gap from 27.78 to 0.06 percentage points and achieves 74.53\% spike-driven Top-1 accuracy. A Triton-fused PT-SSA training kernel retains a $2.91\times$ throughput advantage over recurrent LIF SSA.

cs.NE↗

arXiv2026

Parent Hamiltonian and intrinsic phase transition in non-Hermitian photonic systems

Non-Hermitian systems host phenomena absent in Hermitian physics, but realizing Hamiltonians with intrinsic non-Hermitian properties remains challenging. The theoretical method of non-Hermitian parent Hamiltonian (NH-PH) enables the construction of a non-Hermitian system from a pair of matrix product states (MPSs) with tailored properties. Here, we report the first experimental generation of NH-PHs. This generation starts from MPSs that represent asymmetric Affleck--Kennedy--Lieb--Tasaki (AKLT) states. The construction is validated with single photons via imaginary-time evolution of the generated NH-PH to obtain its left and right ground states. We then characterize the properties of the system by measuring four different order parameters that probe non-reciprocal correlations, chiral imbalance, and conventional antiferromagnetic correlations. Furthermore, extending the framework to a larger system with a different model, we observe an intrinsic non-Hermitian phase transition, manifested by abrupt jumps of an order parameter when the designated zero-energy modes cease to be the globally lowest-energy states. Our work provides the first experimental realization and characterization of non-Hermitian Hamiltonians with controllable and customizable properties, opening new avenues for exploring intrinsic non-Hermitian phenomena across diverse physical platforms.

quant-ph↗

arXiv2026

Non-Hermitian-enhanced quantum sensing in an optical interferometer

The precision of quantum parameter estimation is traditionally constrained by the quantum Cramér-Rao bound, which is based on the Hermitian measurement framework. Recent studies of non-Hermitian systems have suggested new possibilities for enhancing parameter-estimation sensitivity. Here, we experimentally realize quantum parameter estimation using a non-Hermitian observable on a linear optical platform. The parameter is encoded in single-photon probe states and read out with a Sagnac interferometer, which allows us to reconstruct the complex expectation value of the implemented non-Hermitian observable from interference fringes. We observe a reduced error-propagation variance compared with the optimal Hermitian observable for the same probe-state model. This advantage remains visible under amplitude-damping noise. We further analyze the complete optical measurement as a physical positive-operator-valued measure (POVM) and show, through the corresponding classical Fisher information (CFI), that the observed non-Hermitian advantage is consistent with the standard quantum metrological limit when all output ports are included. Our results provide an experimental route to non-Hermitian observable readout and clarify its operational meaning in quantum sensing.

quant-ph↗

arXiv2026

Efficiently Training Time-to-First-Spike Spiking Neural Networks from Scratch

Spiking Neural Networks (SNNs), with their event-driven and biologically inspired mechanisms, are well-suited for energy-efficient neuromorphic hardware. Neural coding, which is critical to SNNs, determines how information is represented via spikes. While Time-to-First-Spike (TTFS) coding uses a single spike per neuron to offer extreme sparsity and energy efficiency, it often suffers from unstable training and low accuracy due to its sparse firing. To address these challenges, we propose a training framework that incorporates parameter initialization, training normalization, a temporal output decoder, and a re-evaluation of the pooling layer. The proposed parameter initialization and training normalization mitigate signal diminishing and gradient vanishing, which helps stabilize training. Our output decoder aggregates temporal spikes to encourage earlier firing, thereby reducing latency. The re-evaluation of the pooling layer demonstrates that max-pooling violates single-spike constraints, which should be avoided, whereas average-pooling preserves them. Experiments show that our framework stabilizes and accelerates training, reduces latency, and achieves state-of-the-art accuracy for step-by-step TTFS SNNs on MNIST (99.48%), Fashion-MNIST (92.90%), CIFAR10 (90.56%), CIFAR100 (70.27%) and DVS Gesture (95.83%).

cs.NE↗

arXiv2026

Chiral Entangled-State Generation through Dissipative Quantum Dynamics

Dissipation, though often detrimental to quantum entanglement, can be manipulated for the preparation of entangled states, wherein ingeniously designed quantum jump processes drive the system toward the desired steady state. Here we venture beyond this paradigm, and demonstrate a new type of entanglement generation in dissipative quantum dynamics. Combining driven-dissipative steady-state engineering and adiabatic passage, we propose a general protocol where the final entangled state depends on the chirality of the evolution path in the parameter space, a scheme that is further extendable to multipartite entanglement. By simulating the Liouvillian dynamics through the quantum Langevin equation for a pair of photons, we experimentally confirm the noise-resistant chiral preparation of various entangled states with high fidelity and concurrence. Our work establishes parametric chiral dynamics as a scalable and robust tool for controllable entanglement generation, paving the way for its applications in quantum information.

quant-ph↗

arXiv2026

Experimental Observation of Dynamical Phase Transitions in a Dephased Photonic Quantum Walk

Dynamical phase transitions in open quantum systems govern how non-equilibrium states relax toward a stationary state. We study these transitions experimentally using a discrete-time photonic quantum walk on a three-node graph. A tunable synthetic gauge flux and calibrated dephasing allow us to control time-reversal symmetry and the detailed balance properties of the effective Markovian dynamics. With detailed balance, we observe a first-order dynamical phase transition marked by a crossing of real Liouvillian eigenvalues. When detailed balance is broken, we observe a second-order dynamical phase transition at an exceptional point where eigenvalues and eigenvectors coalesce. By progressively reducing the dephasing strength, we track the crossover toward the quantum-coherent regime and determine that the transitions persist down to a finite threshold. Our results link Liouvillian spectral topology to relaxation criticality and demonstrate a controllable platform for engineered dissipative dynamics.

quant-ph↗

arXiv2026

Entanglement Dynamics with a Stochastic Non-Hermitian Hamiltonian away from Exceptional Points

Although non-Hermitian dynamics near exceptional points (EPs) provide a route to accelerated entanglement generation, entanglement can also be generated far from EPs at comparable or even higher rates. However, the behavior of such entanglement in open systems remains largely unexplored, rendering it highly susceptible to environmental noise. Here, we study entanglement dynamics in two coupled qubits described by a non-Hermitian Hamiltonian, where the dissipative rates are subject to stochastic noise. We focus on the regime away from EPs. Our approach demonstrates that, even in the presence of classical noise, rich dynamical control can be achieved, enabling highly efficient entanglement generation with a timescale that is significantly shorter than that of both Hermitian systems and EP-based non-Hermitian protocols. Additionally, the timescale is independent of the number of qubits, highlighting favorable scalability for multipartite entanglement generation and facilitating integration into future photonic quantum processors.

quant-ph↗

arXiv2026

Observation of Restored Adiabatic State Transfer in Time-Modulated Non-Hermitian Systems

Exceptional points (EPs) have attracted extensive research interest due to their intriguing properties. One of the hallmarks of EP physics is that dynamically encircling the EPs induces chiral mode switching, arising from the breakdown of adiabaticity due to the presence of a complex spectrum in the system's Hamiltonian. While such chiral mode behavior has been widely observed experimentally, achieving truly adiabatic, and thus symmetric, state transfer, regardless of the winding direction, in time-modulated non-Hermitian systems has remained elusive. In this work, we demonstrate that this long-sought adiabatic state dynamics can indeed be restored. By steering a two-mode photonic setup along specifically designed trajectories in parameter space, we realize conditions where the associated non-Hermitian evolution operator acquires a purely real spectrum. Moreover, our experimental platform enables controlled switching between symmetric (adiabatic) and chiral (non-adiabatic) state-transfer regimes for the same set of initial modes, thus effectively implementing a universal symmetric-asymmetric two-mode switch. Our results therefore open new avenues for harnessing unique topological spectral properties of non-Hermitian systems, paving the way for the practical design of versatile optical wave-manipulation devices and for advancing both classical and quantum information technologies.

physics.optics↗

arXiv2026

Decoherence Resilience of the Non-Hermitian Skin Effect

Decoherence and dissipation, arising from unavoidable interactions with the environment, can exert a dual influence on transport in physical systems, suppressing coherent propagation while inducing diffusion and mitigating localization in disordered systems. Non-Hermitian physics reveals a qualitatively different scenario, in which structured dissipation can induce directional bulk-to-boundary transport, known as the non-Hermitian skin effect (NHSE), that remains robust against disorder. Whether such transport can persist, be enhanced or hindered under decoherence, remains a largely open question. Here we experimentally address this question using photonic quantum walks with two tunable prototypical decoherence channels, dephasing and amplitude damping. Under dephasing, the NHSE survives up to the fully incoherent regime and is observed to even be enhanced by dephasing, yielding drift velocities that exceed those of coherent dynamics. By contrast, amplitude damping shows a pronounced order dependence: applied before the non-Hermitian loss operator, it suppresses and ultimately eliminates the NHSE in the fully incoherent limit; applied afterward, the NHSE persists and can be enhanced at sufficiently large loss strengths. Our work bridges quantum and classical non-Hermitian dynamics, demonstrates the resilience of the NHSE to decoherence, and opens avenues for harnessing decoherence to enhance directional transport in noisy, nonequilibrium systems.

quant-ph↗

arXiv2026

Non-Hermitian Anomalous Scaling Engineering

Non-Hermitian systems exhibit anomalous scaling, a striking departure from conventional bulk laws, rooted in the non-Hermitian skin effect (NHSE). Here, we experimentally uncover this scaling and demonstrate its active control in a temporal photonic lattice. By tracking the real-time evolution of all eigenstates as system size varies, we directly observe scaling-driven spectral reshaping and eigenstate localization, revealing phenomena absent in Hermitian or NHSE-free lattices. In a Su-Schrieffer-Heeger lattice, scaling alone can trigger a non-Hermitian topological phase transition, with edge modes remaining protected. Crucially, Kerr interactions open the frontier of nonlinear non-Hermitian physics: weak nonlinearity accelerates or decelerates anomalous scaling, while strong nonlinearity suppresses it entirely. These results establish the first experimental platform for linear and nonlinear anomalous scaling engineering, paving the way for compact non-Hermitian devices and exploration of nonlinear and many-body non-Hermitian phenomena.

physics.optics↗

arXiv2026

Observation of Braid-Protected Unpaired Exceptional Points

Spectral degeneracies (dubbed nodal points in momentum space) play fundamental roles in understanding exotic properties of light and matter. In lattice systems, unpaired band-structure degeneracies are subject to well-established no-go (doubling) theorems that universally apply to both closed Hermitian systems and open non-Hermitian systems. However, the non-Abelian braid topology of non-Hermitian multi-band systems provides a loophole to these constraints. Here we successfully leverage this loophole in a non-Hermitian three-band system, implementing an unpaired third-order exceptional point (EP3), which manifests as a non-Abelian monopole. We explicitly demonstrate the intricate braiding topology and non-Abelian, path-dependent, fusion rules underlying the unpaired EP3. The experiment uses a new design of single-photon interferometry, enabling eigenstate and spectral resolutions for multi-band systems with widely tunable parameters. Thus, the union of state-of-the-art experiments, fundamental theory, and everyday concepts such as braids pave the way toward the highly exotic non-Abelian topology unique to non-Hermitian settings.

cond-mat.mes-hall↗

arXiv2026

Observation of Criticality-Enhanced Quantum Sensing in Nonunitary Quantum Walks

Quantum physics enables parameter estimation with precisions beyond the capability of classical sensors. Quantum criticality is a key resource for this quantum-enhanced sensing, but experimental realization has been challenging due to the complexity of ground-state preparation and the long time required to reach the steady state near criticality. Here, we experimentally demonstrate critical enhancement in a non-Hermitian topological system using a photonic quantum walk setup. Our system supports two distinct phase transitions at which enhanced sensitivity is observed even at transient times for which the Bayesian inference shows excellent estimation and precision. It is a direct demonstration of criticality-enhanced scaling laws with non-unitary dynamics.

quant-ph↗

arXiv2026

Topological Sensing in the Dynamics of Quantum Walks with Defects

Topological quantum sensing leverages unique topological features to suppress noise and improve the precision of parameter estimation, emerging as a promising tool in both fundamental research and practical application. In this Letter, we propose a sensing protocol that exploits the dynamics of topological quantum walks incorporating localized defects. Unlike conventional schemes that rely on topological protection to suppress disorder and defects, our protocol harnesses the evolution time as a resource to enable precise estimation of the defect parameter. By utilizing topologically nontrivial properties of the quantum walks, the sensing precision can approach the Heisenberg limit. We further demonstrate the performance and robustness of the protocol through Bayesian estimation. Our results show that this approach maintains high precision over a broad range of parameters and exhibits strong robustness against disorder, offering a practical pathway for topologically enhanced quantum metrology.

quant-ph↗

arXiv2025

Calibrating and Rotating: A Unified Framework for Weight Conditioning in PEFT

Parameter-Efficient Fine-Tuning (PEFT) methods are crucial for adapting large pre-trained models. Among these, LoRA is considered a foundational approach. Building on this, the influential DoRA method enhances performance by decomposing weight updates into magnitude and direction. However, its underlying mechanism remains unclear, and it introduces significant computational overhead. In this work, we first identify that DoRA's success stems from its capacity to increase the singular value entropy of the weight update matrix, which promotes a more uniform update distribution akin to full fine-tuning. We then reformulate DoRA into a mathematically equivalent and more efficient matrix form, revealing it as a learnable weight conditioning method. Based on this insight, we propose a unified framework for designing advanced PEFT methods by exploring two orthogonal dimensions: the architectural placement and the transformation type of the conditioning matrix. Within this framework, we introduce two novel methods: (1) \textbf{Pre-Diag}, which applies a diagonal conditioning matrix before the LoRA update to efficiently calibrate the pre-trained weights, thereby enhancing performance while reducing training time; and (2) \textbf{S}kewed \textbf{O}rthogonal \textbf{R}otation \textbf{A}daptation (\textbf{SORA}), which employs a parameter-efficient orthogonal rotation to perform a more powerful, norm-preserving transformation of the feature space. Extensive experiments on natural language understanding and generation tasks demonstrate that our proposed methods achieve superior performance and efficiency compared to both LoRA and DoRA. The code is available at https://github.com/MaeChd/SORA.

cs.LG↗

arXiv2025

Multiplication-Free Parallelizable Spiking Neurons with Efficient Spatio-Temporal Dynamics

Spiking Neural Networks (SNNs) are distinguished from Artificial Neural Networks (ANNs) for their complex neuronal dynamics and sparse binary activations (spikes) inspired by the biological neural system. Traditional neuron models use iterative step-by-step dynamics, resulting in serial computation and slow training speed of SNNs. Recently, parallelizable spiking neuron models have been proposed to fully utilize the massive parallel computing ability of graphics processing units to accelerate the training of SNNs. However, existing parallelizable spiking neuron models involve dense floating operations and can only achieve high long-term dependencies learning ability with a large order at the cost of huge computational and memory costs. To solve the dilemma of performance and costs, we propose the mul-free channel-wise Parallel Spiking Neuron, which is hardware-friendly and suitable for SNNs' resource-restricted application scenarios. The proposed neuron imports the channel-wise convolution to enhance the learning ability, induces the sawtooth dilations to reduce the neuron order, and employs the bit-shift operation to avoid multiplications. The algorithm for the design and implementation of acceleration methods is discussed extensively. Our methods are validated in neuromorphic Spiking Heidelberg Digits voices, sequential CIFAR images, and neuromorphic DVS-Lip vision datasets, achieving superior performance over SOTA spiking neurons. Training speed results demonstrate the effectiveness of our acceleration methods, providing a practical reference for future research. Our code is available at \href{https://github.com/PengXue0812/Multiplication-Free-Parallelizable-Spiking-Neurons-with-Efficient-Spatio-Temporal-Dynamics}{Github}.

cs.NE↗