arXiv subjects

Kai Sun

Publications and source records attributed to Kai Sun.

At least 73 records · Page 4Linked to original sources

arXiv2026

NCCL EP: Towards a Unified Expert Parallel Communication API for NCCL

Mixture-of-Experts (MoE) architectures have become essential for scaling large language models, driving the development of specialized device-initiated communication libraries such as DeepEP, Hybrid-EP, and others. These libraries demonstrate the performance benefits of GPU-initiated RDMA for MoE dispatch and combine operations. This paper presents NCCL EP (Expert Parallelism), a ground-up MoE communication library built entirely on NCCL's Device API. NCCL EP provides unified ncclEpDispatch and ncclEpCombine primitives with both C and Python interfaces, supporting Low-Latency (LL) mode for inference decoding and High-Throughput (HT) mode for training and inference prefill. LL targets small batch sizes (1-128 tokens) using direct all-to-all RDMA+NVLink mesh connectivity with double-buffered communication for overlapping dispatch and combine phases. HT targets large batches (4096+ tokens) using hierarchical communication that aggregates tokens within NVLink domains before inter-node RDMA transmission. Both modes leverage Device API for both intra- and inter-node communications, taking advantage of its topology awareness and optimized GPU-initiated implementation. We evaluate NCCL EP on an H100-based cluster across multi-node configurations, demonstrating competitive LL kernel performance and presenting end-to-end results with vLLM integration. By building MoE communication natively within NCCL, NCCL EP provides a supported path for expert parallelism on current and emerging NVIDIA platforms.

cs.DC↗

arXiv2026

Effects of gas diffusion layer thickness on PEM fuel cells with composite foam-rib flow fields

Gas diffusion layers (GDLs) play a crucial role for the performance of proton exchange membrane fuel cells (PEMFCs). The utilization of composite foam-rib flow fields (CFRFFs) can alter the reactant gas transfer pattern, hence improving the efficiency of under-rib reactant gas transfer and water drainage. The impact of the cathode and anode GDL thicknesses (h_{c,GDL} and h_{a,GDL}) on the performance of CFRFF design is investigated by three-dimensional multiphase non-isothermal numerical simulation in this study. The results indicate that for the conventional rib flow field (CRFF) design, there is an optimal h_{c,GDL} for optimal cell performance, while for the CFRFF design, as h_{c,GDL} becomes thinner, the cell performance increases, and the trend is dominated by the variation of the oxygen concentration. Under a thin GDL, the rib width of the CRFF design should be as small as possible to minimize concentration polarization loss, while the rib width of the CFRFF design can be slightly larger. Furthermore, by decreasing the thickness of h_{a,GDL} in both the CRFF and CFRFF designs, there is an increase in the dissolved water content in the ionomer of the cathode CL and a subsequent decrease in the Ohmic polarization loss.

physics.flu-dyn↗

arXiv2026

MorphSNN: Adaptive Graph Diffusion and Structural Plasticity for Spiking Neural Networks

Spiking Neural Networks (SNNs) currently face a critical bottleneck: while individual neurons exhibit dynamic biological properties, their macro-scopic architectures remain confined within conventional connectivity patterns that are static and hierarchical. This discrepancy between neuron-level dynamics and network-level fixed connectivity eliminates critical brain-like lateral interactions, limiting adaptability in changing environments. To address this, we propose MorphSNN, a backbone framework inspired by biological non-synaptic diffusion and structural plasticity. Specifically, we introduce a Graph Diffusion (GD)mechanism to facilitate efficient undirected signal propagation, complementing the feedforward hierarchy. Furthermore, it incorporates a Spatio-Temporal Structural Plasticity (STSP) mechanism, endowing the network with the capability for instance-specific, dynamic topological reorganization, thereby overcoming the limitations of fixed topologies. Experiments demonstrate that MorphSNN achieves state-of-the-art accuracy on static and neuromorphic datasets; for instance, it reaches 83.35% accuracy on N-Caltech101 with only 5 timesteps. More importantly, its self-evolving topology functions as an intrinsic distribution fingerprint, enabling superior Out-of- Distribution (OOD) detection without auxiliary training. The code is available at anonymous.4open.science/r/MorphSNN-B0BC.

cs.NE↗

arXiv2026

Mitigating Instance Entanglement in Instance-Dependent Partial Label Learning

Partial label learning is a prominent weakly supervised classification task, where each training instance is ambiguously labeled with a set of candidate labels. In real-world scenarios, candidate labels are often influenced by instance features, leading to the emergence of instance-dependent PLL (ID-PLL), a setting that more accurately reflects this relationship. A significant challenge in ID-PLL is instance entanglement, where instances from similar classes share overlapping features and candidate labels, resulting in increased class confusion. To address this issue, we propose a novel Class-specific Augmentation based Disentanglement (CAD) framework, which tackles instance entanglement by both intra- and inter-class regulations. For intra-class regulation, CAD amplifies class-specific features to generate class-wise augmentations and aligns same-class augmentations across instances. For inter-class regulation, CAD introduces a weighted penalty loss function that applies stronger penalties to more ambiguous labels, encouraging larger inter-class distances. By jointly applying intra- and inter-class regulations, CAD improves the clarity of class boundaries and reduces class confusion caused by entanglement. Extensive experimental results demonstrate the effectiveness of CAD in mitigating the entanglement problem and enhancing ID-PLL performance. The code is available at https://github.com/RyanZhaoIc/CAD.git.

cs.CV↗

arXiv2026

Kondo driven suppression of charge density wave in Van der Waals material UTe$_3$

Competing electronic instabilities lie at the heart of emergent phenomena in quantum materials. In low-dimensional metals, Fermi-surface nesting can drive charge density wave (CDW) formation through a Peierls-like mechanism, while in strongly correlated systems, Kondo hybridization reconstructs the electronic structure by entangling localized moments with itinerant electrons. How these two fundamentally different instabilities interact$-$whether they coexist, compete, or mutually exclude each other$-$remains an open question. Here, we present suppression of charge density wave via the Kondo interaction in van der Waals material UTe$_3$. The angle-resolved photoemission spectroscopy (ARPES) data reveals Fermi surface nesting under similar conditions as seen in RETe$_3$ compounds. Despite that, no CDW is found in UTe$_3$ after an extensive search. We demonstrate that strong hybridization between U 5$f$ electrons and Te $p$ states reconstructs the low-energy electronic structure, removes the instability, and preempts CDW formation. Our results reveal a rare example where Kondo hybridization preempts density wave formation, offering a new route to controlling ordering phenomena in correlated 2D materials.

cond-mat.str-el↗

arXiv2026

SceneTransporter: Optimal Transport-Guided Compositional Latent Diffusion for Single-Image Structured 3D Scene Generation

We introduce SceneTransporter, an end-to-end framework for structured 3D scene generation from a single image. While existing methods generate part-level 3D objects, they often fail to organize these parts into distinct instances in open-world scenes. Through a debiased clustering probe, we reveal a critical insight: this failure stems from the lack of structural constraints within the model's internal assignment mechanism. Based on this finding, we reframe the task of structured 3D scene generation as a global correlation assignment problem. To solve this, SceneTransporter formulates and solves an entropic Optimal Transport (OT) objective within the denoising loop of the compositional DiT model. This formulation imposes two powerful structural constraints. First, the resulting transport plan gates cross-attention to enforce an exclusive, one-to-one routing of image patches to part-level 3D latents, preventing entanglement. Second, the competitive nature of the transport encourages the grouping of similar patches, a process that is further regularized by an edge-based cost, to form coherent objects and prevent fragmentation. Extensive experiments show that SceneTransporter outperforms existing methods on open-world scene generation, significantly improving instance-level coherence and geometric fidelity. Code and models will be publicly available at https://2019epwl.github.io/SceneTransporter/.

cs.CV↗

arXiv2026

Layer-dependent antiferromagnetic Chern and axion insulating states in UOTe

Magnetic topological insulators have received significant interest due to their dissipationless edge states, which promise advances in energy-efficient electronic transport. However, the magnetic topological insulator state has typically been found in ferromagnets (FMs) that suffer from low magnetic ordering temperatures and stray fields. Identifying an antiferromagnetic topological insulator that exhibits the quantum anomalous Hall effect (QAHE) with a relatively high Néel temperature has been a longstanding challenge. Here, we focus on the recently discovered van der Waals (vdW) antiferromagnet (AFM) UOTe, which not only features a high Néel temperature ($\sim$150K) but also exhibits intriguing Kondo interaction and topological characteristics. Our systematic analysis of the layer-dependent topological phases based on \textit{ab} initio computations predicts the two-layer UOTe film to be an ideal 2D AFM Chern insulator in which the Hall conductivity is quantized with a fully compensated spin magnetization. By applying an in-plane strain or electric field, we show how the itinerancy of U-5f electrons can be manipulated to trigger a transition between the nontrivial ($C = 1$) and trivial ($C = 0$) phases. Interestingly, the 3-layer UOTe film is found to have zero charge conductance but it hosts a quantized spin Hall conductivity (SHC) with finite magneto-electric coupling, suggesting the presence of an axion insulator-like state. The unique magnetic structure of UOTe supports a layer-tunable topology in which films with an odd number of layers are axion-like insulators, while films with an even number of layers are Chern insulators, and the bulk material is a Dirac semimetal. Our study offers a new intrinsic AFM materials platform for realizing correlated topological phases for next-generation spintronics applications and fundamental science studies.

cond-mat.mtrl-sci↗

arXiv2026

Correlation between 2D Square Ice and 3D Bulk Ice by Critical Crystallization Pressure

Low-dimensional ice trapped in nanocapillaries is a fascinating phenomenon and is ubiquitous in our daily lives. As a decisive factor of the confinement effect, the size of nanocapillary significantly affects the critical crystallization pressure and crystalline structure, especially for multi-layered ices. By choosing square ice as a typical two-dimensional (2D) multi-layered ice pattern and using all-atom molecular dynamics simulations, we further unveil the variation mechanism of critical crystallization pressure with the nanocapillary size. The results show a strong dependence of the critical crystallization pressure on the size of the graphene sheet for monolayer, bilayer, and trilayer square ice. The quasi-macroscopic crystallization pressure, the actual pressure of water molecules, and the freezable region between them are all strongly dependent on the nanocapillary width. As the size of the capillary becomes larger in all three directions, the critical crystallization pressure converges to the true macroscopic crystallization pressure, which is very close to the value of the crystallization pressure for bulk ice. A direct correlation is established between 2D square ice and three-dimensional (3D) bulk ice by the critical crystallization pressure. There is an unfreezable threshold for crystallizing spontaneously in practice when the quasi-macroscopic crystallization pressure is equal to the actual pressure, which can explain the limit of nanocapillary width for multi-layered ice.

physics.chem-ph↗

arXiv2026

Experimental Certification of Quantum Measurements with Maximally Mixed States

So far, certifying quantum devices from their input-output statistics, under minimal assumptions, required the preparation of specific pure quantum states. Recently, Xu et al. [Phys. Rev. Lett. 132, 140201 (2024)] have demonstrated that certain sets of quantum observables can be certified using any state of full rank. However, their method is restricted to ideal conditions. Here, we address this problem and present an experimentally robust method that eliminates the need of preparing states with high fidelity with respect to specific pure states. We demonstrate the feasibility of the method by experimentally certifying photonic devices implementing Peres' set of 24 ququart observables [J. Phys. A 24, L175 (1991)] and Yu and Oh's set of 13 qutrit observables [Phys. Rev. Lett. 108, 030402 (2012)], using maximally mixed states as input. This approach offers a crucial advantage for certifying high-dimensional quantum systems, since it works with maximally mixed and thermal states.

quant-ph↗

arXiv2026

Cumulative Path-Level Semantic Reasoning for Inductive Knowledge Graph Completion

Conventional Knowledge Graph Completion (KGC) methods aim to infer missing information in incomplete Knowledge Graphs (KGs) by leveraging existing information, which struggle to perform effectively in scenarios involving emerging entities. Inductive KGC methods can handle the emerging entities and relations in KGs, offering greater dynamic adaptability. While existing inductive KGC methods have achieved some success, they also face challenges, such as susceptibility to noisy structural information during reasoning and difficulty in capturing long-range dependencies in reasoning paths. To address these challenges, this paper proposes the Cumulative Path-Level Semantic Reasoning for inductive knowledge graph completion (CPSR) framework, which simultaneously captures both the structural and semantic information of KGs to enhance the inductive KGC task. Specifically, the proposed CPSR employs a query-dependent masking module to adaptively mask noisy structural information while retaining important information closely related to the targets. Additionally, CPSR introduces a global semantic scoring module that evaluates both the individual contributions and the collective impact of nodes along the reasoning path within KGs. The experimental results demonstrate that CPSR achieves state-of-the-art performance.

cs.AI↗

arXiv2025

WearVox: An Egocentric Multichannel Voice Assistant Benchmark for Wearables

Wearable devices such as AI glasses are transforming voice assistants into always-available, hands-free collaborators that integrate seamlessly with daily life, but they also introduce challenges like egocentric audio affected by motion and noise, rapid micro-interactions, and the need to distinguish device-directed speech from background conversations. Existing benchmarks largely overlook these complexities, focusing instead on clean or generic conversational audio. To bridge this gap, we present WearVox, the first benchmark designed to rigorously evaluate voice assistants in realistic wearable scenarios. WearVox comprises 3,842 multi-channel, egocentric audio recordings collected via AI glasses across five diverse tasks including Search-Grounded QA, Closed-Book QA, Side-Talk Rejection, Tool Calling, and Speech Translation, spanning a wide range of indoor and outdoor environments and acoustic conditions. Each recording is accompanied by rich metadata, enabling nuanced analysis of model performance under real-world constraints. We benchmark leading proprietary and open-source speech Large Language Models (SLLMs) and find that most real-time SLLMs achieve accuracies on WearVox ranging from 29% to 59%, with substantial performance degradation on noisy outdoor audio, underscoring the difficulty and realism of the benchmark. Additionally, we conduct a case study with two new SLLMs that perform inference with single-channel and multi-channel audio, demonstrating that multi-channel audio inputs significantly enhance model robustness to environmental noise and improve discrimination between device-directed and background speech. Our results highlight the critical importance of spatial audio cues for context-aware voice assistants and establish WearVox as a comprehensive testbed for advancing wearable voice AI research.

cs.CL↗

arXiv2025

High-fidelity realisation of CNOT gate in Majorana-based optical platform

We present the experimental realisation of a robust CNOT quantum gate using Majorana zero modes simulated on a photonic platform. Three Kitaev chains supporting Majorana zero modes at their endpoints are used to encode two logical qubits, and both intra-chain and inter-chain braiding operations are performed to implement the CNOT gate. While the topological encoding of quantum information in Majorana fermions does not offer full topological protection in our non-interacting photonic setting, it nevertheless exhibits a natural resilience to the dominant noise and decoherence effects present in the experiment. Consequently, the fidelity of the CNOT gate is significantly enhanced, surpassing 0.992 and addressing a key limitation in the path toward scalable quantum computation. These results represent a major advancement in topological quantum computing with Majorana fermions and underscore the potential of photonic platforms for realising high-fidelity quantum gates.

quant-ph↗

arXiv2025

Magneto-Moiré Excitons in Twisted Bilayer CrSBr

Moiré superlattices in van der Waals materials have revolutionized the study of electronic and excitonic systems by creating periodic electrostatic potentials. Extending this concept to magnetic materials promises new pathways in merging spintronics with photonics. While moiré magnetism has been revealed with near-field probes and nonlinear optical techniques, the coupling of these magnetic textures to optical excitations - magneto-moiré excitons - remains unexplored. Here, we report the observation of magneto-moiré excitons in twisted bilayer CrSBr, correlated with moiré spin textures that emerge below a critical twist angle of ~2°. The nanoscale moiré spin texture imprints distinct signatures onto the optical spectrum, shifting the exciton energy via a periodic magnetic exchange field. First-principles calculations corroborate that these signatures arise from one-dimensional spin textures governed by the balance of exchange interactions and domain wall energy. Our results demonstrate that moiré magnetism can be used to engineer nanoscale excitonic energy landscapes, providing a new platform for magneto-optical sensing, quantum transduction, and control of non-collinear magnetism and topology through light.

cond-mat.mes-hall↗

arXiv2025

CogniSNN: An Exploration to Random Graph Architecture based Spiking Neural Networks with Enhanced Depth-Scalability and Path-Plasticity

Currently, most spiking neural networks (SNNs) still mimic the chain-like hierarchical architecture in traditional artificial neural networks (ANNs). This method significantly differs from random connections between neurons found in biological brains, limiting the ability to model the evolving mechanisms of neural pathways in biological neural systems, particularly in terms of dynamic depth-scalability and adaptive path-plasticity. This paper develops a new modeling paradigm for SNNs with random graph architecture (RGA), termed Cognition-aware SNN (CogniSNN). Furthermore, we model the depth-scalability and path-plasticity in CogniSNN by introducing a modified spiking residual neural node (ResNode) to counteract network degradation in deeper graph pathways, as well as a critical path-based algorithm that enables CogniSNN to perform path reusability on new tasks leveraging the features of the data and the RGA learned in old tasks. Experiments show that the performance of CogniSNN with redesigned ResNode is comparable, even superior, to current state-of-the-art SNNs on neuromorphic datasets. The critical path-based approach effectively achieves path reuse capability while maintaining expected performance in learning new tasks that are similar to or distinct from the old ones. This study showcases the potential of RGA-based SNNs and paves a new path for modeling the fusion of computational neuroscience and deep intelligent agents. The code is available at github.com/Yongsheng124/CogniSNN.

cs.NE↗

arXiv2025

Probing Topological Surface States and Conduction via Extended Defects in (Bi$_{1-x}$Sb$_x$)$_2$Te$_3$ Films

(Bi$_{1-x}$Sb$_x$)$_2$Te$_3$ alloys are non-degenerate topological insulators (TIs) whose Dirac point (DP) can be tuned within the bulk bandgap by varying the composition, effectively reducing bulk conduction while allowing surface carrier conduction. Magnetotransport measurements of a series of (Bi$_{1-x}$Sb$_x$)$_2$Te$_3$ thin films indicate electron-dominated conduction, with weak anti-localization attributed to topological surface states (TSSs). Due to the similarity of phase coherence lengths and twin boundary spacings ($\sim$100 nm), we consider the role of twin boundaries as additional conducting paths. Density functional theory calculations reveal an enhanced density of states near the Fermi level at $60^\circ$ twin boundaries, with 2D carrier concentration in excess of $3 \times 10^{13}$ cm$^{-2}$. Furthermore, an analysis of the longitudinal magnetoconductivity yields an upper bound of $7.3 \times 10^{-4}$ S for twin boundary conductivity, resulting in a carrier mobility as high as $142$ cm$^2$/(V$\cdot$s). We discuss the role of twin boundaries in facilitating a transition from a massive Dirac cone dispersion to gapless, topologically protected surface states. Understanding the role of twin boundaries on carrier conduction in non-degenerate TIs is critical for the development of novel TI-based electronic devices.

cond-mat.mtrl-sci↗

arXiv2025

A Voronoi Diagram-Based Approach for AC Optimal Power Flow

The primary goal of Optimal Power Flow (OPF) is to optimize the operation of a power system while meeting the demand and adhering to operational constraints. This paper presents a new approach for AC OPF. First, the approach constructs a Voronoi diagram by distributing multiple sample points representing potential solutions throughout the search space. Then, it recursively adds new sample points including a tentative optimal point from the continuous gradient-projection method, a point in the most sparsely populated region to ensure high fidelity, and the connecting point, until the stopping criterion is met. The proposed approach is illustrated in detail using the IEEE 9-bus system and then validated on the IEEE 39-bus and 118-bus systems to verify the quality of the obtained solution.

math.OC↗

arXiv2025

CogniSNN: Enabling Neuron-Expandability, Pathway-Reusability, and Dynamic-Configurability with Random Graph Architectures in Spiking Neural Networks

Spiking neural networks (SNNs), regarded as the third generation of artificial neural networks, are expected to bridge the gap between artificial intelligence and computational neuroscience. However, most mainstream SNN research directly adopts the rigid, chain-like hierarchical architecture of traditional artificial neural networks (ANNs), ignoring key structural characteristics of the brain. Biological neurons are stochastically interconnected, forming complex neural pathways that exhibit Neuron-Expandability, Pathway-Reusability, and Dynamic-Configurability. In this paper, we introduce a new SNN paradigm, named Cognition-aware SNN (CogniSNN), by incorporating Random Graph Architecture (RGA). Furthermore, we address the issues of network degradation and dimensional mismatch in deep pathways by introducing an improved pure spiking residual mechanism alongside an adaptive pooling strategy. Then, we design a Key Pathway-based Learning without Forgetting (KP-LwF) approach, which selectively reuses critical neural pathways while retaining historical knowledge, enabling efficient multi-task transfer. Finally, we propose a Dynamic Growth Learning (DGL) algorithm that allows neurons and synapses to grow dynamically along the internal temporal dimension. Extensive experiments demonstrate that CogniSNN achieves performance comparable to, or even surpassing, current state-of-the-art SNNs on neuromorphic datasets and Tiny-ImageNet. The Pathway-Reusability enhances the network's continuous learning capability across different scenarios, while the dynamic growth algorithm improves robustness against interference and mitigates the fixed-timestep constraints during neuromorphic chip deployment. This work demonstrates the potential of SNNs with random graph structures in advancing brain-inspired intelligence and lays the foundation for their practical application on neuromorphic hardware.

cs.NE↗

arXiv2025

Two-phase flow in porous metal foam flow fields of PEM fuel cells

Porous metal foam (PMF) flow field is a potential option for proton exchange membrane fuel cells (PEMFCs) due to its excellent capabilities in gas distribution and water drainage. However, the gas-liquid two-phase flow in the PMF flow field on the pore scale is still unclear. In this study, we investigate the gas-liquid two-phase flow in the PMF flow field. Film, plug, and ligament flows are found in the hydrophilic PMF flow field, while slug and droplet flows are found in the hydrophobic PMF flow field. The results suggest that optimizing the pore size, increasing the metal foam surface hydrophobicity, and optimizing the operating condition are helpful for the water management of the PMF flow field. The frequency analysis of the pressure drop also shows that the dominant frequency can be used as an indicator to analyze the transition between different flow patterns.

physics.flu-dyn↗