arXiv subjects

Xin Cheng

Publications and source records attributed to Xin Cheng.

At least 73 records · Page 4Linked to original sources

arXiv2025

EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics

Imitating how humans move their gaze in a visual scene is a vital research problem for both visual understanding and psychology, kindling crucial applications such as building alive virtual characters. Previous studies aim to predict gaze trajectories when humans are free-viewing an image, searching for required targets, or looking for clues to answer questions in an image. While these tasks focus on visual-centric scenarios, humans move their gaze also along with audio signal inputs in more common scenarios. To fill this gap, we introduce a new task that predicts human gaze trajectories in a visual scene with synchronized audio inputs and provide a new dataset containing 20k gaze points from 8 subjects. To effectively integrate audio information and simulate the dynamic process of human gaze motion, we propose a novel learning framework called EyEar (Eye moving while Ear listening) based on physics-informed dynamics, which considers three key factors to predict gazes: eye inherent motion tendency, vision salient attraction, and audio semantic attraction. We also propose a probability density score to overcome the high individual variability of gaze trajectories, thereby improving the stabilization of optimization and the reliability of the evaluation. Experimental results show that EyEar outperforms all the baselines in the context of all evaluation metrics, thanks to the proposed components in the learning model.

cs.MM↗

arXiv2025

PersonaMath: Boosting Mathematical Reasoning via Persona-Driven Data Augmentation

While closed-source Large Language Models (LLMs) demonstrate strong mathematical problem-solving abilities, open-source models still face challenges with such tasks. To bridge this gap, we propose a data augmentation approach and introduce PersonaMathQA, a dataset derived from MATH and GSM8K, on which we train the PersonaMath models. Our approach consists of two stages: the first stage focuses on learning from Persona Diversification, and the second stage emphasizes learning from Reflection. In the first stage, we regenerate detailed chain-of-thought (CoT) solutions as instructions using a closed-source LLM and introduce a persona-driven data augmentation technique. This technique innovatively classifies personas based on occupations, significantly enhancing the dataset's diversity and quality. In the second stage, we incorporate reflection to fully leverage more challenging and valuable questions. Evaluation of our PersonaMath models on MATH and GSM8K reveals that the PersonaMath-7B model (based on Qwen2.5-7B) achieves an accuracy of 61.2% on MATH and 87.8% on GSM8K, surpassing all baseline methods and achieving state-of-the-art performance. Notably, our dataset contains only 128.9K data points-merely 32.6% of MetaMathQA and 49.5% of MathInstruct-yet our model outperforms these baselines, demonstrating the high quality and diversity of our dataset, which enables more efficient model training. We open-source the PersonaMathQA dataset, PersonaMath models, and our code for public usage.

cs.CL↗

arXiv2025

Understanding observational characteristics of solar flare current sheets

The elongated bright structures above solar flare loops are suggested to be current sheets, where magnetic reconnection takes place. Observations have revealed various characteristics of the current sheet; however, their physical origin remains to be ascertained. In this study we aim to reveal the relations of observational characteristics of current sheets with the fundamental processes of magnetic reconnection. Using high-resolution 3D magnetohydrodynamic simulations of turbulent magnetic reconnection within a solar flare current sheet, we synthesized the remote-sensing observations of the current sheet and determined their physical properties. Turbulent magnetic reconnection can significantly broaden the apparent width of the current sheet, which is much larger than the realistic physical width because of the superposition effect. The differential emission measures of the current sheet have two peaks; the high-temperature component is spatially related to confirmed small-scale reconnection sites, showing that the current sheet is directly heated by reconnection. Moreover, we demonstrate that strong turbulence can cause the nonthermal broadening of spectral lines at both the current sheet and flare loop-top regions. A strong correlation between them in time is also observed. Our 3D turbulent magnetic reconnection flare model can be used to interpret primary observational characteristics of the elongated bright current sheets of solar flares.

astro-ph.SR↗

arXiv2025

On Turán problems for suspension hypergraphs

For a given graph $F$, the $r$-uniform suspension of $F$ is the $r$-uniform hypergraph obtained from $F$ by taking $r-2$ new vertices and adding them to every edge. In this paper, we consider Turán problems on suspension hypergraphs, and we obtain several general and exact results.

math.CO↗

arXiv2025

Flaring Activities of Fast Rotating Stars have Solar-like Latitudinal Distribution

The dynamo theory has always been one of the biggest mysteries in stellar physics. One key reason for its uncertainty is poor knowledge of the dynamo process on stars except the Sun. The most important observation feature of solar dynamo is that active regions only appear at low latitudes, which provides a crucial constraint to the dynamo theory, while Doppler imaging, the current technique to spatially resolve stellar hemisphere, is difficult to distinguish the equatorial region . Hence, the latitudinal distribution of active regions (LDAR) of stars is ambiguous and controversial, mainly due to the limit of the current technique for spatially resolving the stellar surface. Fast rotating stars, which are young and active, are thought to operate with a different dynamo process than the Sun. We study their LDAR and compare them with the Sun to reveal the underlying dynamo process. Flares are drastic and observational activity events, which occur in active regions. Here, we propose a new method to study how the apparent flaring activity varies with respect to the inclination to determine the LDAR of fast rotating stars.We find that the LDAR of fast rotating stars is consistent with that of the Sun, contrary to expectations. Our results provide a crucial constraint to stellar dynamo, indicating that the solar-like dynamo also applies to fast rotating stars, even spanning different stages of their evolution.

astro-ph.SR↗

arXiv2024

LoVA: Long-form Video-to-Audio Generation

Video-to-audio (V2A) generation is important for video editing and post-processing, enabling the creation of semantics-aligned audio for silent video. However, most existing methods focus on generating short-form audio for short video segment (less than 10 seconds), while giving little attention to the scenario of long-form video inputs. For current UNet-based diffusion V2A models, an inevitable problem when handling long-form audio generation is the inconsistencies within the final concatenated audio. In this paper, we first highlight the importance of long-form V2A problem. Besides, we propose LoVA, a novel model for Long-form Video-to-Audio generation. Based on the Diffusion Transformer (DiT) architecture, LoVA proves to be more effective at generating long-form audio compared to existing autoregressive models and UNet-based diffusion models. Extensive objective and subjective experiments demonstrate that LoVA achieves comparable performance on 10-second V2A benchmark and outperforms all other baselines on a benchmark with long-form video input.

cs.SD↗

arXiv2024

xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token

This paper introduces xRAG, an innovative context compression method tailored for retrieval-augmented generation. xRAG reinterprets document embeddings in dense retrieval--traditionally used solely for retrieval--as features from the retrieval modality. By employing a modality fusion methodology, xRAG seamlessly integrates these embeddings into the language model representation space, effectively eliminating the need for their textual counterparts and achieving an extreme compression rate. In xRAG, the only trainable component is the modality bridge, while both the retriever and the language model remain frozen. This design choice allows for the reuse of offline-constructed document embeddings and preserves the plug-and-play nature of retrieval augmentation. Experimental results demonstrate that xRAG achieves an average improvement of over 10% across six knowledge-intensive tasks, adaptable to various language model backbones, ranging from a dense 7B model to an 8x7B Mixture of Experts configuration. xRAG not only significantly outperforms previous context compression methods but also matches the performance of uncompressed models on several datasets, while reducing overall FLOPs by a factor of 3.53. Our work pioneers new directions in retrieval-augmented generation from the perspective of multimodality fusion, and we hope it lays the foundation for future efficient and scalable retrieval-augmented systems

cs.CL↗

arXiv2024

Cycle-Constrained Adversarial Denoising Convolutional Network for PET Image Denoising: Multi-Dimensional Validation on Large Datasets with Reader Study and Real Low-Dose Data

Positron emission tomography (PET) is a critical tool for diagnosing tumors and neurological disorders but poses radiation risks to patients, particularly to sensitive populations. While reducing injected radiation dose mitigates this risk, it often compromises image quality. To reconstruct full-dose-quality images from low-dose scans, we propose a Cycle-constrained Adversarial Denoising Convolutional Network (Cycle-DCN). This model integrates a noise predictor, two discriminators, and a consistency network, and is optimized using a combination of supervised loss, adversarial loss, cycle consistency loss, identity loss, and neighboring Structural Similarity Index (SSIM) loss. Experiments were conducted on a large dataset consisting of raw PET brain data from 1,224 patients, acquired using a Siemens Biograph Vision PET/CT scanner. Each patient underwent a 120-seconds brain scan. To simulate low-dose PET conditions, images were reconstructed from shortened scan durations of 30, 12, and 5 seconds, corresponding to 1/4, 1/10, and 1/24 of the full-dose acquisition, respectively, using a custom-developed GPU-based image reconstruction software. The results show that Cycle-DCN significantly improves average Peak Signal-to-Noise Ratio (PSNR), SSIM, and Normalized Root Mean Square Error (NRMSE) across three dose levels, with improvements of up to 56%, 35%, and 71%, respectively. Additionally, it achieves contrast-to-noise ratio (CNR) and Edge Preservation Index (EPI) values that closely align with full-dose images, effectively preserving image details, tumor shape, and contrast, while resolving issues with blurred edges. The results of reader studies indicated that the images restored by Cycle-DCN consistently received the highest ratings from nuclear medicine physicians, highlighting their strong clinical relevance.

eess.IV↗

arXiv2024

Simultaneous Eruption and Shrinkage of Pre-existing Flare Loops during a Subsequent Solar Eruption

We investigated two consecutive solar eruption events in the solar active region (AR) 12994 at the solar eastern limb on 2022 April 15. We found that the flare loops formed by the first eruption were involved in the second eruption. During the initial stage of the second flare, the middle part of these flare loops (E-loops) erupted outward along with the flux ropes below, while the parts of the flare loops (I-loops1 and I-loops2) on either side of the E-loops first rose and then contracted. Approximately 1 hour after the eruption, the heights of I-loops1 and I-loops2 decreased by 9 Mm and 45 Mm, respectively, compared to before the eruption. Their maximum descent velocities were 30 km/s and 130 km/s, respectively. The differential emission measure (DEM) results indicate that the plasma above I-loops1 and I-loops2 began to be heated about 23 minutes and 44 minutes after the start of the second flare, respectively. Within 20 minutes, the plasma temperature in these regions increased from ~3 MK to 6 MK. We proposed an adiabatic heating mechanism that magnetic energy would be converted into thermal and kinetic energy when the pre-stretched loops contract. Our calculations show that the magnetic energy required to heat the two high-temperature regions are 10^29-10^30 erg, which correspond to a loss of field strength of 2-3 G.

astro-ph.SR↗

arXiv2024

ChipExpert: The Open-Source Integrated-Circuit-Design-Specific Large Language Model

The field of integrated circuit (IC) design is highly specialized, presenting significant barriers to entry and research and development challenges. Although large language models (LLMs) have achieved remarkable success in various domains, existing LLMs often fail to meet the specific needs of students, engineers, and researchers. Consequently, the potential of LLMs in the IC design domain remains largely unexplored. To address these issues, we introduce ChipExpert, the first open-source, instructional LLM specifically tailored for the IC design field. ChipExpert is trained on one of the current best open-source base model (Llama-3 8B). The entire training process encompasses several key stages, including data preparation, continue pre-training, instruction-guided supervised fine-tuning, preference alignment, and evaluation. In the data preparation stage, we construct multiple high-quality custom datasets through manual selection and data synthesis techniques. In the subsequent two stages, ChipExpert acquires a vast amount of IC design knowledge and learns how to respond to user queries professionally. ChipExpert also undergoes an alignment phase, using Direct Preference Optimization, to achieve a high standard of ethical performance. Finally, to mitigate the hallucinations of ChipExpert, we have developed a Retrieval-Augmented Generation (RAG) system, based on the IC design knowledge base. We also released the first IC design benchmark ChipICD-Bench, to evaluate the capabilities of LLMs across multiple IC design sub-domains. Through comprehensive experiments conducted on this benchmark, ChipExpert demonstrated a high level of expertise in IC design knowledge Question-and-Answer tasks.

cs.AR↗

arXiv2024

Flexible and Adaptable Summarization via Expertise Separation

A proficient summarization model should exhibit both flexibility -- the capacity to handle a range of in-domain summarization tasks, and adaptability -- the competence to acquire new knowledge and adjust to unseen out-of-domain tasks. Unlike large language models (LLMs) that achieve this through parameter scaling, we propose a more parameter-efficient approach in this study. Our motivation rests on the principle that the general summarization ability to capture salient information can be shared across different tasks, while the domain-specific summarization abilities need to be distinct and tailored. Concretely, we propose MoeSumm, a Mixture-of-Expert Summarization architecture, which utilizes a main expert for gaining the general summarization capability and deputy experts that selectively collaborate to meet specific summarization task requirements. We further propose a max-margin loss to stimulate the separation of these abilities. Our model's distinct separation of general and domain-specific summarization abilities grants it with notable flexibility and adaptability, all while maintaining parameter efficiency. MoeSumm achieves flexibility by managing summarization across multiple domains with a single model, utilizing a shared main expert and selected deputy experts. It exhibits adaptability by tailoring deputy experts to cater to out-of-domain few-shot and zero-shot scenarios. Experimental results on 11 datasets show the superiority of our model compared with recent baselines and LLMs. We also provide statistical and visual evidence of the distinct separation of the two abilities in MoeSumm (https://github.com/iriscxy/MoE_Summ).

cs.CL↗

arXiv2024

P$^2$-ViT: Power-of-Two Post-Training Quantization and Acceleration for Fully Quantized Vision Transformer

Vision Transformers (ViTs) have excelled in computer vision tasks but are memory-consuming and computation-intensive, challenging their deployment on resource-constrained devices. To tackle this limitation, prior works have explored ViT-tailored quantization algorithms but retained floating-point scaling factors, which yield non-negligible re-quantization overhead, limiting ViTs' hardware efficiency and motivating more hardware-friendly solutions. To this end, we propose \emph{P$^2$-ViT}, the first \underline{P}ower-of-Two (PoT) \underline{p}ost-training quantization and acceleration framework to accelerate fully quantized ViTs. Specifically, {as for quantization,} we explore a dedicated quantization scheme to effectively quantize ViTs with PoT scaling factors, thus minimizing the re-quantization overhead. Furthermore, we propose coarse-to-fine automatic mixed-precision quantization to enable better accuracy-efficiency trade-offs. {In terms of hardware,} we develop {a dedicated chunk-based accelerator} featuring multiple tailored sub-processors to individually handle ViTs' different types of operations, alleviating reconfigurable overhead. Additionally, we design {a tailored row-stationary dataflow} to seize the pipeline processing opportunity introduced by our PoT scaling factors, thereby enhancing throughput. Extensive experiments consistently validate P$^2$-ViT's effectiveness. {Particularly, we offer comparable or even superior quantization performance with PoT scaling factors when compared to the counterpart with floating-point scaling factors. Besides, we achieve up to $\mathbf{10.1\times}$ speedup and $\mathbf{36.8\times}$ energy saving over GPU's Turing Tensor Cores, and up to $\mathbf{1.84\times}$ higher computation utilization efficiency against SOTA quantization-based ViT accelerators. Codes are available at \url{https://github.com/shihuihong214/P2-ViT}.

cs.AI↗

arXiv2024

Leveraging 2D Information for Long-term Time Series Forecasting with Vanilla Transformers

Time series prediction is crucial for understanding and forecasting complex dynamics in various domains, ranging from finance and economics to climate and healthcare. Based on Transformer architecture, one approach involves encoding multiple variables from the same timestamp into a single temporal token to model global dependencies. In contrast, another approach embeds the time points of individual series into separate variate tokens. The former method faces challenges in learning variate-centric representations, while the latter risks missing essential temporal information critical for accurate forecasting. In our work, we introduce GridTST, a model that combines the benefits of two approaches using innovative multi-directional attentions based on a vanilla Transformer. We regard the input time series data as a grid, where the $x$-axis represents the time steps and the $y$-axis represents the variates. A vertical slicing of this grid combines the variates at each time step into a \textit{time token}, while a horizontal slicing embeds the individual series across all time steps into a \textit{variate token}. Correspondingly, a \textit{horizontal attention mechanism} focuses on time tokens to comprehend the correlations between data at various time steps, while a \textit{vertical}, variate-aware \textit{attention} is employed to grasp multivariate correlations. This combination enables efficient processing of information across both time and variate dimensions, thereby enhancing the model's analytical strength. % We also integrate the patch technique, segmenting time tokens into subseries-level patches, ensuring that local semantic information is retained in the embedding. The GridTST model consistently delivers state-of-the-art performance across various real-world datasets.

cs.LG↗

arXiv2024

Eruption of a million-Kelvin warm magnetic flux rope on the Sun

Solar magnetic flux rope (MFR) plays a central role in the physics of coronal mass ejections (CMEs). It mainly includes a cold filament at typical chromospheric temperatures (10000 K) and a hot channel at high coronal temperatures (10 MK). The warm MFR at quiescent coronal temperatures of a million Kelvin is, however, rarely reported. In this study, using multiwavelength images from Atmospheric Imaging Assembly (AIA) on board the Solar Dynamic Observatory (SDO) and Extreme Ultraviolet Imager (EUVI) on board the Solar Terrestrial Relations Observatory-A (STEREO-A), we present an eruption of a warm channel, that represents an MFR with quiescent coronal temperatures (0.6-2.5 MK). On 2022 May 8, we observed the failed eruption of a hot channel, with the average temperature and emission measure (EM) of 10 MK and 1.1*1028 cm^-5, using AIA high-temperature images in active region (AR) 13007. This failed eruption was associated with a C8.2 flare, with no CME. Subsequently, we observed a warm channel that appeared in AIA and EUVI low-temperature images, rather than AIA high-temperature images. It then erupted, and transformed toward a semi-circular shape. An associated C2.1 flare, along with the signatures of magnetic reconnection in AIA high-temperature images, were identified. Additionally, we observed a CME associated with this event. Compared with the hot channel, the warm channel is cooler and rarer with the average temperature and EM of 1.7 (1.6) MK and 2.0*1026 (2.3*1026) cm^-5. All the results suggest an unambiguous observation of the million-Kelvin warm MFR, that erupted as a CME, and fill a gap in the temperature domain of coronal MFRs.

astro-ph.SR↗

arXiv2024

Facility Assignment with Fair Cost Sharing: Equilibrium and Mechanism Design

In the one-dimensional facility assignment problem, m facilities and n agents are positioned along the real line. Each agent will be assigned to a single facility to receive service. Each facility incurs a building cost, which is shared equally among the agents utilizing it. Additionally, each agent independently bears a connection cost to access a facility. Thus, an agent's cost is the sum of the connection cost and her portion of the building cost. The social cost is the total cost of all agents. Notably, the optimal assignment that minimizes the social cost can be found in polynomial time. In this paper, we study the problem from two game-theoretical settings regarding the strategy space of agents and the rule the assignment. In both settings, agents act strategically to minimize their individual costs. In our first setting, the strategy space of agents is the set of facilities, granting agents the freedom to select any facility. Consequently, the self-formed assignment can exhibit instability, as agents may deviate to other facilities. We focus on the computation of an equilibrium assignment, where no agent has an incentive to unilaterally change her choice. We show that we can compute a pure Nash equilibrium in polynomial time. In our second setting, agents report their positions to a mechanism for assignment to facilities. The strategy space of agents becomes the set of all positions. Our interest lies in strategyproof mechanisms. It is essential to note that the preference induced by the agents' cost function is more complex as it depends on how other agents are assigned. We establish a strong lower bound against all strategyproof and anonymous mechanisms: none can achieve a bounded social cost approximation ratio. Nonetheless, we identify a class of non-trivial strategyproof mechanisms for any n and m that is unanimous and anonymous.

cs.GT↗

arXiv2024

A model for heating the super-hot corona in solar active regions

What physical mechanisms heat the outer solar or stellar atmosphere to million-Kelvin temperatures is a fundamental but long-standing open question. In particular, the solar corona in active region cores contains an even hotter component reaching ten million Kelvin, manifesting as persistent coronal loops in extreme ultraviolet and soft X-ray images, which imposes a more stringent energy budget. Here, we present a self-consistent coronal heating model using a state-of-the-art three-dimensional radiative magnetohydrodynamics simulation. We find that the continuous magnetic flux emergence in active regions keeps driving magnetic reconnections that release energy impulsively but, on time average, persistently. As a result, numerous sub-structures are heated to ten million Kelvin and then evolve independently, which collectively form long-lived and stable coronal loops as in observations. This provides a heating model explaining the origin of the super-hot coronal plasma and the persistence of hot coronal loops in emerging active regions.

astro-ph.SR↗

arXiv2024

A Method for Determining the Locations and Configurations of Magnetic Reconnection within 3D Turbulent Plasmas

Context. Three-dimensional (3D) reconnection is an important mechanism for efficiently releasing energy during astrophysical eruptive events, which is difficult to be quantitatively analyzed especially within turbulent plasmas. Aims. In this paper, an efficient method for identifying locations and configurations of 3D reconnection from MHD data is developed. Methods. This method analyzes the local nonideal electric field and magnetic structure at an arbitrary position. As only performing algebraical manipulations on the discrete field data and avoiding computationally expensive operations like field-line tracing and root-finding, this method naturally possesses high efficiency. To validate this method, we apply it to the 3D data from a high-resolution simulation of a Harris-sheet reconnection and a data-driven simulation of a coronal flux rope eruption. Results. It is shown that this method can precisely identify the local structures of discrete magnetic field. Through the information of nonideal electric field and the geometric attributes of magnetic field, the local structures of reconnection sites can be effectively and comprehensively determined. For fine turbulent processes, both qualitative pictures and quantitative statistical properties of small-scale reconnection structures can be obtained. For large-scale solar simulations, macro-scale magnetic structures such as flux ropes and eruption current sheets can also be recognized. Conclusions. We develop a powerful method to analyze multi-scale structures of 3D reconnection. It can be applied not only in MHD simulations but also in kinetic simulations, plasma experiments, and in-situ observations.

astro-ph.SR↗

arXiv2024

StyleChat: Learning Recitation-Augmented Memory in LLMs for Stylized Dialogue Generation

Large Language Models (LLMs) demonstrate superior performance in generative scenarios and have attracted widespread attention. Among them, stylized dialogue generation is essential in the context of LLMs for building intelligent and engaging dialogue agent. However the ability of LLMs is data-driven and limited by data bias, leading to poor performance on specific tasks. In particular, stylized dialogue generation suffers from a severe lack of supervised data. Furthermore, although many prompt-based methods have been proposed to accomplish specific tasks, their performance in complex real-world scenarios involving a wide variety of dialog styles further enhancement. In this work, we first introduce a stylized dialogue dataset StyleEval with 38 styles by leveraging the generative power of LLMs comprehensively, which has been carefully constructed with rigorous human-led quality control. Based on this, we propose the stylized dialogue framework StyleChat via recitation-augmented memory strategy and multi-task style learning strategy to promote generalization ability. To evaluate the effectiveness of our approach, we created a test benchmark that included both a generation task and a choice task to comprehensively evaluate trained models and assess whether styles and preferences are remembered and understood. Experimental results show that our proposed framework StyleChat outperforms all the baselines and helps to break the style boundary of LLMs.

cs.CL↗