arXiv subjects

Chen Yang

Publications and source records attributed to Chen Yang.

At least 91 records · Page 5Linked to original sources

arXiv2026

Decentralized Attention Fails Centralized Signals: Rethinking Transformers for Medical Time Series

Accurate analysis of medical time series (MedTS) data, such as electroencephalography (EEG) and electrocardiography (ECG), plays a pivotal role in healthcare applications, including the diagnosis of brain and heart diseases. MedTS data typically exhibit two critical patterns: temporal dependencies within individual channels and channel dependencies across multiple channels. While recent advances in deep learning have leveraged Transformer-based models to effectively capture temporal dependencies, they often struggle with modeling channel dependencies. This limitation stems from a structural mismatch: MedTS signals are inherently centralized, whereas the Transformer's attention mechanism is decentralized, making it less effective at capturing global synchronization and unified waveform patterns. To address this mismatch, we propose CoTAR (Core Token Aggregation-Redistribution), a centralized MLP-based module designed to replace decentralized attention. Instead of allowing all tokens to interact directly, as in standard attention, CoTAR introduces a global core token that serves as a proxy to facilitate inter-token interactions, thereby enforcing a centralized aggregation and redistribution strategy. This design not only better aligns with the centralized nature of MedTS signals but also reduces computational complexity from quadratic to linear. Experiments on five benchmarks validate the superiority of our method in both effectiveness and efficiency, achieving up to a 11.6% improvement on the APAVA dataset, while using only 33% of the memory and 20% of the inference time compared to the previous state of the art. Code and all training scripts are available at https://github.com/Levi-Ackman/TeCh.

cs.LG↗

arXiv2026

Koopman Identification of Nonlinear Systems via Reservoir Liftings

Learning tractable linear representations of nonlinear dynamical systems via Koopman operator theory is often hindered by dictionary selection, temporal memory encoding, and numerical ill-conditioning. Inspired by Reservoir Computing (RC) paradigm, this paper introduces the RC-Koopman framework, which interprets reservoir as a stateful, finite-dimensional Koopman dictionary whose temporal depth is explicitly controlled by its spectral radius. We show that the Echo State Property (ESP) guarantees well-posedness and favorable numerical conditioning of the lifted Koopman approximation. A correlation-based spectral radius selection algorithm aligns reservoir memory with dominant system timescales. Analysis reveals how the finite memory of the reservoir determines which Koopman eigenfunctions remain observable from the lifted features. Evaluation on synthetic benchmarks demonstrates that RC-Koopman achieves a favorable balance between reconstruction accuracy of the underlying nonlinear dynamics and dynamical stability, compared to Extended Dynamic Mode Decomposition (EDMD) and Hankel-based lifting approaches. Code available at: https://github.com/NEAR-the-future/RC-Koopman.git

cs.LG↗

arXiv2026

Strongly Coupled Sectors in Inflation: Gapped Theories of Unparticles

We consider a novel scenario for a strongly coupled spectator sector during inflation, that of a higher dimensional conformal field theory with large anomalous dimensions -- ``unparticles'' -- and compactify the extra dimensions. More specifically, we take generalized free fields in five dimensions, where the extra dimension is compactified to a circle. Due to the usual Kaluza-Klein mechanism, the resulting excitations carry properties of both particles and unparticles, so we dub this scenario ``gapped unparticles''. We derive a two-point function of the gapped unparticles by performing dimensional reduction. We then compute, in the collapsed limit, the four-point correlation function of conformally coupled scalars exchanging a gapped unparticle, which are used as seed functions to obtain the correlation function of primordial density perturbations. The phenomenology of the resulting correlators presents some novel features, such as oscillations with an envelope controlled by the anomalous dimension, rather than the usual value of 3/2. Depending on the value of the five-dimensional scaling dimension and effective mass of the gapped unparticles, we find a clear strategy to distinguish gapped unparticles from heavy massive scalars. If we assume the interactions are localized on a brane, gapped unparticles with different effective masses will share a universal coupling, and their exchanges produce an interesting interference pattern.

hep-th↗

arXiv2026

Evidence for Quark Confinement in the Proton

The strong interaction is the fundamental force that holds quarks and the gluon force carriers together to form protons and neutrons and also binds the atomic nucleus. The theory governing quark-gluon interactions is Quantum Chromodynamics (QCD). A wide variety of experimental data teaches us that quarks and gluons cannot be observed in isolation, a phenomenon known as confinement that is unique to QCD. But no one has used QCD to mathematically prove confinement. Here we show how to define and measure the force on quarks in the proton using available experimental data. Direct evidence for confinement is obtained because the force is found to be attractive and constant for a wide range of quark positions. This work guides future experimental efforts at future Electron-Ion Colliders aimed at obtaining a rigorous quantitative understanding of confinement and the origin of nuclear mass.

hep-ph↗

arXiv2026

Birds of a Feather Cluster Nearby: a Proximity-Aware Geo-Codebook for Local Service Recommendation

Generative recommendation systems are increasingly adopted in local service platforms, where semantic relevance alone is insufficient without strict geographic feasibility. A key technical challenge lies in semantic ID (SID) tokenization, which directly impacts the recommendation performance. However, existing semantic codebooks neglect geographic constraints, often resulting in recommendations that are semantically relevant yet geographically unreachable. To address this limitation, we propose Pro-GEO, a Proximity-aware GEO-codebook. Pro-GEO establishes a geo-centroid local coordinate system to capture intra-cluster spatial relationships and a geo-rotary position encoding mechanism that models geographic proximity as orthogonal rotational transformations in the high-dimensional embedding. This design enables semantic and spatial signals to be jointly modeled in a balanced manner, without reducing geographic information to a weak auxiliary feature. Extensive experiments conducted on a large-scale industrial dataset reveal that Pro-GEO significantly outperforms state-of-the-art methods. In particular, Pro-GEO reduces the average geographic clustering distance by 45.60% and achieves a 1.87% improvement in Hit@50, highlighting its effectiveness for real-world local service recommendation.

cs.IR↗

arXiv2026

Channel Attention-Guided Cross-Modal Knowledge Distillation for Referring Image Segmentation

Referring image segmentation (RIS) requires accurate segmentation of target regions in images according to language descriptions, which is a cross-modal task integrating vision and language. Existing RIS methods typically employ large-scale vision and language encoding models to improve performance, but their enormous parameter size severely restricts deployment in scenarios with limited computing resources. To solve this problem, this paper proposes a channel attention-guided cross-modal knowledge distillation method, which transfers the high-order fine-grained correlations between vision and language learned by the teacher network, as well as the correlations between semantic components represented by each channel, to the student network. Compared with the traditional pixel-wise relational distillation, this method not only enables the student to learn the knowledge of the teacher, but also retains part of its independent learning ability, alleviating the transfer of learning bias. Experimental results on two public datasets show that the proposed distillation method does not introduce additional parameters during inference and can achieve significant performance improvement for the student model.

cs.CV↗

arXiv2026

TowerDataset: A Heterogeneous Benchmark for Transmission Corridor Segmentation with a Global-Local Fusion Framework

Fine-grained semantic segmentation of transmission-corridor point clouds is fundamental for intelligent power-line inspection. However, current progress is limited by realistic data scarcity and the difficulty of modeling global corridor structure and local geometric details in long, heterogeneous scenes. Existing public datasets usually provide only a few coarse categories or short cropped scenes which overlook long-range structural dependencies, severe long-tail distributions, and subtle distinctions among safety-critical components. As a result, current methods are difficult to evaluate under realistic inspection settings, and their ability to preserve and integrate complementary global and local cues remains unclear. To address the above challenges, we introduce TowerDataset, a heterogeneous benchmark for transmission-corridor segmentation. TowerDataset contains 661 real-world scenes and about 2.466 billion points. It preserves long corridor extents, defines a fine-grained 22-class taxonomy, and provides standardized splits and evaluation protocols. In addition, we present a global-local fusion framework which preserves and fuses whole-scene and local-detail information. A whole-scene branch with NoCrop training and prototypical contrastive learning captures long-range topology and contextual dependencies. A block-wise local branch retains fine geometric structures. Both predictions are then fused and refined by geometric validation. This design allows the model to exploit both global relationships and local shape details when recognizing rare and confusing components. Experiments on TowerDataset and two public benchmarks demonstrate the challenge of the proposed benchmark and the robustness of our framework in real, complex, and heterogeneous transmission-corridor scenes. The dataset will be released soon at https://huggingface.co/datasets/tccx18/Towerdataset/tree/main.

cs.CV↗

arXiv2026

Aziz and Howl's Gravity-Induced Entanglement Channel is Essentially Classical Mechanics

Aziz and Howl argued that a classical gravitational field can generate quantum entanglement through a quantum-field-theoretic channel mediated by virtual matter propagation. However, their claimed channel is more naturally and accurately understood as semiclassical wavepacket motion in an external gravitational field, rather than as a distinctively quantum-field-theoretic entangling effect. Moreover, the result of their perturbative computation is incorrectly magnified: they selected a discontinuous wavefunction with infinite kinetic energy as the initial state and simultaneously treated it as stationary. Once a correct treatment using Gaussian wavepacket is adapted, the resulting effect will be negligibly small.

quant-ph↗

arXiv2026

Tuning Language Models for Robust Prediction of Diverse User Behaviors

Predicting user behavior is essential for intelligent assistant services, yet deep learning models often struggle to capture long-tailed behaviors. Large language models (LLMs), with their pretraining on vast corpora containing rich behavioral knowledge, offer promise. However, existing fine-tuning approaches tend to overfit to frequent ``anchor'' behaviors, reducing their ability to predict less common ``tail'' behaviors. In this paper, we introduce BehaviorLM, a progressive fine-tuning approach that addresses this issue. In the first stage, LLMs are fine-tuned on anchor behaviors while preserving general behavioral knowledge. In the second stage, fine-tuning uses a balanced subset of all behaviors based on sample difficulty to improve tail behavior predictions without sacrificing anchor performance. Experimental results on two real-world datasets demonstrate that BehaviorLM robustly predicts both anchor and tail behaviors and effectively leverages LLM behavioral knowledge to master tail behavior prediction with few-shot examples.

cs.CL↗

arXiv2026

Advances and Challenges in Solar Flare Prediction: A Review

Solar flares, as one of the most prominent manifestations of solar activity, have a profound impact on both the Earth's space environment and human activities. As a result, accurate solar flare prediction has emerged as a central topic in space weather research. In recent years, substantial progress has been made in the field of solar flare forecasting, driven by the rapid advancements in space observation technology and the continuous improvement of data processing capabilities. This paper presents a comprehensive review of the current state of research in this area, with a particular focus on tracing the evolution of data-driven approaches -- which have progressed from early statistical learning techniques to more sophisticated machine learning and deep learning paradigms, and most recently, to the emergence of Multimodal Large Models (MLMs). Furthermore, this study examines the realistic performance of existing flare forecasting platforms, elucidating their limitations in operational space weather applications and thereby offering a practical reference for future advancements in technological optimization and system design.

astro-ph.SR↗

arXiv2026

Realtime-VLA V2: Learning to Run VLAs Fast, Smooth, and Accurate

In deployment of the VLA models to real-world robotic tasks, execution speed matters. In previous work arXiv:2510.26742 we analyze how to make neural computation of VLAs on GPU fast. However, we leave the question of how to actually deploy the VLA system on the real robots open. In this report we describe a set of practical techniques to achieve the end-to-end result of running a VLA-driven robot at an impressive speed in real world tasks that require both accuracy and dexterity. The stack of technology ranges across calibration, planning & control, and learning based method to identify optimal execution speed. In the tasks we show, the robot even executes in a speed on par with casual human operation and approaching the hardware limit of our lightweight arm. The unaccelerated videos and inference traces are provided in https://dexmal.github.io/realtime-vla-v2/.

cs.RO↗

arXiv2026

Searching for Unparticles with the Cosmic Microwave Background

Multi-field models of inflation typically assume that interactions between particles can be treated perturbatively. Strongly-coupled models provide an intriguing alternative and may offer novel inflationary phenomenology. We study the "unparticle" scenario, where the inflaton is weakly mixed with a strongly-coupled sector, specified by a (gapless) conformal field theory. For certain choices of conformal scaling dimension, $Δ$, the exchange of unparticles leads to distinctive non-Gaussian features in the primordial curvature distribution, including bispectra with enhanced squeezed limits and oscillations close to the equilateral regime. Efficiently analyzing these models using Cosmic Microwave Background (CMB) data is a challenge since the shapes are non-factorizable in momenta and often highly degenerate with single-field self-interactions. Here, we overcome these limitations using a library of tools, including neural-network factorization schemes and optimal CMB estimators. Our pipeline condenses 161 non-separable templates into just 7 factorizable forms, with negligible loss of signal-to-noise. We apply the model to the Planck data, asking two key questions: (1) can we detect unparticles? (2) can we distinguish them from single-field self-interactions? Across $1\leq Δ\leq 9$, we find a maximal signal-to-noise of $1.2σ$, implying no evidence for new physics. We also place the first CMB constraints on the modified consistency-condition-satisfying orthogonal bispectrum with $f^{\rm orth^*}_{\rm NL} = -12\pm12$. While many unparticle models are degenerate with single-field shapes, values of $Δ$ close to half-integers have very different shapes, offering an intriguing future discovery channel. The methods developed herein can be directly applied to other classes of templates, motivating the exploration of models beyond the standard weakly-coupled paradigm.

astro-ph.CO↗

arXiv2026

Text-Image Conditioned 3D Generation

High-quality 3D assets are essential for VR/AR, industrial design, and entertainment, motivating growing interest in generative models that create 3D content from user prompts. Most existing 3D generators, however, rely on a single conditioning modality: image-conditioned models achieve high visual fidelity by exploiting pixel-aligned cues but suffer from viewpoint bias when the input view is limited or ambiguous, while text-conditioned models provide broad semantic guidance yet lack low-level visual detail. This limits how users can express intent and raises a natural question: can these two modalities be combined for more flexible and faithful 3D generation? Our diagnostic study shows that even simple late fusion of text- and image-conditioned predictions outperforms single-modality models, revealing strong cross-modal complementarity. We therefore formalize Text-Image Conditioned 3D Generation, which requires joint reasoning over a visual exemplar and a textual specification. To address this task, we introduce TIGON, a minimalist dual-branch baseline with separate image- and text-conditioned backbones and lightweight cross-modal fusion. Extensive experiments show that text-image conditioning consistently improves over single-modality methods, highlighting complementary vision-language guidance as a promising direction for future 3D generation research. Project page: https://jumpat.github.io/tigon-page

cs.CV↗

arXiv2026

Geometric Imbalance in Semi-Supervised Node Classification

Class imbalance in graph data presents a significant challenge for effective node classification, particularly in semi-supervised scenarios. In this work, we formally introduce the concept of geometric imbalance, which captures how message passing on class-imbalanced graphs leads to geometric ambiguity among minority-class nodes in the riemannian manifold embedding space. We provide a rigorous theoretical analysis of geometric imbalance on the riemannian manifold and propose a unified framework that explicitly mitigates it through pseudo-label alignment, node reordering, and ambiguity filtering. Extensive experiments on diverse benchmarks show that our approach consistently outperforms existing methods, especially under severe class imbalance. Our findings offer new theoretical insights and practical tools for robust semi-supervised node classification.

cs.LG↗

arXiv2026

IBCapsNet: Information Bottleneck Capsule Network for Noise-Robust Representation Learning

Capsule networks (CapsNets) are superior at modeling hierarchical spatial relationships but suffer from two critical limitations: high computational cost due to iterative dynamic routing and poor robustness under input corruptions. To address these issues, we propose IBCapsNet, a novel capsule architecture grounded in the Information Bottleneck (IB) principle. Instead of iterative routing, IBCapsNet employs a one-pass variational aggregation mechanism, where primary capsules are first compressed into a global context representation and then processed by class-specific variational autoencoders (VAEs) to infer latent capsules regularized by the KL divergence. This design enables efficient inference while inherently filtering out noise. Experiments on MNIST, Fashion-MNIST, SVHN and CIFAR-10 show that IBCapsNet matches CapsNet in clean-data accuracy (achieving 99.41% on MNIST and 92.01% on SVHN), yet significantly outperforms it under four types of synthetic noise - demonstrating average improvements of +17.10% and +14.54% for clamped additive and multiplicative noise, respectively. Moreover, IBCapsNet achieves 2.54x faster training and 3.64x higher inference throughput compared to CapsNet, while reducing model parameters by 4.66%. Our work bridges information-theoretic representation learning with capsule networks, offering a principled path toward robust, efficient, and interpretable deep models. Code is available at https://github.com/cxiang26/IBCapsnet

cs.CV↗

arXiv2026

A Survey of Large Language Models

Language is essentially a complex, intricate system of human expressions governed by grammatical rules. It poses a significant challenge to develop capable AI algorithms for comprehending and grasping a language. As a major approach, language modeling has been widely studied for language understanding and generation in the past two decades, evolving from statistical language models to neural language models. Recently, pre-trained language models (PLMs) have been proposed by pre-training Transformer models over large-scale corpora, showing strong capabilities in solving various NLP tasks. Since researchers have found that model scaling can lead to performance improvement, they further study the scaling effect by increasing the model size to an even larger size. Interestingly, when the parameter scale exceeds a certain level, these enlarged language models not only achieve a significant performance improvement but also show some special abilities that are not present in small-scale language models. To discriminate the difference in parameter scale, the research community has coined the term large language models (LLM) for the PLMs of significant size. Recently, the research on LLMs has been largely advanced by both academia and industry, and a remarkable progress is the launch of ChatGPT, which has attracted widespread attention from society. The technical evolution of LLMs has been making an important impact on the entire AI community, which would revolutionize the way how we develop and use AI algorithms. In this survey, we review the recent advances of LLMs by introducing the background, key findings, and mainstream techniques. In particular, we focus on four major aspects of LLMs, namely pre-training, adaptation tuning, utilization, and capacity evaluation. Besides, we also summarize the available resources for developing LLMs and discuss the remaining issues for future directions.

cs.CL↗

arXiv2026

Asymmetric Linear-Combination-of-Unitaries Realization of Quantum Convolution via Modular Adders

Discrete circular convolution over $\mathbb{Z}/N\mathbb{Z}$ is a linear operator and can be implemented on quantum hardware within the linear-combination-of-unitaries (LCU) framework. In this work, we make this connection explicit through an asymmetric-LCU formulation: circular convolution is the postselected block of a circuit whose controlled-shift unitary is modular addition on computational-basis states. The asymmetry is essential: fixing the postselection state to the uniform state $|u\rangle$ while supplying the kernel state $|\mathbf{b}\rangle$ as the input ancilla naturally preserves the complex coefficients $b_i$ within the block, whereas a symmetric overlap would yield $|b_i|^2$ weights and erase their phases. Accordingly, when $|\mathbf{a}\rangle$ and $|\mathbf{b}\rangle$ are supplied by upstream quantum routines, the convolution subroutine requires only the fixed uncompute $\mathrm{PREP}_u^\dagger$, completely avoiding the need for a kernel-dependent inverse preparation $\mathrm{PREP}_b^\dagger$. We then introduce a reversal matrix $J_n=X^{\otimes n}$ and define reflected shifts $\widetilde{L}_{i,n}=L_{i,n}J_n$. This symmetrization yields a recursive operator algebra for convolution that is natively compatible with LCU/block-encoding workflows. The resulting symmetrized operator differs from circular convolution only by one known input-side $J_n$ layer. Crucially, for real-valued kernels, the resulting operator $H_n(\mathbf{b})=\sum_i b_i\widetilde{L}_{i,n}$ is Hermitian, providing a direct Hermitian interface for quantum singular value transformation (QSVT) and related spectral transformations. Based on this framework, we present a transparent recursive construction, paired with an exactly equivalent optimized bitwise compilation of the same $\mathrm{SELECT}$ block. Finally, we evaluate implementation trade-offs and resource scaling under explicit cost-model conventions.

quant-ph↗

arXiv2026

Robust Wildfire Forecasting under Partial Observability: From Reconstruction to Prediction

Satellite-derived fire observations are the primary input for learning-based wildfire spread prediction, yet they are inherently incomplete due to cloud cover, smoke obscuration, and sensor artifacts. This partial observability introduces a domain gap between the clean data used to train forecasting models and the degraded inputs encountered during deployment, often leading to unreliable predictions. To address this challenge, we formulate wildfire forecasting under partial observability using a two-stage probabilistic framework that decouples observation recovery from spatiotemporal prediction. Stage-I reconstructs plausible fire maps from corrupted observations via conditional inpainting, while Stage-II models wildfire dynamics on the recovered sequences using a spatiotemporal forecasting network. We consider four network architectures for the reconstruction module-a Residual U-Net (MaskUNet), a Conditional VAE (MaskCVAE), a cross-attention Vision Transformer (MaskViT), and a discrete diffusion model (MaskD3PM)-spanning CNN-based, latent-variable, attention-based, and diffusion-based approaches. We evaluate the performance of the two-stage approach on the WildfireSpreadTS (WSTS) dataset under various settings, including pixel-wise and block-wise masking, eight corruption levels (10%-80%), four fire scenarios, and leave-one-year-out cross-validation. Results show that all learning-based recovery models substantially outperform non-learning baselines, with MaskCVAE and MaskUNet achieving the strongest overall performance. Importantly, inserting the reconstruction stage before forecasting significantly mitigates the domain gap, restoring next-day prediction accuracy to near-clean-input levels even under severe information loss.

eess.IV↗