arXiv subjects

Yixuan Huang

Publications and source records attributed to Yixuan Huang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Multi-Dimensional Evaluation of LLMs for Grammatical Error Correction

Automated assistants for Grammatical Error Correction are now embedded in educational platforms serving millions of learners, yet three critical gaps remain in this domain: (1) latest-generation Large Language Models (LLMs) lack comprehensive evaluation on grammar correction tasks; (2) whether combining these LLMs improves correction quality is unexplored; and (3) the extent to which reference-based metrics underestimate GEC system performance has not been adequately quantified. In this study, first, we evaluate latest-generation LLMs on edit precision, fluency preservation, and meaning retention, showing fine-tuned GPT-4o achieves state-of-the-art performance across all three dimensions. Second, through grammatical error type analysis we demonstrate that individual LLMs exhibit highly similar error correction patterns ($ρ=0.947$). Third, we show that reference-based metrics underestimate GEC performance with 73.76% of GPT-4o corrections different from gold standards being equally valid or even superior. These GEC evaluation findings equip educators with guidance for selecting GEC assistants that enhance rather than constrain student linguistic development. We make our data, code, and models publicly available.

cs.CL↗

arXiv2026

KinDER: A Physical Reasoning Benchmark for Robot Learning and Planning

Robotic systems that interact with the physical world must reason about kinematic and dynamic constraints imposed by their own embodiment, their environment, and the task at hand. We introduce KinDER, a benchmark for Kinematic and Dynamic Embodied Reasoning that targets physical reasoning challenges arising in robot learning and planning. KinDER comprises 25 procedurally generated environments, a Gymnasium-compatible Python library with parameterized skills and demonstrations, and a standardized evaluation suite with 13 implemented baselines spanning task and motion planning, imitation learning, reinforcement learning, and foundation-model-based approaches. The environments are designed to isolate five core physical reasoning challenges: basic spatial relations, nonprehensile multi-object manipulation, tool use, combinatorial geometric constraints, and dynamic constraints, disentangled from perception, language understanding, and application-specific complexity. Empirical evaluation shows that existing methods struggle to solve many of the environments, indicating substantial gaps in current approaches to physical reasoning. We additionally include real-to-sim-to-real experiments on a mobile manipulator to assess the correspondence between simulation and real-world physical interaction. KinDER is fully open-sourced and intended to enable systematic comparison across diverse paradigms for advancing physical reasoning in robotics. Website and code: https://prpl-group.com/kinder-site/

cs.RO↗

arXiv2026

MMCORE: MultiModal COnnection with Representation Aligned Latent Embeddings

We present MMCORE, a unified framework designed for multimodal image generation and editing. MMCORE leverages a pre-trained Vision-Language Model (VLM) to predict semantic visual embeddings via learnable query tokens, which subsequently serve as conditioning signals for a diffusion model. This streamlined design effectively transfers the rich understanding and reasoning capabilities of VLMs into the visual generation process. By obviating the need for deep fusion between autoregressive and diffusion models or training from scratch, MMCORE significantly reduces computational overhead while maintaining high-fidelity synthesis. MMCORE seamlessly integrates text-to-image synthesis with interleaved image generation, demonstrating robust multimodal comprehension in complex scenarios such as spatial reasoning and visual grounding. Comprehensive evaluations indicate that MMCORE consistently outperforms state-of-the-art baselines across a broad spectrum of text-to-image and single/multi-image editing benchmarks.

cs.CV↗

arXiv2026

Hierarchical DLO Routing with Reinforcement Learning and In-Context Vision-language Models

Long-horizon routing tasks of deformable linear objects (DLOs), such as cables and ropes, are common in industrial assembly lines and everyday life. These tasks are particularly challenging because they require robots to manipulate DLO with long-horizon planning and reliable skill execution. Successfully completing such tasks demands adapting to their nonlinear dynamics, decomposing abstract routing goals, and generating multi-step plans composed of multiple skills, all of which require accurate high-level reasoning during execution. In this paper, we propose a fully autonomous hierarchical framework for solving challenging DLO routing tasks. Given an implicit or explicit routing goal expressed in language, our framework leverages vision-language models~(VLMs) for in-context high-level reasoning to synthesize feasible plans, which are then executed by low-level skills trained via reinforcement learning. To improve robustness over long horizons, we further introduce a failure recovery mechanism that reorients the DLO into insertion-feasible states. Our approach generalizes to diverse scenes involving object attributes, spatial descriptions, implicit language commands, and \myred{extended 5-clip settings}. It achieves an overall success rate of 92\% across long-horizon routing scenarios. Please refer to our project page: https://icra2026-dloroute.github.io/DLORoute/

cs.RO↗

arXiv2026

Interacting ground states of moiré ladders

Moiré materials have emerged as a rich platform for exploring strong correlation effects in low dimensions, with twisted bilayer graphene (TBG) as a paradigmatic example. To distill the essential ingredients driving moiré-induced phases, a simplified one-dimensional analog -- a two-leg ladder with spatially modulated interleg hopping and a uniform magnetic flux -- was recently introduced. This model, which we refer to as the moiré ladder, features a nearly flat lowest-energy band in a suitable parameter regime, capturing the band-flattening mechanism of TBG. We investigate the ground-state phase diagram of the moiré ladder using a combination of bosonization and density matrix renormalization group (DMRG) techniques, and systematically disentangle the respective roles of the flux and the hopping modulation. At half filling, previous numerical work identified a metal-insulator transition at finite interaction strength and an unexpected ferromagnetic ground state. Revisiting this, we show that the metal-insulator transition can be understood perturbatively within bosonization, governed by the number of Fermi points. In contrast, the ferromagnetic correlations are nonperturbative and require both flux and spatial modulation -- neither alone is sufficient. We extend our analysis to other fillings: one-quarter, three-quarters, slightly above half filling (half filling plus two electrons), and slightly below half filling (half filling minus two electrons). At moderate interactions, we observe ferromagnetism below half filling and antiferromagnetism above; at stronger interactions, ferromagnetism dominates across all studied fillings. Crucially, the analysis demonstrates that periodic interleg hopping alone does not engender new correlated phases; the magnetic flux is essential for the observed unconventional behavior.

cond-mat.str-el↗

arXiv2026

RaDAR: Relation-aware Diffusion-Asymmetric Graph Contrastive Learning for Recommendation

Collaborative filtering (CF) recommendation has been significantly advanced by integrating Graph Neural Networks (GNNs) and Graph Contrastive Learning (GCL). However, (i) random edge perturbations often distort critical structural signals and degrade semantic consistency across augmented views, and (ii) data sparsity hampers the propagation of collaborative signals, limiting generalization. To tackle these challenges, we propose RaDAR (Relation-aware Diffusion-Asymmetric Graph Contrastive Learning Framework for Recommendation Systems), a novel framework that combines two complementary view generation mechanisms: a graph generative model to capture global structure and a relation-aware denoising model to refine noisy edges. RaDAR introduces three key innovations: (1) asymmetric contrastive learning with global negative sampling to maintain semantic alignment while suppressing noise; (2) diffusion-guided augmentation, which employs progressive noise injection and denoising for enhanced robustness; and (3) relation-aware edge refinement, dynamically adjusting edge weights based on latent node semantics. Extensive experiments on three public benchmarks demonstrate that RaDAR consistently outperforms state-of-the-art methods, particularly under noisy and sparse conditions.

cs.LG↗

arXiv2026

LAP: Language-Action Pre-Training Enables Zero-shot Cross-Embodiment Transfer

A long-standing goal in robotics is a generalist policy that can be deployed zero-shot on new robot embodiments without per-embodiment adaptation. Despite large-scale multi-embodiment pre-training, existing Vision-Language-Action models (VLAs) remain tightly coupled to their training embodiments and typically require costly fine-tuning. We introduce Language-Action Pre-training (LAP), a simple recipe that represents low-level robot actions directly in natural language, aligning action supervision with the pre-trained vision-language model's input-output distribution. LAP requires no learned tokenizer, no costly annotation, and no embodiment-specific architectural design. Based on LAP, we present LAP-3B, which to the best of our knowledge is the first VLA to achieve substantial zero-shot transfer to previously unseen robot embodiments without any embodiment-specific fine-tuning. Across multiple novel robots and manipulation tasks, LAP-3B attains over 50% average zero-shot success, delivering roughly a 2x improvement over the strongest prior VLAs. We further show that LAP enables efficient adaptation and favorable scaling, while unifying action prediction and VQA in a shared language-action format that yields additional gains through co-training.

cs.RO↗

arXiv2026

Physics-Informed Implicit Neural Representation for Wireless Imaging in RIS-Aided ISAC System

Wireless imaging has become a vital function in future integrated sensing and communication (ISAC) systems. However, traditional model-based and data-driven deep learning imaging methods face challenges related to multipath extraction, dataset acquisition, and multi-scenario adaptation. To overcome these limitations, this study innovatively combines implicit neural representation (INR) with explicit physical models to realize wireless imaging in reconfigurable intelligent surface (RIS)-aided ISAC systems. INR employs neural networks (NNs) to project physical locations to voxel values, which is indirectly supervised by measurements of channel state information with physics-informed loss functions. The continuous shape and scattering characteristics of targets are embedded into NN parameters through training, enabling arbitrary image resolutions and off-grid voxel value prediction. Additionally, three issues related to INR-based imager are further addressed. First, INR is generalized to enable efficient imaging under multipath interference by jointly learning image and multipath information. Second, the imaging speed and accuracy for dynamic targets are enhanced by embedding prior image information. Third, imaging results are employed to assist in RIS phase design for improved communication performance. Extensive simulations demonstrate that the proposed INR-based imager significantly outperforms traditional model-based methods with super-resolution abilities, and the focal length characteristics of the imaging system is revealed. Moreover, communication performance can benefit from the imaging results. Part of the source code for this paper can be accessed at https://github.com/kiwi1944/INRImager

cs.IT↗

arXiv2026

Composable Visual Tokenizers with Generator-Free Diagnostics of Learnability

We introduce CompTok, a training framework for learning visual tokenizers whose tokens are enhanced for compositionality. CompTok uses a token-conditioned diffusion decoder. By employing an InfoGAN-style objective, where we train a recognition model to predict the tokens used to condition the diffusion decoder using the decoded images, we enforce the decoder to not ignore any of the tokens. To promote compositional control, besides the original images, CompTok also trains on tokens formed by swapping token subsets between images, enabling more compositional control of the token over the decoder. As the swapped tokens between images do not have ground truth image targets, we apply a manifold constraint via an adversarial flow regularizer to keep unpaired swap generations on the natural-image distribution. The resulting tokenizer not only achieves state-of-the-art performance on image class-conditioned generation, but also demonstrates properties such as swapping tokens between images to achieve high level semantic editing of an image. Additionally, we propose two metrics that measures the landscape of the token space that can be useful to describe not only the compositionality of the tokens, but also how easy to learn the landscape is for a generator to be trained on this space. We show in experiments that CompTok can improve on both of the metrics as well as supporting state-of-the-art generators for class conditioned generation.

cs.CV↗

arXiv2026

RIS-Aided Cooperative ISAC Network for Imaging-Based Low-Altitude Surveillance

The low-altitude economy is integral to the advancement of numerous sectors, necessitating the development of advanced low-altitude surveillance techniques. Nevertheless, conventional methods encounter limitations of high deployment costs and low signal strength. This study proposes a reconfigurable intelligent surface (RIS)-aided cooperative integrated sensing and communication (ISAC) network for low-altitude surveillance. This network employs RISs to reflect ISAC signals into low-altitude space for sensing. To enhance signal strength, we employ active RIS (ARIS) to amplify the signals. Moreover, in order to avoid error propagation and data association in traditional sensing methods, we model low-altitude surveillance as an imaging problem based on compressed sensing theory, which can be solved through the subspace pursuit algorithm. We derive the Cramer-Rao lower bound (CRLB) of the proposed RIS-aided low-altitude imaging system and analyze the impacts of various system parameters on sensing performance, providing guidance for ISAC system configuration. Numerical results show that ARIS outperforms passive RIS under identical power constraints, achieving effective imaging and target detection at altitudes up to 300 meters.

cs.IT↗

arXiv2025

Multi-Agent Pointer Transformer: Seq-to-Seq Reinforcement Learning for Multi-Vehicle Dynamic Pickup-Delivery Problems

This paper addresses the cooperative Multi-Vehicle Dynamic Pickup and Delivery Problem with Stochastic Requests (MVDPDPSR) and proposes an end-to-end centralized decision-making framework based on sequence-to-sequence, named Multi-Agent Pointer Transformer (MAPT). MVDPDPSR is an extension of the vehicle routing problem and a spatio-temporal system optimization problem, widely applied in scenarios such as on-demand delivery. Classical operations research methods face bottlenecks in computational complexity and time efficiency when handling large-scale dynamic problems. Although existing reinforcement learning methods have achieved some progress, they still encounter several challenges: 1) Independent decoding across multiple vehicles fails to model joint action distributions; 2) The feature extraction network struggles to capture inter-entity relationships; 3) The joint action space is exponentially large. To address these issues, we designed the MAPT framework, which employs a Transformer Encoder to extract entity representations, combines a Transformer Decoder with a Pointer Network to generate joint action sequences in an AutoRegressive manner, and introduces a Relation-Aware Attention module to capture inter-entity relationships. Additionally, we guide the model's decision-making using informative priors to facilitate effective exploration. Experiments on 8 datasets demonstrate that MAPT significantly outperforms existing baseline methods in terms of performance and exhibits substantial computational time advantages compared to classical operations research methods.

cs.LG↗

arXiv2025

Seedream 4.0: Toward Next-generation Multimodal Image Generation

We introduce Seedream 4.0, an efficient and high-performance multimodal image generation system that unifies text-to-image (T2I) synthesis, image editing, and multi-image composition within a single framework. We develop a highly efficient diffusion transformer with a powerful VAE which also can reduce the number of image tokens considerably. This allows for efficient training of our model, and enables it to fast generate native high-resolution images (e.g., 1K-4K). Seedream 4.0 is pretrained on billions of text-image pairs spanning diverse taxonomies and knowledge-centric concepts. Comprehensive data collection across hundreds of vertical scenarios, coupled with optimized strategies, ensures stable and large-scale training, with strong generalization. By incorporating a carefully fine-tuned VLM model, we perform multi-modal post-training for training both T2I and image editing tasks jointly. For inference acceleration, we integrate adversarial distillation, distribution matching, and quantization, as well as speculative decoding. It achieves an inference time of up to 1.8 seconds for generating a 2K image (without a LLM/VLM as PE model). Comprehensive evaluations reveal that Seedream 4.0 can achieve state-of-the-art results on both T2I and multimodal image editing. In particular, it demonstrates exceptional multimodal capabilities in complex tasks, including precise image editing and in-context reasoning, and also allows for multi-image reference, and can generate multiple output images. This extends traditional T2I systems into an more interactive and multidimensional creative tool, pushing the boundary of generative AI for both creativity and professional applications. We further scale our model and data as Seedream 4.5. Seedream 4.0 and Seedream 4.5 are accessible on Volcano Engine https://www.volcengine.com/experience/ark?launch=seedream.

cs.CV↗

arXiv2025

Classification of autoimmune diseases from Peripheral blood TCR repertoires by multimodal multi-instance learning

T cell receptor (TCR) repertoires encode critical immunological signatures for autoimmune diseases, yet their clinical application remains limited by sequence sparsity and low witness rates. We developed EAMil, a multi-instance deep learning framework that leverages TCR sequencing data to diagnose systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) with exceptional accuracy. By integrating PrimeSeq feature extraction with ESMonehot encoding and enhanced gate attention mechanisms, our model achieved state-of-the-art performance with AUCs of 98.95% for SLE and 97.76% for RA. EAMil successfully identified disease-associated genes with over 90% concordance with established differential analyses and effectively distinguished disease-specific TCR genes. The model demonstrated robustness in classifying multiple disease categories, utilizing the SLEDAI score to stratify SLE patients by disease severity as well as to diagnose the site of damage in SLE patients, and effectively controlling for confounding factors such as age and gender. This interpretable framework for immune receptor analysis provides new insights for autoimmune disease detection and classification with broad potential clinical applications across immune-mediated conditions.

cs.LG↗

arXiv2025

Bipartite holes, degree sums and Hamilton cycles

The {\em bipartite-hole-number} of a graph $G$, denoted as $\widetildeα(G)$, is the minimum number $k$ such that there exist integers $a$ and $b$ with $a + b = k+1$ such that for any two disjoint sets $A, B \subseteq V(G)$, there is an edge between $A$ and $B$. McDiarmid and Yolov initiated research on bipartite holes by extending Dirac's classical theorem on minimum degree and Hamiltonian cycles. They showed that a graph on at least three vertices with $δ(G) \ge \widetildeα(G)$ is Hamiltonian. Later, Draganić, Munhá Correia and Sudakov proved that $δ\ge \widetildeα(G)$ implies that $G$ is pancyclic, unless $G = K_{\frac n2, \frac n2}$. This extended the result of McDiarmid and Yolov and generalized a theorem of Bondy on pancyclicity. In this paper, we show that a $2$-connected graph $G$ is Hamiltonian if $σ_2(G) \ge 2 \widetildeα(G) - 1$, and that a connected graph $G$ contains a cycle through all vertices of degree at least $\widetildeα(G)$. Both results extended McDiarmid and Yolov's result. As a step toward proving pancyclicity, we show that if an $n$-vertex graph $G$ satisfies $σ_2(G) \ge 2 \widetildeα(G) - 1$, then it either contains a triangle or it is $K_{\frac n2, \frac n2}$. Finally, we discuss the relationship between connectivity and the bipartite hole number.

math.CO↗

arXiv2025

Reconstruction of C_4-free graphs from the set of closed neighborhoods and digital convexity

Fomin, Kratochvíl, Lokshtanov, Mancini, and Telle showed that every $C_{4}$-free graph is reconstructible from the \emph{multiset} of closed neighborhoods. We strengthen their result proving that every $C_{4}$-free graph is reconstructible from the \emph{set} of closed neighborhoods. This extends the work of Lafrance et al.\ by showing that all $C_{4}$-free graphs, and hence all graphs of girth at least five, are reconstructible from their digitally convex sets. A subset $S$ of vertices in a graph $G$ is digitally convex if, for every vertex $v \notin S$, there is a private neighbor of $v$. We establish that reconstruction from digitally convex sets is equivalent to reconstruction from the set of closed neighborhoods.

math.CO↗

arXiv2025

Spin Seebeck Effect of Triangular-lattice Spin Supersolid

Using thermal tensor-network approach, we investigate the spin Seebeck effect (SSE) of the triangular-lattice quantum antiferromagnet hosting spin supersolid phase. We focus on the low-temperature scaling behaviors of the normalized spin current across the interface. For the 1D Heisenberg chain, we find a negative spinon spin in the bulk current with algebraic temperature scaling; at low fields, boundary effects induce a second sign reversal at lower temperatures. These benchmark results are consistent with field-theoretical analysis. On the triangular lattice, spin frustration dramatically enhances the low-temperature SSE, with distinct spin-current signatures -- particularly the sign reversal and characteristic temperature dependence -- distinguishing different spin states. Remarkably, we discover a persistent, negative spin current in the spin supersolid phase, which saturates to a non-zero value in the low-temperature limit and can be ascribed to the Goldstone-mode-mediated spin supercurrents. Moreover, a universal scaling $T^{d/z}$ is found at the U(1)-symmetric polarization quantum critical points. These distinct quantum spin transport traits provide sensitive spin current probes for spin supersolid states in quantum magnets such as Na$_2$BaCo(PO$_4$)$_2$. Furthermore, our results also establish spin supersolids as a tunable quantum platform for spin caloritronics in the ultralow-temperature regime.

cond-mat.str-el↗

arXiv2025

Fail2Progress: Learning from Real-World Robot Failures with Stein Variational Inference

Skill effect models for long-horizon manipulation tasks are prone to failures in conditions not covered by training data distributions. Therefore, enabling robots to reason about and learn from failures is necessary. We investigate the problem of efficiently generating a dataset targeted to observed failures. After fine-tuning a skill effect model on this dataset, we evaluate the extent to which the model can recover from failures and minimize future failures. We propose Fail2Progress, an approach that leverages Stein variational inference to generate multiple simulation environments in parallel, enabling efficient data sample generation similar to observed failures. Our method is capable of handling several challenging mobile manipulation tasks, including transporting multiple objects, organizing a constrained shelf, and tabletop organization. Through large-scale simulation and real-world experiments, we demonstrate that our approach excels at learning from failures across different numbers of objects. Furthermore, we show that Fail2Progress outperforms several baselines.

cs.RO↗

arXiv2025

Magnetic anisotropy effect on stabilizing magnetization plateaus of kagome strip chain Heisenberg antiferromagnets

We investigate the anisotropic effect of magnetization plateaus in the antiferromagnetic Heisenberg model on a kagome strip chain. The kagome strip chain Heisenberg model, composed of a hexagonal net of triangles forming five-site unit cells, exhibits four magnetization plateaus in the presence of an applied magnetic field. Using numerical density matrix renormalization group method, we find that the magnetization plateaus are stable against anisotropic interactions in the same direction of the applied magnetic field, but the plateaus become much smaller with anisotropic interactions in other directions. We further show the anisotropic effect of the magnon excitations of the 0.6 plateau state using linear spin wave theory. The magnon bandwidth remains small when tuning the anisotropic interactions along the field where the magnetization plateau is stable, while the band becomes more dispersive with anisotropic interactions perpendicular to the field. In addition, upon tuning down the interaction strength for the two lower legs below a critical value, the Hamiltonian of the kagome strip chain is dominated by two separate spin chains. This can be used to determine the effective lattice structure in materials with strong distortions. Our results enhance the theoretical understanding of the anisotropic effect and the nature of magnetization plateaus in kagome strip chain materials, which can contribute to the design and manipulation of kagome materials with tailored properties.

cond-mat.str-el↗