arXiv subjects

Yue Yin

Publications and source records attributed to Yue Yin.

At least 37 records · Page 2Linked to original sources

arXiv2025

An Efficient Medical Image Classification Method Based on a Lightweight Improved ConvNeXt-Tiny Architecture

Intelligent analysis of medical imaging plays a crucial role in assisting clinical diagnosis. However, achieving efficient and high-accuracy image classification in resource-constrained computational environments remains challenging. This study proposes a medical image classification method based on an improved ConvNeXt-Tiny architecture. Through structural optimization and loss function design, the proposed method enhances feature extraction capability and classification performance while reducing computational complexity. Specifically, the method introduces a dual global pooling (Global Average Pooling and Global Max Pooling) feature fusion strategy into the ConvNeXt-Tiny backbone to simultaneously preserve global statistical features and salient response information. A lightweight channel attention module, termed Squeeze-and-Excitation Vector (SEVector), is designed to improve the adaptive allocation of channel weights while minimizing parameter overhead. Additionally, a Feature Smoothing Loss is incorporated into the loss function to enhance intra-class feature consistency and suppress intra-class variance. Under CPU-only conditions (8 threads), the method achieves a maximum classification accuracy of 89.10% on the test set within 10 training epochs, exhibiting a stable convergence trend in loss values. Experimental results demonstrate that the proposed method effectively improves medical image classification performance in resource-limited settings, providing a feasible and efficient solution for the deployment and promotion of medical imaging analysis models.

cs.CV↗

arXiv2025

Oxidation States in Solids from Data-Driven Paradigms

The oxidation state (OS) is an essential chemical concept that embodies chemical intuition but cannot be computed with well-defined physical laws. We establish a data-driven paradigm, with its implementation as Tsinghua Oxidation States in Solids (TOSS), to explicitly compute the OSs in crystal structures as the emergent properties from large-sized datasets based on Bayesian maximum a posteriori probability (MAP). TOSS employs two looping structures over the large-sized dataset of crystal structures to obtain an emergent library of distance distributions as the foundation for chemically intuitive understanding and then determine the OSs by minimizing a loss function for each structure based on MAP and distance distributions in the whole dataset. The application of TOSS to a dataset of $\mathrm{>}$1,000,000 crystal structures delivers a superior success rate, and using the resulting OSs as the dataset, we further train a data-driven alternative to TOSS based on graph convolutional networks. We expect TOSS and the ML-model-based alternative to find a wide spectrum of applications, and this work also demonstrates an encouraging example for the data-driven paradigms to explicitly compute the chemical intuition for tackling complex problems in chemistry.

physics.chem-ph↗

arXiv2025

A Study of the Spectral Properties of Gamma-Ray Bursts with the Main and Second Bursts

The origins of the main burst and second burst of gamma-ray bursts (GRBs) and the composition of their jets remain uncertain. To explore this complex subject more thoroughly, we conduct a spectral analysis on 18 GRBs with a main and a second burst observed by Fermi/GBM. First, we employ Bayesian time-resolved spectral analysis to compare the spectral components of the main and the second burst, finding that $83.3\%$ of the main and second bursts contain a thermal component. $67\%$ of the GRBs, the thermal component gradually decreased from the main to the second burst and the number of spectra exceeding the "Synchrotron line-of-death" is significantly higher in the main burst than in the second burst. Subsequently, we ascertain that for both the main and second bursts, $71.4\%$ of the low-energy spectral index $α$ and $77.8\%$ of the peak energy $E_{p}$ evolve in a similar fashion. There are $50.0\%$ and $72.2\%$ of the GRBs exhibit comparable correlations for the $Flux-α$ and $α-E_{p}$, respectively. For $Flux-E_{p}$ both the main and second burst show a positive correlation. Moreover, from the perspective of the temporal evolution of characteristic radii, the transition from the main to the second burst appeared to be seamless. Finally, we find that both the main and the second burst follow the same Amati relation and Yonetoku relation. Our analysis strongly indicates that the second burst is a continuation of the main burst and is highly likely to share a common physical origin.

astro-ph.HE↗

arXiv2025

Haloscope Searching for Dark Photons at Q-band with a Novel Coupling Tuning Structure

Laboratory searching for dark matter is crucial for understanding several fundamental conundrums in physics and cosmology. Most cavity-based haloscope searches focus on the frequency range below 10 GHz, while the parameter space with higher frequency remains rarely explored, due to the challenges lying in the fabrication of microwave cavities. Here we report the first Q-band haloscope searching for dark photons with a 33.141 GHz cavity. A novel coupling tuning structure separated from the cavity was designed so as not to degrade the quality factor of the cavity. We have established the most stringent constraints $χ<2.5\times10^{-12}$ at a confidence level of 90$\%$ in the frequency range from 33.139 GHz to 33.143 GHz, corresponding to the mass of dark photons ranging from 137.05 $μ$eV to 137.07 $μ$eV. The results surpass the previous astronomical constraints by nearly three orders of magnitude. This work has demonstrated the feasibility of dark matter haloscopes at Q band. In the future, the constraints can be further improved by more than one order of magnitude through low-temperature experiments, and the setup can be extended to search for axions, axion-like particles, and high-frequency gravitational waves.

astro-ph.CO↗

arXiv2025

Dynamic Learning and Productivity for Data Analysts: A Bayesian Hidden Markov Model Perspective

Data analysts are essential in organizations, transforming raw data into insights that drive decision-making and strategy. This study explores how analysts' productivity evolves on a collaborative platform, focusing on two key learning activities: writing queries and viewing peer queries. While traditional research often assumes static models, where performance improves steadily with cumulative learning, such models fail to capture the dynamic nature of real-world learning. To address this, we propose a Hidden Markov Model (HMM) that tracks how analysts transition between distinct learning states based on their participation in these activities. Using an industry dataset with 2,001 analysts and 79,797 queries, this study identifies three learning states: novice, intermediate, and advanced. Productivity increases as analysts advance to higher states, reflecting the cumulative benefits of learning. Writing queries benefits analysts across all states, with the largest gains observed for novices. Viewing peer queries supports novices but may hinder analysts in higher states due to cognitive overload or inefficiencies. Transitions between states are also uneven, with progression from intermediate to advanced being particularly challenging. This study advances understanding of into dynamic learning behavior of knowledge worker and offers practical implications for designing systems, optimizing training, enabling personalized learning, and fostering effective knowledge sharing.

cs.SI↗

arXiv2025

InfoBid: A Simulation Framework for Studying Information Disclosure in Auctions with Large Language Model-based Agents

In online advertising systems, publishers often face a trade-off in information disclosure strategies: while disclosing more information can enhance efficiency by enabling optimal allocation of ad impressions, it may lose revenue potential by decreasing uncertainty among competing advertisers. Similar to other challenges in market design, understanding this trade-off is constrained by limited access to real-world data, leading researchers and practitioners to turn to simulation frameworks. The recent emergence of large language models (LLMs) offers a novel approach to simulations, providing human-like reasoning and adaptability without necessarily relying on explicit assumptions about agent behavior modeling. Despite their potential, existing frameworks have yet to integrate LLM-based agents for studying information asymmetry and signaling strategies, particularly in the context of auctions. To address this gap, we introduce InfoBid, a flexible simulation framework that leverages LLM agents to examine the effects of information disclosure strategies in multi-agent auction settings. Using GPT-4o, we implemented simulations of second-price auctions with diverse information schemas. The results reveal key insights into how signaling influences strategic behavior and auction outcomes, which align with both economic and social learning theories. Through InfoBid, we hope to foster the use of LLMs as proxies for human economic and social agents in empirical studies, enhancing our understanding of their capabilities and limitations. This work bridges the gap between theoretical market designs and practical applications, advancing research in market simulations, information design, and agent-based reasoning while offering a valuable tool for exploring the dynamics of digital economies.

cs.GT↗

arXiv2024

Human-Centered Design for AI-based Automatically Generated Assessment Reports: A Systematic Review

This paper provides a comprehensive review of the design and implementation of automatically generated assessment reports (AutoRs) for formative use in K-12 Science, Technology, Engineering, and Mathematics (STEM) classrooms. With the increasing adoption of technology-enhanced assessments, there is a critical need for human-computer interactive tools that efficiently support the interpretation and application of assessment data by teachers. AutoRs are designed to provide synthesized, interpretable, and actionable insights into students' performance, learning progress, and areas for improvement. Guided by cognitive load theory, this study emphasizes the importance of reducing teachers' cognitive demands through user-centered and intuitive designs. It highlights the potential of diverse information presentation formats such as text, visual aids, and plots and advanced functionalities such as live and interactive features to enhance usability. However, the findings also reveal that many existing AutoRs fail to fully utilize these approaches, leading to high initial cognitive demands and limited engagement. This paper proposes a conceptual framework to inform the design, implementation, and evaluation of AutoRs, balancing the trade-offs between usability and functionality. The framework aims to address challenges in engaging teachers with technology-enhanced assessment results, facilitating data-driven decision-making, and providing personalized feedback to improve the teaching and learning process.

cs.HC↗

arXiv2024

A Study of the Spectral properties of Gamma-Ray Bursts with the Precursors and Main bursts

There is no consensus yet on whether the precursor and the main burst of gamma-ray bursts (GRBs) have the same origin, and their jet composition is still unclear. In order to further investigate this issue, we systematically search 21 Fermi GRBs with both precursor and main burst for spectral analysis. We first perform Bayesian time-resolved spectral analysis and find that almost all the precursors and the main bursts (94.4$\%$) exhibit thermal components, and the vast majority of them have low-energy spectral index ($α$) (72.2$\%$) that exceed the limit of synchrotron radiation. We then analyse the evolution and correlation of the spectral parameters and find that approximately half of the $α$ (50$\%$) of the precursors and the main bursts evolve in a similar pattern, while peak energy ($E_{p}$) (55.6$\%$) behave similarly, and their evolution is mainly characterized by flux tracking; for the $α-F$ (the flux) relation, more than half of the precursors and the main bursts (61.1$\%$) exhibit roughly similar patterns; the $E_{p}-F$ relation in both the precursor and main burst (100$\%$) exhibits a positive correlation of at least moderate strength. Next, we constrain the outflow properties of the precursors and the main bursts and find that most of them exhibit typical properties of photosphere radiation. Finally, we compare the time-integrated spectra of the precursors and the main bursts and find that nearly all of them are located in similar regions of the Amati relation and follow the Yonetoku relation. Therefore, we conclude that main bursts are continuations of precursors and they may share a common physical origin.

astro-ph.HE↗

arXiv2023

Component Segmentation of Engineering Drawings Using Graph Convolutional Networks

We present a data-driven framework to automate the vectorization and machine interpretation of 2D engineering part drawings. In industrial settings, most manufacturing engineers still rely on manual reads to identify the topological and manufacturing requirements from drawings submitted by designers. The interpretation process is laborious and time-consuming, which severely inhibits the efficiency of part quotation and manufacturing tasks. While recent advances in image-based computer vision methods have demonstrated great potential in interpreting natural images through semantic segmentation approaches, the application of such methods in parsing engineering technical drawings into semantically accurate components remains a significant challenge. The severe pixel sparsity in engineering drawings also restricts the effective featurization of image-based data-driven methods. To overcome these challenges, we propose a deep learning based framework that predicts the semantic type of each vectorized component. Taking a raster image as input, we vectorize all components through thinning, stroke tracing, and cubic bezier fitting. Then a graph of such components is generated based on the connectivity between the components. Finally, a graph convolutional neural network is trained on this graph data to identify the semantic type of each component. We test our framework in the context of semantic segmentation of text, dimension and, contour components in engineering drawings. Results show that our method yields the best performance compared to recent image, and graph-based segmentation methods.

cs.CV↗

arXiv2022

Integrable semi-discretisation of the Drinfel'd--Sokolov hierarchies

We propose a novel semi-discrete Kadomtsev--Petviashvili equation with two discrete and one continuous independent variables, which is integrable in the sense of having the standard and adjoint Lax pairs, from the direct linearisation framework. By performing reductions on the semi-discrete Kadomtsev--Petviashvili equation, new semi-discrete versions of the Drinfel'd--Sokolov hierarchies associated with Kac--Moody Lie algebras $A_r^{(1)}$, $A_{2r}^{(2)}$, $C_r^{(1)}$ and $D_{r+1}^{(2)}$ are successfully constructed. A Lax pair involving the fraction of $\mathbb{Z}_\mathcal{N}$ graded matrices is also found for each of the semi-discrete Drinfel'd--Sokolov equations. Furthermore, the direct linearisation construction guarantees the existence of exact solutions of all the semi-discrete equations discussed in the paper, providing another insight into their integrability in addition to the analysis of Lax pairs.

nlin.SI↗

arXiv2022

A Model-Agnostic Causal Learning Framework for Recommendation using Search Data

Machine-learning based recommender systems(RSs) has become an effective means to help people automatically discover their interests. Existing models often represent the rich information for recommendation, such as items, users, and contexts, as embedding vectors and leverage them to predict users' feedback. In the view of causal analysis, the associations between these embedding vectors and users' feedback are a mixture of the causal part that describes why an item is preferred by a user, and the non-causal part that merely reflects the statistical dependencies between users and items, for example, the exposure mechanism, public opinions, display position, etc. However, existing RSs mostly ignored the striking differences between the causal parts and non-causal parts when using these embedding vectors. In this paper, we propose a model-agnostic framework named IV4Rec that can effectively decompose the embedding vectors into these two parts, hence enhancing recommendation results. Specifically, we jointly consider users' behaviors in search scenarios and recommendation scenarios. Adopting the concepts in causal analysis, we embed users' search behaviors as instrumental variables (IVs), to help decompose original embedding vectors in recommendation, i.e., treatments. IV4Rec then combines the two parts through deep neural networks and uses the combined results for recommendation. IV4Rec is model-agnostic and can be applied to a number of existing RSs such as DIN and NRHUB. Experimental results on both public and proprietary industrial datasets demonstrate that IV4Rec consistently enhances RSs and outperforms a framework that jointly considers search and recommendation.

cs.IR↗

arXiv2022

RecipeSnap -- a lightweight image-to-recipe model

In this paper we want to address the problem of automation for recognition of photographed cooking dishes and generating the corresponding food recipes. Current image-to-recipe models are computation expensive and require powerful GPUs for model training and implementation. High computational cost prevents those existing models from being deployed on portable devices, like smart phones. To solve this issue we introduce a lightweight image-to-recipe prediction model, RecipeSnap, that reduces memory cost and computational cost by more than 90% while still achieving 2.0 MedR, which is in line with the state-of-the-art model. A pre-trained recipe encoder was used to compute recipe embeddings. Recipes from Recipe1M dataset and corresponding recipe embeddings are collected as a recipe library, which are used for image encoder training and image query later. We use MobileNet-V2 as image encoder backbone, which makes our model suitable to portable devices. This model can be further developed into an application for smart phones with a few effort. A comparison of the performance between this lightweight model to other heavy models are presented in this paper. Code, data and models are publicly accessible on github.

cs.CV↗

arXiv2022

SerialTrack: ScalE and Rotation Invariant Augmented Lagrangian Particle Tracking

We present a new particle tracking algorithm to accurately resolve large deformation and rotational motion fields, which takes advantage of both local and global particle tracking algorithms. We call this method the ScalE and Rotation Invariant Augmented Lagrangian Particle Tracking (SerialTrack). This method builds an iterative scale and rotation invariant topology-based feature for each particle within a multi-scale tracking algorithm. The global kinematic compatibility condition is applied as a global augmented Lagrangian constraint to enhance the tracking accuracy. An open source software package implementing this numerical approach to track both 2D and 3D, incremental and cumulative deformation fields is provided.

cs.RO↗

arXiv2021

Linear integral equations and two-dimensional Toda systems

The direct linearisation framework is presented for the two-dimensional Toda equations associated with the infinite-dimensional Lie algebras $A_\infty$, $B_\infty$ and $C_\infty$, as well as the Kac--Moody algebras $A_{r}^{(1)}$, $A_{2r}^{(2)}$, $C_{r}^{(1)}$ and $D_{r+1}^{(2)}$ for arbitrary integers $r\in\mathbb{Z}^+$, from the aspect of a set of linear integral equations in a certain form. Such a scheme not only provides a unified perspective to understand the underlying integrability structure, but also induces the direct linearising type solution potentially leading to the universal solution space, for each class of the two-dimensional Toda system. As particular applications of this framework to the two-dimensional Toda lattices, we rediscover the Lax pairs and the adjoint Lax pairs and simultaneously construct the generalised Cauchy matrix solutions.

nlin.SI↗

arXiv2021

On the origin of GeV spectral break for Fermi blazars: 3C 454.3

The GeV break in spectra of the blazar 3C 454.3 is a special observation feature that has been discovered by the {\it Fermi}-LAT. The origin of the GeV break in the spectra is still under debate. In order to explore the possible source of GeV spectral break in 3C 454.3, a one-zone homogeneous leptonic jet model, as well as the {\it McFit} technique are utilized for fitting the quasi-simultaneous multi-waveband spectral energy distribution (SED) of 3C 454.3. The outside border of the broad-line region (BLR) and inner dust torus are chosen to contribute radiation in the model as external, seed photons to the external-Compton process, considering the observed $γ$-ray radiation. The combination of two components, namely the Compton-scattered BLR and dust torus radiation, assuming a broken power-law distribution of emitted particles, provides a proper fitting to the multi-waveband SED of 3C 454.3 detected 2008 Aug 3 - Sept 2 and explains the GeV spectral break. We propose that the spectral break of 3C 454.3 may originate from an inherent break in the energy distribution of the emitted particles and the Klein-Nishina effect. A comparison is performed between the energy density of the 'external' photon field for the whole BLR $U_{\rm BLR}$ achieved via model fitting and that constrained from the BLR data. The distance from the position of the $γ$-ray radiation area of 3C 454.3 to the central black hole could be constrained at $\sim 0.78$pc ($\sim 4.00 R_{\rm BLR}$, the size of the BLR).

astro-ph.HE↗

arXiv2020

Duluth at SemEval-2020 Task 7: Using Surprise as a Key to Unlock Humorous Headlines

We use pretrained transformer-based language models in SemEval-2020 Task 7: Assessing the Funniness of Edited News Headlines. Inspired by the incongruity theory of humor, we use a contrastive approach to capture the surprise in the edited headlines. In the official evaluation, our system gets 0.531 RMSE in Subtask 1, 11th among 49 submissions. In Subtask 2, our system gets 0.632 accuracy, 9th among 32 submissions.

cs.CL↗

arXiv2020

Confirmed width-Eiso and width-Liso relations in GRB: comparison with the Amati and Yonetoku relations

In this paper, we select a sample including 141 BEST time-integrated F spectra and 145 BEST peak flux P spectra observed by the Konus-Wind with known redshift to recheck the connection between the spectral width and $E_{iso}$ as well as $L_{iso}$. We define six types of absolute spectral widths. It is found that all of the rest-frame absolute spectral widths are strongly positive correlated with $E_{iso}$ as well as $L_{iso}$ for the long burst for both the F and P spectra. All of the short bursts are the outliers for width-$E_{iso}$ relation and most of the short bursts are consistent with the long bursts for the width-$L_{iso}$ relation for both F and P spectra. Moreover, all of the location energy, $E_{2}$ and $E_{1}$, corresponding to various spectral widths are also positive correlated with $E_{iso}$ as well as $L_{iso}$. We compare all of the relations with the Amati and Yonetoku relations and find the width-$E_{iso}$ and width-$L_{iso}$ relations when the widths are at about 90\% maximum of the $EF_{E}$ spectra almost overlap with Amati relation and Yonetoku relation, respectively. The correlations of $E_{2}-E_{iso}$, $E_{1}-E_{iso}$ and $E_{2}-L_{iso}$, $E_{1}-L_{iso}$ when the location energies are at 99\% maximum of the $EF_{E}$ spectra are very close to the Amati and Yonetoku relations, respectively. Therefore, we confirm the existence of tight width-$E_{iso}$ and width-$L_{iso}$ relations for long bursts. We further show that the spectral shape is indeed related to $E_{iso}$ and $L_{iso}$. The Amati and Yonetoku relations are not necessarily the best relationships to relate the energy to the $E_{iso}$ and $L_{iso}$. They may be the special cases of the width-$E_{iso}$ and width-$L_{iso}$ relations or the energy-$E_{iso}$ and energy-$L_{iso}$ relations.

astro-ph.HE↗

arXiv2020

An empirical "high-confidence" candidate zone for $Fermi$ BL Lacertae objects

In the third catalog of active galactic nuclei detected by the $Fermi$ Large Area Telescope Clean (3LAC) sample, there are 402 blazars candidates of uncertain type (BCU). The proposed analysis will help to evaluate the potential optical classification flat spectrum radio quasars (FSRQs) versus BL Lacertae (BL Lacs) objects of BCUs, which can help to understand which is the most elusive class of blazar hidden in the Fermi sample. By studying the 3LAC sample, we found some critical values of $γ$-ray photon spectral index ($Γ_{\rm ph}$), variability index (VI) and radio flux (${\rm F_R}$) of the sources separate known FSRQs and BL Lac objects. We further utilize those values to defined an empirical "high-confidence" candidate zone that can be used to classify the BCUs. Within such a zone ($Γ_{\rm ph}<2.187$, log${\rm F_R}<2.258$ and ${ \rm logVI <1.702}$), we found that 120 BCUs can be classified BL Lac candidates with a higher degree of confidence (with a misjudged rate $<1\%$). Our results suggest that an empirical "high confidence" diagnosis is possible to distinguish the BL Lacs from the Fermi observations based on only on the direct observational data of $Γ_{\rm ph}$, VI and ${\rm F_R}$.

astro-ph.HE↗