arXiv subjects

Yu Zhang

Publications and source records attributed to Yu Zhang.

At least 523 records · Page 29Linked to original sources

arXiv2025

Scalable MatMul-free Language Modeling

Large Language Models (LLMs) have fundamentally altered how we approach scaling in machine learning. However, these models pose substantial computational and memory challenges, primarily due to the reliance on matrix multiplication (MatMul) within their attention and feed-forward (FFN) layers. We demonstrate that MatMul operations can be eliminated from LLMs while maintaining strong performance, even at billion-parameter scales. Our MatMul-free models, tested on models up to 2.7B parameters, are comparable to state-of-the-art pre-trained Transformers, and the performance gap narrows as model size increases. Our approach yields significant memory savings: a GPU-efficient implementation reduces memory consumption by up to 61% during training and over 10x during inference. When adapted for a multi-chip neuromorphic system, the model leverages asynchronous processing to achieve 4x higher throughput with 10x less energy than edge GPUs.

cs.CL↗

arXiv2025

A Two-armed Bandit Framework for A/B Testing

A/B testing is widely used in modern technology companies for policy evaluation and product deployment, with the goal of comparing the outcomes under a newly-developed policy against a standard control. Various causal inference and reinforcement learning methods developed in the literature are applicable to A/B testing. This paper introduces a two-armed bandit framework designed to improve the power of existing approaches. The proposed procedure consists of three main steps: (i) employing doubly robust estimation to generate pseudo-outcomes, (ii) utilizing a two-armed bandit framework to construct the test statistic, and (iii) applying a permutation-based method to compute the $p$-value. We demonstrate the efficacy of the proposed method through asymptotic theories, numerical experiments and real-world data from a ridesharing company, showing its superior performance in comparison to existing methods.

stat.ML↗

arXiv2025

ProactiveVA: Proactive Visual Analytics with LLM-Based UI Agent

Visual analytics (VA) is typically applied to complex data, thus requiring complex tools. While visual analytics empowers analysts in data analysis, analysts may get lost in the complexity occasionally. This highlights the need for intelligent assistance mechanisms. However, even the latest LLM-assisted VA systems only provide help when explicitly requested by the user, making them insufficiently intelligent to offer suggestions when analysts need them the most. We propose a ProactiveVA framework in which LLM-powered UI agent monitors user interactions and delivers context-aware assistance proactively. To design effective proactive assistance, we first conducted a formative study analyzing help-seeking behaviors in user interaction logs, identifying when users need proactive help, what assistance they require, and how the agent should intervene. Based on this analysis, we distilled key design requirements in terms of intent recognition, solution generation, interpretability and controllability. Guided by these requirements, we develop a three-stage UI agent pipeline including perception, reasoning, and acting. The agent autonomously perceives users' needs from VA interaction logs, providing tailored suggestions and intuitive guidance through interactive exploration of the system. We implemented the framework in two representative types of VA systems, demonstrating its generalizability, and evaluated the effectiveness through an algorithm evaluation, case and expert study and a user study. We also discuss current design trade-offs of proactive VA and areas for further exploration.

cs.HC↗

arXiv2025

Interaction of Magnetic Fields with Spinons in a Fractionalized State

The 4d-electron trimer lattice exhibits either a quantum spin liquid (QSL) or a heavy-fermion strange metal (HFSM) phase, depending on Nb content. In the QSL state, itinerant spinons act as effective heat carriers, enhancing thermal conductivity. Strikingly, applying a magnetic field up to 14 T causes an abrupt increase in heat capacity by as much as 5000% below 150 mK, disrupting the linear temperature dependence characteristic of both phases. Meanwhile, AC magnetic susceptibility and electrical resistivity remain nearly unchanged, while thermal conductivity is suppressed by up to 40% below 4 K. These observations suggest that spinons, though charge-neutral, are highly sensitive to magnetic fields at low temperatures. We propose that the field induces Anderson localization of spinons, leading to emergent non-magnetic two-level systems that account for the rapid rise in heat capacity. These findings uncover a previously unexplored regime of spinon dynamics, governed by field-induced localization and distinct from conventional magnetic or transport signatures. Comments: To appear in npj Quantum Materials.

cond-mat.str-el↗

arXiv2025

Non-Orthogonal AFDM: A Promising Spectrum-Efficient Waveform for 6G High-Mobility Communications

This paper proposes a spectrum-efficient nonorthogonal affine frequency division multiplexing (AFDM) waveform for reliable high-mobility communications in the upcoming sixth-generation (6G) mobile systems. Our core idea is to introduce a compression factor to enable controllable subcarrier overlapping in chirp-based AFDM modulation. To mitigate intercarrier interference (ICI), we introduce linear precoding at the transmitter and an iterative detection scheme at the receiver. Simulation results demonstrate that these techniques can effectively reduce interference and maintain robust bit error rate (BER) performance even under aggressive compression factors and high-mobility channel conditions. The proposed non-orthogonal AFDM waveform offers a promising solution for next-generation wireless networks, balancing spectrum efficiency and Doppler resilience in highly dynamic environments.

eess.SP↗

arXiv2025

Tensorial Spin-Phonon Relaxation Reveals Mode-Selective Relaxation Pathways in a Single-Molecule Magnet

Understanding and controlling spin relaxation in molecular qubits is essential for developing chemically tunable quantum information platforms. We present a fully first-principles framework for computing the spin relaxation tensor in a single-molecule magnet, \ce{VOPc(OH)8}, by combining density functional theory with a mode-resolved open-system formalism. By expanding the spin Hamiltonian in vibrational normal modes and evaluating both linear and quadratic spin-phonon coupling tensors via finite differences of the $g$-tensor, we construct a relaxation tensor that enters a Lindblad-type quantum master equation. Our formalism captures both direct (one-phonon) and resonant-Raman (two-phonon) relaxation processes. Numerical analysis reveals a highly mode-selective structure: only three vibrational modes dominate longitudinal ($T_1$) decoherence, while a single mode accounts for the majority of transverse ($T_2$) relaxation. The computed relaxation times show excellent agreement with experimental measurements, without any empirical fitting. These results demonstrate that first-principles spin-phonon tensors can provide predictive insight into decoherence pathways and guide the rational design of molecular qubits.

quant-ph↗

arXiv2025

X-Intelligence 3.0: Training and Evaluating Reasoning LLM for Semiconductor Display

Large language models (LLMs) have recently achieved significant advances in reasoning and demonstrated their advantages in solving challenging problems. Yet, their effectiveness in the semiconductor display industry remains limited due to a lack of domain-specific training and expertise. To bridge this gap, we present X-Intelligence 3.0, the first high-performance reasoning model specifically developed for the semiconductor display industry. This model is designed to deliver expert-level understanding and reasoning for the industry's complex challenges. Leveraging a carefully curated industry knowledge base, the model undergoes supervised fine-tuning and reinforcement learning to enhance its reasoning and comprehension capabilities. To further accelerate development, we implemented an automated evaluation framework that simulates expert-level assessments. We also integrated a domain-specific retrieval-augmented generation (RAG) mechanism, resulting in notable performance gains on benchmark datasets. Despite its relatively compact size of 32 billion parameters, X-Intelligence 3.0 outperforms SOTA DeepSeek-R1-671B across multiple evaluations. This demonstrates its exceptional efficiency and establishes it as a powerful solution to the longstanding reasoning challenges faced by the semiconductor display industry.

cs.CL↗

arXiv2025

GTSinger: A Global Multi-Technique Singing Corpus with Realistic Music Scores for All Singing Tasks

The scarcity of high-quality and multi-task singing datasets significantly hinders the development of diverse controllable and personalized singing tasks, as existing singing datasets suffer from low quality, limited diversity of languages and singers, absence of multi-technique information and realistic music scores, and poor task suitability. To tackle these problems, we present GTSinger, a large global, multi-technique, free-to-use, high-quality singing corpus with realistic music scores, designed for all singing tasks, along with its benchmarks. Particularly, (1) we collect 80.59 hours of high-quality singing voices, forming the largest recorded singing dataset; (2) 20 professional singers across nine widely spoken languages offer diverse timbres and styles; (3) we provide controlled comparison and phoneme-level annotations of six commonly used singing techniques, helping technique modeling and control; (4) GTSinger offers realistic music scores, assisting real-world musical composition; (5) singing voices are accompanied by manual phoneme-to-audio alignments, global style labels, and 16.16 hours of paired speech for various singing tasks. Moreover, to facilitate the use of GTSinger, we conduct four benchmark experiments: technique-controllable singing voice synthesis, technique recognition, style transfer, and speech-to-singing conversion. The demos can be found at http://aaronz345.github.io/GTSingerDemo/. We provide the dataset and the code for processing data and conducting benchmarks at https://huggingface.co/datasets/AaronZ345/GTSinger and https://github.com/AaronZ345/GTSinger.

eess.AS↗

arXiv2025

The impact of stellar rotations and binaries on the shape of upper main sequence near turn off in open cluster NGC\,6067

We present the analysis of the extended Main Sequence Turn-Off (eMSTO) in the open cluster NGC\,6067. We derive the projected rotational velocity, \textit{v}sin\textit{i}, of the stars belonging to the eMSTO region of the main sequence (MS) utilizing \textit{Gaia}-ESO spectra. Our results reveal a positive correlation between \textit{v}sin\textit{i} and the color of eMSTO stars, where fast-rotating stars predominantly occupy the red part of the MS while slow-rotating ones prefer a bluer side of the MS. The gravity-darkening effect might be a reason for this correlation. We find that most of the close binaries present in the eMSTO population would be slow-rotating due to the tidal-locking phenomenon. We identify four double-lined spectroscopic binaries (SB2) featuring slow-rotating companions, further supporting this tidal-locking hypothesis. However, the spatial distribution and the cumulative radial distribution indicate a higher concentration of red eMSTO stars in the cluster's central region than their bluer counterparts. This suggests that tidal locking is less likely to be the cause of the observed spread in rotation rates among eMSTO stars. Instead, we propose that star-disk interactions during the pre-main-sequence phase might have played a crucial role in spreading the rotation rates of stars, leading to the eMSTO phenomenon in NGC\,6067.

astro-ph.SR↗

arXiv2025

Spontaneous Spatial Cognition Emerges during Egocentric Video Viewing through Non-invasive BCI

Humans possess a remarkable capacity for spatial cognition, allowing for self-localization even in novel or unfamiliar environments. While hippocampal neurons encoding position and orientation are well documented, the large-scale neural dynamics supporting spatial representation, particularly during naturalistic, passive experience, remain poorly understood. Here, we demonstrate for the first time that non-invasive brain-computer interfaces (BCIs) based on electroencephalography (EEG) can decode spontaneous, fine-grained egocentric 6D pose, comprising three-dimensional position and orientation, during passive viewing of egocentric video. Despite EEG's limited spatial resolution and high signal noise, we find that spatially coherent visual input (i.e., continuous and structured motion) reliably evokes decodable spatial representations, aligning with participants' subjective sense of spatial engagement. Decoding performance further improves when visual input is presented at a frame rate of 100 ms per image, suggesting alignment with intrinsic neural temporal dynamics. Using gradient-based backpropagation through a neural decoding model, we identify distinct EEG channels contributing to position -- and orientation specific -- components, revealing a distributed yet complementary neural encoding scheme. These findings indicate that the brain's spatial systems operate spontaneously and continuously, even under passive conditions, challenging traditional distinctions between active and passive spatial cognition. Our results offer a non-invasive window into the automatic construction of egocentric spatial maps and advance our understanding of how the human mind transforms everyday sensory experience into structured internal representations.

q-bio.NC↗

arXiv2025

Nexus-Gen: Unified Image Understanding, Generation, and Editing via Prefilled Autoregression in Shared Embedding Space

Unified multimodal generative models aim to integrate image understanding and generation abilities, offering significant advantages in harnessing multimodal corpora, particularly interleaved text-image data. However, existing unified models exhibit limitations in image synthesis quality, autoregressive error accumulation, and image editing capability. In this work, we propose Nexus-Gen, a novel architecture that unifies image understanding, generation, and editing tasks in a shared image embedding space. This shared space serves as a bridge for the autoregressive and diffusion models, which seamlessly integrates their complementary strengths in cross-modal modeling. To mitigate the severe error accumulation during autoregressive embedding prediction, we propose a novel prefilled autoregression strategy that aligns training-inference dynamics by prefilling input sequences with learnable embeddings. After multi-stage and multi-task training on our constructed large-scale dataset with 26.3 million samples, Nexus-Gen achieves state-of-the-art performance on the evaluation benchmarks spanning image understanding, generation and editing tasks. All models, datasets, and source codes are released in https://github.com/modelscope/Nexus-Gen to facilitate further advancements across the field.

cs.CV↗

arXiv2025

Assessing Color Vision Test in Large Vision-language Models

With the widespread adoption of large vision-language models, the capacity for color vision in these models is crucial. However, the color vision abilities of large visual-language models have not yet been thoroughly explored. To address this gap, we define a color vision testing task for large vision-language models and construct a dataset \footnote{Anonymous Github Showing some of the data https://anonymous.4open.science/r/color-vision-test-dataset-3BCD} that covers multiple categories of test questions and tasks of varying difficulty levels. Furthermore, we analyze the types of errors made by large vision-language models and propose fine-tuning strategies to enhance their performance in color vision tests.

cs.CV↗

arXiv2025

Accelerating Automatic Program Repair with Dual Retrieval-Augmented Fine-Tuning and Patch Generation on Large Language Models

Automated Program Repair (APR) is essential for ensuring software reliability and quality while enhancing efficiency and reducing developers' workload. Although rule-based and learning-based APR methods have demonstrated their effectiveness, their performance was constrained by the defect type of repair, the quality of training data, and the size of model parameters. Recently, Large Language Models (LLMs) combined with Retrieval-Augmented-Generation (RAG) have been increasingly adopted in APR tasks. However, current code LLMs and RAG designs neither fully address code repair tasks nor consider code-specific features. To overcome these limitations, we propose SelRepair, a novel APR approach with integration of a fine-tuned LLM with a newly-designed dual RAG module. This approach uses a bug-fix pair dataset for fine-tuning and incorporates semantic and syntactic/structural similarity information through an RAG selection gate. This design ensures relevant information is retrieved efficiently, thereby reducing token length and inference time. Evaluations on Java datasets show SelRepair outperforms other APR methods, achieving 26.29% and 17.64% in terms of exact match (EM) on different datasets while reducing inference time by at least 6.42% with controlled input lengths.

cs.SE↗

arXiv2025

Structural Representations and Identification of Marginal Policy Effects

This paper investigates the structural interpretation of the marginal policy effect (MPE) within nonseparable models. We demonstrate that, for a smooth functional of the outcome distribution, the MPE equals its functional derivative evaluated at the outcome-conditioned weighted average structural derivative. This equivalence is definitional rather than identification-based. Building on this theoretical result, we propose an alternative identification strategy for the MPE that complements existing methods.

econ.EM↗

arXiv2025

STARS: A Unified Framework for Singing Transcription, Alignment, and Refined Style Annotation

Recent breakthroughs in singing voice synthesis (SVS) have heightened the demand for high-quality annotated datasets, yet manual annotation remains prohibitively labor-intensive and resource-intensive. Existing automatic singing annotation (ASA) methods, however, primarily tackle isolated aspects of the annotation pipeline. To address this fundamental challenge, we present STARS, which is, to our knowledge, the first unified framework that simultaneously addresses singing transcription, alignment, and refined style annotation. Our framework delivers comprehensive multi-level annotations encompassing: (1) precise phoneme-audio alignment, (2) robust note transcription and temporal localization, (3) expressive vocal technique identification, and (4) global stylistic characterization including emotion and pace. The proposed architecture employs hierarchical acoustic feature processing across frame, word, phoneme, note, and sentence levels. The novel non-autoregressive local acoustic encoders enable structured hierarchical representation learning. Experimental validation confirms the framework's superior performance across multiple evaluation dimensions compared to existing annotation approaches. Furthermore, applications in SVS training demonstrate that models utilizing STARS-annotated data achieve significantly enhanced perceptual naturalness and precise style control. This work not only overcomes critical scalability challenges in the creation of singing datasets but also pioneers new methodologies for controllable singing voice synthesis. Audio samples are available at https://gwx314.github.io/stars-demo/.

cs.SD↗

arXiv2025

Pre-Training Graph Contrastive Masked Autoencoders are Strong Distillers for EEG

Effectively utilizing extensive unlabeled high-density EEG data to improve performance in scenarios with limited labeled low-density EEG data presents a significant challenge. In this paper, we address this challenge by formulating it as a graph transfer learning and knowledge distillation problem. We propose a Unified Pre-trained Graph Contrastive Masked Autoencoder Distiller, named EEG-DisGCMAE, to bridge the gap between unlabeled and labeled as well as high- and low-density EEG data. Our approach introduces a novel unified graph self-supervised pre-training paradigm, which seamlessly integrates the graph contrastive pre-training with the graph masked autoencoder pre-training. Furthermore, we propose a graph topology distillation loss function, allowing a lightweight student model trained on low-density data to learn from a teacher model trained on high-density data during pre-training and fine-tuning. This method effectively handles missing electrodes through contrastive distillation. We validate the effectiveness of EEG-DisGCMAE across four classification tasks using two clinical EEG datasets with abundant data. The source code is available at https://github.com/weixinxu666/EEG_DisGCMAE.

cs.LG↗

arXiv2025

PAST: A multimodal single-cell foundation model for histopathology and spatial transcriptomics in cancer

While pathology foundation models have transformed cancer image analysis, they often lack integration with molecular data at single-cell resolution, limiting their utility for precision oncology. Here, we present PAST, a pan-cancer single-cell foundation model trained on 20 million paired histopathology images and single-cell transcriptomes spanning multiple tumor types and tissue contexts. By jointly encoding cellular morphology and gene expression, PAST learns unified cross-modal representations that capture both spatial and molecular heterogeneity at the cellular level. This approach enables accurate prediction of single-cell gene expression, virtual molecular staining, and multimodal survival analysis directly from routine pathology slides. Across diverse cancers and downstream tasks, PAST consistently exceeds the performance of existing approaches, demonstrating robust generalizability and scalability. Our work establishes a new paradigm for pathology foundation models, providing a versatile tool for high-resolution spatial omics, mechanistic discovery, and precision cancer research.

q-bio.QM↗

arXiv2025

SimVecVis: A Dataset for Enhancing MLLMs in Visualization Understanding

Current multimodal large language models (MLLMs), while effective in natural image understanding, struggle with visualization understanding due to their inability to decode the data-to-visual mapping and extract structured information. To address these challenges, we propose SimVec, a novel simplified vector format that encodes chart elements such as mark type, position, and size. The effectiveness of SimVec is demonstrated by using MLLMs to reconstruct chart information from SimVec formats. Then, we build a new visualization dataset, SimVecVis, to enhance the performance of MLLMs in visualization understanding, which consists of three key dimensions: bitmap images of charts, their SimVec representations, and corresponding data-centric question-answering (QA) pairs with explanatory chain-of-thought (CoT) descriptions. We finetune state-of-the-art MLLMs (e.g., MiniCPM and Qwen-VL), using SimVecVis with different dataset dimensions. The experimental results show that it leads to substantial performance improvements of MLLMs with good spatial perception capabilities (e.g., MiniCPM) in data-centric QA tasks. Our dataset and source code are available at: https://github.com/VIDA-Lab/SimVecVis.

cs.HC↗