arXiv subjects

Yu Wang

Publications and source records attributed to Yu Wang.

At least 433 records · Page 24Linked to original sources

arXiv2025

VocalBench-zh: Decomposing and Benchmarking the Speech Conversational Abilities in Mandarin Context

The development of multi-modal large language models (LLMs) leads to intelligent approaches capable of speech interactions. As one of the most widely spoken languages globally, Mandarin is supported by most models to enhance their applicability and reach. However, the scarcity of comprehensive speech-to-speech (S2S) benchmarks in Mandarin contexts impedes systematic evaluation for developers and hinders fair model comparison for users. In this work, we propose VocalBench-zh, an ability-level divided evaluation suite adapted to Mandarin context consisting of 10 well-crafted subsets and over 10K high-quality instances, covering 12 user-oriented characters. The evaluation experiment on 14 mainstream models reveals the common challenges for current routes, and highlights the need for new insights into next-generation speech interactive systems. The evaluation codes and datasets will be available at https://github.com/SJTU-OmniAgent/VocalBench-zh.

cs.CL↗

arXiv2025

Probing critical phenomena in open quantum systems using atom arrays

At continuous phase transitions, quantum many-body systems exhibit scale-invariance and complex, emergent universal behavior. Most strikingly, at a quantum critical point, correlations decay as a power law, with exponents determined by a set of universal scaling dimensions. Experimentally probing such power-law correlations is extremely challenging, owing to the complex interplay between decoherence, the vanishing energy gap, and boundary effects. Here, we employ a Rydberg quantum simulator to adiabatically prepare critical ground states of both a one-dimensional ring and a two-dimensional square lattice. By accounting for and tuning the openness of our quantum system, which is well-captured by the introduction of a single phenomenological length scale, we are able to directly observe power-law correlations and extract the corresponding scaling dimensions. Moreover, in two dimensions, we observe a decoupling between phase transitions in the bulk and on the boundary, allowing us to identify two distinct boundary universality classes. Our work demonstrates that direct adiabatic preparation of critical states in quantum simulators can complement recent approaches to studying quantum criticality using the Kibble-Zurek mechanism or digital quantum circuits.

quant-ph↗

arXiv2025

VocalNet-M2: Advancing Low-Latency Spoken Language Modeling via Integrated Multi-Codebook Tokenization and Multi-Token Prediction

Current end-to-end spoken language models (SLMs) have made notable progress, yet they still encounter considerable response latency. This delay primarily arises from the autoregressive generation of speech tokens and the reliance on complex flow-matching models for speech synthesis. To overcome this, we introduce VocalNet-M2, a novel low-latency SLM that integrates a multi-codebook tokenizer and a multi-token prediction (MTP) strategy. Our model directly generates multi-codebook speech tokens, thus eliminating the need for a latency-inducing flow-matching model. Furthermore, our MTP strategy enhances generation efficiency and improves overall performance. Extensive experiments demonstrate that VocalNet-M2 achieves a substantial reduction in first chunk latency (from approximately 725ms to 350ms) while maintaining competitive performance across mainstream SLMs. This work also provides a comprehensive comparison of single-codebook and multi-codebook strategies, offering valuable insights for developing efficient and high-performance SLMs for real-time interactive applications.

cs.CL↗

arXiv2025

Low-Energy Free-Electron Nonclassical Lasing

Harnessing a beam of slow free electrons in artificial photonic structures offers a powerful, tunable platform for generating nonclassical light without the need for heavy physical equipment. Here we present a theory of nonclassical lasing, demonstrating how incoherent electrons in photonic crystal cavities can coherently emit photons through collective dynamics. When photon emission rate exceeds cavity losses, nonclassical lasing with sub-Poissonian photon statistics emerges, driven by multi-photon Rabi oscillations. At specific coupling strengths, quantum state trapping effect emerges, producing high-fidelity Fock states at room temperature (e.g. nearly 90%-fidelity of four photon Fock state). Notably, the frequency of the emitted photons can be readily tuned via the velocity of the injected electrons to match cavity modes. This approach supports photonic integration and offers a scalable, energy-efficient platform for room-temperature quantum light sources and advanced studies in quantum electrodynamics.

physics.optics↗

arXiv2025

Evaluating Cell AI Foundation Models in Kidney Pathology with Human-in-the-Loop Enrichment

Training AI foundation models has emerged as a promising large-scale learning approach for addressing real-world healthcare challenges, including digital pathology. While many of these models have been developed for tasks like disease diagnosis and tissue quantification using extensive and diverse training datasets, their readiness for deployment on some arguably simplest tasks, such as nuclei segmentation within a single organ (e.g., the kidney), remains uncertain. This paper seeks to answer this key question, "How good are we?", by thoroughly evaluating the performance of recent cell foundation models on a curated multi-center, multi-disease, and multi-species external testing dataset. Additionally, we tackle a more challenging question, "How can we improve?", by developing and assessing human-in-the-loop data enrichment strategies aimed at enhancing model performance while minimizing the reliance on pixel-level human annotation. To address the first question, we curated a multicenter, multidisease, and multispecies dataset consisting of 2,542 kidney whole slide images (WSIs). Three state-of-the-art (SOTA) cell foundation models-Cellpose, StarDist, and CellViT-were selected for evaluation. To tackle the second question, we explored data enrichment algorithms by distilling predictions from the different foundation models with a human-in-the-loop framework, aiming to further enhance foundation model performance with minimal human efforts. Our experimental results showed that all three foundation models improved over their baselines with model fine-tuning with enriched data. Interestingly, the baseline model with the highest F1 score does not yield the best segmentation outcomes after fine-tuning. This study establishes a benchmark for the development and deployment of cell vision foundation models tailored for real-world data applications.

cs.CV↗

arXiv2025

Strain-engineered nanoscale spin polarization reversal in diamond nitrogen-vacancy centers

The ability to control solid-state quantum emitters is fundamental to advancing quantum technologies. The performance of these systems is fundamentally governed by their spin-dependent photodynamics, yet conventional control methods using cavities offer limited access to key non-radiative processes. Here we demonstrate that anisotropic lattice strain serves as a powerful tool for manipulating spin dynamics in solid-state systems. Under high pressure, giant shear strain gradients trigger a complete reversal of the intrinsic spin polarization, redirecting ground-state population from $|0\rangle$ to $|\pm 1\rangle$ manifold. We show that this reprogramming arises from strain-induced mixing of the NV center's excited states and dramatic alteration of intersystem crossing, which we quantify through a combination of opto-magnetic spectroscopy and a theoretical model that disentangles symmetry-preserving and symmetry-breaking strain contributions. Furthermore, the polarization reversal is spatially mapped with a transition region below 120 nm, illustrating sub-diffraction-limit control. Our work establishes strain engineering as a powerful tool for tailoring quantum emitter properties, opening avenues for programmable quantum light sources, high-density spin-based memory, and hybrid quantum photonic devices.

quant-ph↗

arXiv2025

R2R: Efficiently Navigating Divergent Reasoning Paths with Small-Large Model Token Routing

Large Language Models (LLMs) achieve impressive reasoning capabilities at the cost of substantial inference overhead, posing substantial deployment challenges. Although distilled Small Language Models (SLMs) significantly enhance efficiency, their performance suffers as they fail to follow LLMs' reasoning paths. Luckily, we reveal that only a small fraction of tokens genuinely diverge reasoning paths between LLMs and SLMs. Most generated tokens are either identical or exhibit neutral differences, such as minor variations in abbreviations or expressions. Leveraging this insight, we introduce **Roads to Rome (R2R)**, a neural token routing method that selectively utilizes LLMs only for these critical, path-divergent tokens, while leaving the majority of token generation to the SLM. We also develop an automatic data generation pipeline that identifies divergent tokens and generates token-level routing labels to train the lightweight router. We apply R2R to combine R1-1.5B and R1-32B models from the DeepSeek family, and evaluate on challenging math, coding, and QA benchmarks. With an average activated parameter size of 5.6B, R2R surpasses the average accuracy of R1-7B by 1.6x, outperforming even the R1-14B model. Compared to R1-32B, it delivers a 2.8x wall-clock speedup with comparable performance, advancing the Pareto frontier of test-time scaling efficiency. Our code is available at https://github.com/thu-nics/R2R.

cs.CL↗

arXiv2025

Topological transition and emergent elasticity of dislocation in skyrmion lattice: Beyond Kittel's magnetic-polar analogy

Magnetic and polar skyrmions exhibit topologically protected quasiparticle behavior, including emergent fields, deformation, and the formation of a densely packed skyrmion lattice, beyond conventional domain configurations described by Kittel's law. Analogous to atomic crystals, lattice defects, especially dislocations and their associated strain fields, are crucial for understanding the lattice behavior of skyrmions; however, their features and roles remain insufficiently understood. Here, we show that magnetic skyrmion dislocations develop a core-split structure due to a significant skyrmion elongation up to 180% of their original length, reaching a topological transition from a single skyrmion to two half-skyrmions. Despite such a distinct structure, the long-range strain fields around the dislocation perfectly obey conventional Volterra's elasticity theory, in contrast to polar skyrmion lattices, where skyrmion deformations cause a breakdown of the elasticity theory. Furthermore, an energetic analysis shows that Dzyaloshinskii-Moriya interaction drives the large skyrmion deformation of the dislocation core. Our findings not only clarify the coexistence of topological core-reconstruction and a robust long-range elastic field of dislocations in magnetic skyrmion lattices, but also reveal that magnetic and electric domains, long regarded as dual and analogous, exhibit fundamental differences when extended into the regime of collective topological quasiparticles.

cond-mat.mtrl-sci↗

arXiv2025

FedMGP: Personalized Federated Learning with Multi-Group Text-Visual Prompts

In this paper, we introduce FedMGP, a new paradigm for personalized federated prompt learning in vision-language models. FedMGP equips each client with multiple groups of paired textual and visual prompts, enabling the model to capture diverse, fine-grained semantic and instance-level cues. A diversity loss is introduced to drive each prompt group to specialize in distinct and complementary semantic aspects, ensuring that the groups collectively cover a broader range of local characteristics. During communication, FedMGP employs a dynamic prompt aggregation strategy based on similarity-guided probabilistic sampling: each client computes the cosine similarity between its prompt groups and the global prompts from the previous round, then samples s groups via a softmax-weighted distribution. This soft selection mechanism preferentially aggregates semantically aligned knowledge while still enabling exploration of underrepresented patterns effectively balancing the preservation of common knowledge with client-specific features. Notably, FedMGP maintains parameter efficiency by redistributing a fixed prompt capacity across multiple groups, achieving state-of-the-art performance with the lowest communication parameters among all federated prompt learning methods. Theoretical analysis shows that our dynamic aggregation strategy promotes robust global representation learning by reinforcing shared semantics while suppressing client-specific noise. Extensive experiments demonstrate that FedMGP consistently outperforms prior approaches in both personalization and domain generalization across diverse federated vision-language benchmarks. The code will be released on https://github.com/weihao-bo/FedMGP.git.

cs.CV↗

arXiv2025

SC-LoRA: Balancing Efficient Fine-tuning and Knowledge Preservation via Subspace-Constrained LoRA

Parameter-Efficient Fine-Tuning (PEFT) methods, particularly Low-Rank Adaptation (LoRA), are indispensable for efficiently customizing Large Language Models (LLMs). However, vanilla LoRA suffers from slow convergence speed and knowledge forgetting problems. Recent studies have leveraged the power of designed LoRA initialization, to enhance the fine-tuning efficiency, or to preserve knowledge in the pre-trained LLM. However, none of these works can address the two cases at the same time. To this end, we introduce Subspace-Constrained LoRA (SC-LoRA), a novel LoRA initialization framework engineered to navigate the trade-off between efficient fine-tuning and knowledge preservation. We achieve this by constraining the output of trainable LoRA adapters in a low-rank subspace, where the context information of fine-tuning data is most preserved while the context information of preserved knowledge is least retained, in a balanced way. Such constraint enables the trainable weights to primarily focus on the main features of fine-tuning data while avoiding damaging the preserved knowledge features. We provide theoretical analysis on our method, and conduct extensive experiments including safety preservation and world knowledge preservation, on various downstream tasks. In our experiments, SC-LoRA succeeds in delivering superior fine-tuning performance while markedly diminishing knowledge forgetting, surpassing contemporary LoRA initialization methods.

cs.LG↗

arXiv2025

PartnerMAS: An LLM Hierarchical Multi-Agent Framework for Business Partner Selection on High-Dimensional Features

High-dimensional decision-making tasks, such as business partner selection, involve evaluating large candidate pools with heterogeneous numerical, categorical, and textual features. While large language models (LLMs) offer strong in-context reasoning capabilities, single-agent or debate-style systems often struggle with scalability and consistency in such settings. We propose PartnerMAS, a hierarchical multi-agent framework that decomposes evaluation into three layers: a Planner Agent that designs strategies, Specialized Agents that perform role-specific assessments, and a Supervisor Agent that integrates their outputs. To support systematic evaluation, we also introduce a curated benchmark dataset of venture capital co-investments, featuring diverse firm attributes and ground-truth syndicates. Across 140 cases, PartnerMAS consistently outperforms single-agent and debate-based multi-agent baselines, achieving up to 10--15\% higher match rates. Analysis of agent reasoning shows that planners are most responsive to domain-informed prompts, specialists produce complementary feature coverage, and supervisors play an important role in aggregation. Our findings demonstrate that structured collaboration among LLM agents can generate more robust outcomes than scaling individual models, highlighting PartnerMAS as a promising framework for high-dimensional decision-making in data-rich domains.

cs.MA↗

arXiv2025

How Close Are We? Limitations and Progress of AI Models in Banff Lesion Scoring

The Banff Classification provides the global standard for evaluating renal transplant biopsies, yet its semi-quantitative nature, complex criteria, and inter-observer variability present significant challenges for computational replication. In this study, we explore the feasibility of approximating Banff lesion scores using existing deep learning models through a modular, rule-based framework. We decompose each Banff indicator - such as glomerulitis (g), peritubular capillaritis (ptc), and intimal arteritis (v) - into its constituent structural and inflammatory components, and assess whether current segmentation and detection tools can support their computation. Model outputs are mapped to Banff scores using heuristic rules aligned with expert guidelines, and evaluated against expert-annotated ground truths. Our findings highlight both partial successes and critical failure modes, including structural omission, hallucination, and detection ambiguity. Even when final scores match expert annotations, inconsistencies in intermediate representations often undermine interpretability. These results reveal the limitations of current AI pipelines in replicating computational expert-level grading, and emphasize the importance of modular evaluation and computational Banff grading standard in guiding future model development for transplant pathology.

cs.CV↗

arXiv2025

Dust and Water in V883 Ori: Relics of a Retreating Snowline

V883 Ori is an FU-Orionis-type outburst system characterized by a shoulder at 50-70 au in its ALMA band 6 and 7 intensity profiles. Previously, this feature was attributed to dust pile-up from pebble disintegration at the water snowline. However, recent multi-wavelength observations show continuity in the spectral index across the expected snowline region, disfavoring abrupt changes in grain properties. Moreover, extended water emission is detected beyond 80 au, pointing to a snowline further out. This Letter aims to explain both features with a model in which the snowline is receding. We construct a 2D disk model that solves the cooling and subsequent vapor recondensation during the post-outburst dimming phase. Our results show that both the intensity shoulder and the extended water emission are natural relics of a retreating snowline: the shoulder arises from excess surface density generated by vapor recondensation at the moving condensation front, while the outer water vapor reservoir persists due to the long recondensation timescales of $10^{2}-10^{3}$ yr at the disk atmosphere. As V883 Ori continues to fade, we predict that the intensity shoulder will migrate inward by an observationally significant amount of 10 au over about 25 years.

astro-ph.EP↗

arXiv2025

Quantum Advantage via Efficient Post-processing on Qudit Classical Shadow tomography

The computation of $\operatorname{tr}(AB)$ is essential in quantum science and artificial intelligence, yet classical methods for $ d $-dimensional matrices $ A $ and $ B $ require $ O(d^2) $ complexity, which becomes infeasible for exponentially large systems. We introduce a quantum approach based on qudit shadow tomography that reduces both computational and storage complexities to $ O(\text{poly}(\log d)) $ in specific cases. The proposed method applies to quantum density matrices $ A $ and Hermitian matrices $ B $ with given $\operatorname{tr}(B)$ and $\operatorname{tr}(B^2)$ bounded by a constant (referred to as BN-observables). It guarantees at least a quadratic speedup ($O(d^2) \to O(d)$) in the worst case and achieves exponential speedup for approximately average cases. For any $ n $-qubit stabilizer state $ρ$ and arbitrary BN-observable $ O $, we show that $\operatorname{tr}(ρO)$ can be efficiently estimated with $\text{poly}(n)$ computations. Moreover, our approach significantly reduces the post-processing complexity in shadow tomography using random Clifford measurements, and it is applicable to arbitrary dimensions $ d $. These advances open new avenues for efficient high-dimensional data analysis and modeling.

quant-ph↗

arXiv2025

A light-induced charge order mode in a metastable cuprate ladder

We report the observation of an emergent charge order mode in the optically-excited cuprate ladder Sr$_{14}$Cu$_{24}$O$_{41}$. Near-infrared light in the ladder plane drives a symmetry-protected electronic metastable state together with a partial melting of the equilibrium charge order. Our time-resolved resonant inelastic x-ray scattering measurements at the upper Hubbard band reveal a gapless collective excitation dispersing from the charge-order wavevector up to 0.8 eV with a slope on the order of the quasiparticle velocity. These findings reveal a regime where correlated carriers acquire itinerant character at finite momentum, and charge order becomes dynamically fluctuating, offering a platform to explore light-induced pairing instabilities.

cond-mat.str-el↗

arXiv2025

Learning the PTM Code through a Coarse-to-Fine, Mechanism-Aware Framework

Post-translational modifications (PTMs) form a combinatorial "code" that regulates protein function, yet deciphering this code - linking modified sites to their catalytic enzymes - remains a central unsolved problem in understanding cellular signaling and disease. We introduce COMPASS-PTM, a mechanism-aware, coarse-to-fine learning framework that unifies residue-level PTM profiling with enzyme-substrate assignment. COMPASS-PTM integrates evolutionary representations from protein language models with physicochemical priors and a crosstalk-aware prompting mechanism that explicitly models inter-PTM dependencies. This design allows the model to learn biologically coherent patterns of cooperative and antagonistic modifications while addressing the dual long-tail distribution of PTM data. Across multiple proteome-scale benchmarks, COMPASS-PTM establishes new state-of-the-art performance, including a 122% relative F1 improvement in multi-label site prediction and a 54% gain in zero-shot enzyme assignment. Beyond accuracy, the model demonstrates interpretable generalization, recovering canonical kinase motifs and predicting disease-associated PTM rewiring caused by missense variants. By bridging statistical learning with biochemical mechanism, COMPASS-PTM unifies site-level and enzyme-level prediction into a single framework that learns the grammar underlying protein regulation and signaling.

cs.CE↗

arXiv2025

Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents

We present Game-TARS, a generalist game agent trained with a unified, scalable action space anchored to human-aligned native keyboard-mouse inputs. Unlike API- or GUI-based approaches, this paradigm enables large-scale continual pre-training across heterogeneous domains, including OS, web, and simulation games. Game-TARS is pre-trained on over 500B tokens with diverse trajectories and multimodal data. Key techniques include a decaying continual loss to reduce causal confusion and an efficient Sparse-Thinking strategy that balances reasoning depth and inference cost. Experiments show that Game-TARS achieves about 2 times the success rate over the previous sota model on open-world Minecraft tasks, is close to the generality of fresh humans in unseen web 3d games, and outperforms GPT-5, Gemini-2.5-Pro, and Claude-4-Sonnet in FPS benchmarks. Scaling results on training-time and test-time confirm that the unified action space sustains improvements when scaled to cross-game and multimodal data. Our results demonstrate that simple, scalable action representations combined with large-scale pre-training provide a promising path toward generalist agents with broad computer-use abilities.

cs.AI↗

arXiv2025

Distilled Decoding 1: One-step Sampling of Image Auto-regressive Models with Flow Matching

Autoregressive (AR) models have achieved state-of-the-art performance in text and image generation but suffer from slow generation due to the token-by-token process. We ask an ambitious question: can a pre-trained AR model be adapted to generate outputs in just one or two steps? If successful, this would significantly advance the development and deployment of AR models. We notice that existing works that try to speed up AR generation by generating multiple tokens at once fundamentally cannot capture the output distribution due to the conditional dependencies between tokens, limiting their effectiveness for few-step generation. To address this, we propose Distilled Decoding (DD), which uses flow matching to create a deterministic mapping from Gaussian distribution to the output distribution of the pre-trained AR model. We then train a network to distill this mapping, enabling few-step generation. DD doesn't need the training data of the original AR model, making it more practical. We evaluate DD on state-of-the-art image AR models and present promising results on ImageNet-256. For VAR, which requires 10-step generation, DD enables one-step generation (6.3$\times$ speed-up), with an acceptable increase in FID from 4.19 to 9.96. For LlamaGen, DD reduces generation from 256 steps to 1, achieving an 217.8$\times$ speed-up with a comparable FID increase from 4.11 to 11.35. In both cases, baseline methods completely fail with FID>100. DD also excels on text-to-image generation, reducing the generation from 256 steps to 2 for LlamaGen with minimal FID increase from 25.70 to 28.95. As the first work to demonstrate the possibility of one-step generation for image AR models, DD challenges the prevailing notion that AR models are inherently slow, and opens up new opportunities for efficient AR generation. The project website is at https://imagination-research.github.io/distilled-decoding.

cs.CV↗