arXiv subjects

Duo Xu

Publications and source records attributed to Duo Xu.

At least 37 records · Page 2Linked to original sources

arXiv2025

ChartM$^3$: A Multi-Stage Code-Driven Pipeline for Constructing Multi-Dimensional and Multi-Step Visual Reasoning Data in Chart Comprehension

Complex chart understanding tasks demand advanced visual recognition and reasoning capabilities from multimodal large language models (MLLMs). However, current research provides limited coverage of complex chart scenarios and computation-intensive reasoning tasks prevalent in real-world applications. This study proposes an automated multi-stage code-driven pipeline for systematically generating visual reasoning datasets to address these limitations. The pipeline integrates retrieval-augmented generation (RAG) to retrieve professional chart templates and employs chain-of-thought (CoT) strategies to generate reasoning codes that simulate real data distributions, thereby driving chart rendering and question-related statistical computations. Through model-based evaluation, the pipeline enhances chart diversity and data quality. Using this framework, we construct ChartM$^3$, a multi-dimensional and multi-step dataset containing 38K charts and 142K Q&A pairs for training, along with 2,871 high-quality evaluation samples for enabling practical performance assessment. Supervised fine-tuning (SFT) and reinforcement learning (RL) experiments demonstrate that our dataset significantly improves reasoning capabilities and cross-domain generalization performance, enabling smaller models to achieve performance comparable to larger-scale models in complex chart comprehension.

cs.CV↗

arXiv2025

Discovering "Words" in Music: Unsupervised Learning of Compositional Sparse Code for Symbolic Music

This paper presents an unsupervised machine learning algorithm that identifies recurring patterns -- referred to as ``music-words'' -- from symbolic music data. These patterns are fundamental to musical structure and reflect the cognitive processes involved in composition. However, extracting these patterns remains challenging because of the inherent semantic ambiguity in musical interpretation. We formulate the task of music-word discovery as a statistical optimization problem and propose a two-stage Expectation-Maximization (EM)-based learning framework: 1. Developing a music-word dictionary; 2. Reconstructing the music data. When evaluated against human expert annotations, the algorithm achieved an Intersection over Union (IoU) score of 0.61. Our findings indicate that minimizing code length effectively addresses semantic ambiguity, suggesting that human optimization of encoding systems shapes musical semantics. This approach enables computers to extract ``basic building blocks'' from music data, facilitating structural analysis and sparse encoding. The method has two primary applications. First, in AI music, it supports downstream tasks such as music generation, classification, style transfer, and improvisation. Second, in musicology, it provides a tool for analyzing compositional patterns and offers insights into the principle of minimal encoding across diverse musical styles and composers.

cs.SD↗

arXiv2025

Experimental study of turbulent mixing in a T-shaped mixer

One of the most widespread canonical devices for fluid mixing is the T-shaped mixer, in which two opposing miscible liquid streams meet at a junction and then mix along a main channel. Laminar steady and time-periodic flows in T-shaped mixers have been thoroughly studied, but turbulent flows have received much less scrutiny despite their prevalence in applications. We here introduce a novel experimental setup with a hydraulic diameter of four centimetres that enables the optical study of turbulent mixing at small scales. Using this setup, we perform two-dimensional particle image velocimetry and planar laser-induced fluorescence measurements. First, we successfully replicate characteristic flow regimes observed in micro-scale T-shaped mixers at low Reynolds numbers. We then focus on the turbulent regime and characterize the turbulent kinetic energy and dissipation along the mixing channel. Further, we measure the scalar concentration variance and its corresponding probability density function and spectra. The latter exhibits an incipient Batchelor scaling. We estimate the mechanical-to-scalar timescale ratio and examine the link between the turbulent velocity and scalar fields. The measurement data are compared with model predictions and correlations used in engineering practice, and with our own direct numerical simulations performed with a spectral-element code.

physics.flu-dyn↗

arXiv2025

A framework for learning symbolic turbulence models from indirect observation data via neural networks and feature importance analysis

Learning symbolic turbulence models from indirect observation data is of significant interest as it not only improves the accuracy of posterior prediction but also provides explicit model formulations with good interpretability. However, it typically resorts to gradient-free evolutionary algorithms, which can be relatively inefficient compared to gradient-based approaches, particularly when the Reynolds-averaged Navier-Stokes (RANS) simulations are involved in the training process. In view of this difficulty, we propose a framework that uses neural networks and the associated feature importance analysis to improve the efficiency of symbolic turbulence modeling. In doing so, the gradient-based method can be used to efficiently learn neural network-based representations of Reynolds stress from indirect data, which is further transformed into simplified mathematical expressions with symbolic regression. Moreover, feature importance analysis is introduced to accelerate the convergence of symbolic regression by excluding insignificant input features. The proposed training strategy is tested in the flow in a square duct, where it correctly learns underlying analytic models from indirect velocity data. Further, the method is applied in the flow over the periodic hills, demonstrating that the feature importance analysis can significantly improve the training efficiency and learn symbolic turbulence models with satisfactory generalizability.

physics.flu-dyn↗

arXiv2025

ImplicitCell: Resolution Cell Modeling of Joint Implicit Volume Reconstruction and Pose Refinement in Freehand 3D Ultrasound

Freehand 3D ultrasound enables volumetric imaging by tracking a conventional ultrasound probe during freehand scanning, offering enriched spatial information that improves clinical diagnosis. However, the quality of reconstructed volumes is often compromised by tracking system noise and irregular probe movements, leading to artifacts in the final reconstruction. To address these challenges, we propose ImplicitCell, a novel framework that integrates Implicit Neural Representation (INR) with an ultrasound resolution cell model for joint optimization of volume reconstruction and pose refinement. Three distinct datasets are used for comprehensive validation, including phantom, common carotid artery, and carotid atherosclerosis. Experimental results demonstrate that ImplicitCell significantly reduces reconstruction artifacts and improves volume quality compared to existing methods, particularly in challenging scenarios with noisy tracking data. These improvements enhance the clinical utility of freehand 3D ultrasound by providing more reliable and precise diagnostic information.

eess.IV↗

arXiv2025

First Search for Pulsed CH Maser Emission Stimulated by a Pulsar

We present the first search for pulsed CH maser emission potentially stimulated by PSR J1644$-$4559, conducted using the ultra-wide-bandwidth low-frequency receiver on Murriyang, CSIRO's Parkes Radio Telescope. Observations targeted three CH $Λ$-doublet transitions at 3264, 3335, and 3349 MHz, with a variability timescale of 78 ms. We detected ten CH emission features at 3335 and 3349 MHz, and seven features at 3264 MHz, during both pulsar-ON and pulsar-OFF phases. The observed velocities align with the OH emission and absorption reported by a previous study, suggesting a close spatial association between CH and OH molecules. The derived column densities for CH clouds within the Parkes beam range from $0.05$ to $9.8 \times 10^{13}$ cm$^{-2}$, indicating that these clouds are likely in diffuse and translucent states. Upper limits for CH column densities within the pulsar beam ranged from $0.3$ to $9.8 \times 10^{13}$ cm$^{-2}$. Comparison of these column densities suggests that CH clouds may exhibit clumpiness and substructure. No significant stimulated emission feature was detected in the optical depth spectra. Additionally, as part of our search for pulsed stimulated emission, we investigated the potential CH absorption of the pulsar signal and found none, in agreement with astrophysical expectations. The upper limits for the potential maser amplification factors towards PSR J1644$-$4559 at 3264, 3335, and 3349 MHz are 1.014, 1.009, and 1.009, respectively. This study demonstrates the feasibility of detecting pulsed CH maser emission in the interstellar medium stimulated by pulsar photons.

astro-ph.GA↗

arXiv2025

Exploring Magnetic Fields in Molecular Clouds through Denoising Diffusion Probabilistic Models

Accurately measuring magnetic field strength in the interstellar medium, including giant molecular clouds (GMCs), remains a significant challenge. We present a machine learning approach using Denoising Diffusion Probabilistic Models (DDPMs) to estimate magnetic field strength from synthetic observables such as column density, dust continuum polarization vector orientation angles, and line-of-sight (LOS) nonthermal velocity dispersion. We trained three versions of the DDPM model: the 1-channel DDPM (using only column density), the 2-channel DDPM (incorporating both column density and polarization angles), and the 3-channel DDPM (which combines column density, polarization angles, and LOS nonthermal velocity dispersion). We assessed the models on both synthetic test samples and new simulation data that were outside the training set's distribution. The 3-channel DDPM consistently outperformed both the other DDPM variants and the power-law fitting approach based on column density alone, demonstrating its robustness in handling previously unseen data. Additionally, we compared the performance of the Davis-Chandrasekhar-Fermi (DCF) methods, both classical and modified, to the DDPM predictions. The classical DCF method overestimated the magnetic field strength by approximately an order of magnitude. Although the modified DCF method showed improvement over the classical version, it still fell short of the precision achieved by the 3-channel DDPM.

astro-ph.GA↗

arXiv2024

Learning Hidden Subgoals under Temporal Ordering Constraints in Reinforcement Learning

In real-world applications, the success of completing a task is often determined by multiple key steps which are distant in time steps and have to be achieved in a fixed time order. For example, the key steps listed on the cooking recipe should be achieved one-by-one in the right time order. These key steps can be regarded as subgoals of the task and their time orderings are described as temporal ordering constraints. However, in many real-world problems, subgoals or key states are often hidden in the state space and their temporal ordering constraints are also unknown, which make it challenging for previous RL algorithms to solve this kind of tasks. In order to address this issue, in this work we propose a novel RL algorithm for {\bf l}earning hidden {\bf s}ubgoals under {\bf t}emporal {\bf o}rdering {\bf c}onstraints (LSTOC). We propose a new contrastive learning objective which can effectively learn hidden subgoals (key states) and their temporal orderings at the same time, based on first-occupancy representation and temporal geometric sampling. In addition, we propose a sample-efficient learning strategy to discover subgoals one-by-one following their temporal order constraints by building a subgoal tree to represent discovered subgoals and their temporal ordering relationships. Specifically, this tree can be used to improve the sample efficiency of trajectory collection, fasten the task solving and generalize to unseen tasks. The LSTOC framework is evaluated on several environments with image-based observations, showing its significant improvement over baseline methods.

cs.LG↗

arXiv2024

Generalization of Compositional Tasks with Logical Specification via Implicit Planning

In this study, we address the challenge of learning generalizable policies for compositional tasks defined by logical specifications. These tasks consist of multiple temporally extended sub-tasks. Due to the sub-task inter-dependencies and sparse reward issue in long-horizon tasks, existing reinforcement learning (RL) approaches, such as task-conditioned and goal-conditioned policies, continue to struggle with slow convergence and sub-optimal performance in generalizing to compositional tasks. To overcome these limitations, we introduce a new hierarchical RL framework that enhances the efficiency and optimality of task generalization. At the high level, we present an implicit planner specifically designed for generalizing compositional tasks. This planner selects the next sub-task and estimates the multi-step return for completing the remaining task to complete from the current state. It learns a latent transition model and performs planning in the latent space by using a graph neural network (GNN). Subsequently, the high-level planner's selected sub-task guides the low-level agent to effectively handle long-horizon tasks, while the multi-step return encourages the low-level policy to account for future sub-task dependencies, enhancing its optimality. We conduct comprehensive experiments to demonstrate the framework's advantages over previous methods in terms of both efficiency and optimality.

cs.AI↗

arXiv2024

LLM-Augmented Symbolic Reinforcement Learning with Landmark-Based Task Decomposition

One of the fundamental challenges in reinforcement learning (RL) is to take a complex task and be able to decompose it to subtasks that are simpler for the RL agent to learn. In this paper, we report on our work that would identify subtasks by using some given positive and negative trajectories for solving the complex task. We assume that the states are represented by first-order predicate logic using which we devise a novel algorithm to identify the subtasks. Then we employ a Large Language Model (LLM) to generate first-order logic rule templates for achieving each subtask. Such rules were then further fined tuned to a rule-based policy via an Inductive Logic Programming (ILP)-based RL agent. Through experiments, we verify the accuracy of our algorithm in detecting subtasks which successfully detect all of the subtasks correctly. We also investigated the quality of the common-sense rules produced by the language model to achieve the subtasks. Our experiments show that our LLM-guided rule template generation can produce rules that are necessary for solving a subtask, which leads to solving complex tasks with fewer assumptions about predefined first-order logic predicates of the environment.

cs.AI↗

arXiv2024

Constraints-Informed Neural-Laguerre Approximation of Nonlinear MPC with Application in Power Electronics

This paper considers learning online (implicit) nonlinear model predictive control (MPC) laws using neural networks and Laguerre functions. Firstly, we parameterize the control sequence of nonlinear MPC using Laguerre functions, which typically yields a smoother control law compared to the original nonlinear MPC law. Secondly, we employ neural networks to learn the coefficients of the Laguerre nonlinear MPC solution, which comes with several benefits, namely the dimension of the learning space is dictated by the number of Laguerre functions and the complete predicted input sequence can be used to learn the coefficients. To mitigate constraints violation for neural approximations of nonlinear MPC, we develop a constraints-informed loss function that penalizes the violation of polytopic state constraints during learning. Box input constraints are handled by using a clamp function in the output layer of the neural network. We demonstrate the effectiveness of the developed framework on a nonlinear buck-boost converter model with sampling rates in the sub-millisecond range, where online nonlinear MPC would not be able to run in real time. The developed constraints-informed neural-Laguerre approximation yields similar performance with long-horizon online nonlinear MPC, but with execution times of a few microseconds, as validated on a field-programmable gate array (FPGA) platform.

math.OC↗

arXiv2024

Finite Control Set Model Predictive Control with Limit Cycle Stability Guarantees

This paper considers the design of finite control set model predictive control (FCS-MPC) for discrete-time switched affine systems. Existing FCS-MPC methods typically pursue practical stability guarantees, which ensure convergence to a bounded invariant set that contains a desired steady state. As such, current FCS-MPC methods result in unpredictable steady-state behavior due to arbitrary switching among the available finite control inputs. Motivated by this, we present a FCS-MPC design that aims to stabilize a steady-state limit cycle compatible with a desired output reference via a suitable cost function. We provide conditions in terms of periodic terminal costs and finite control set control laws that guarantee asymptotic stability of the developed limit cycle FCS-MPC algorithm. Moreover, we develop conditions for recursive feasibility of limit cycle FCS-MPC in terms of periodic terminal sets and we provide systematic methods for computing ellipsoidal and polytopic periodically invariant sets that contain a desired steady-state limit cycle. Compared to existing periodic terminal ingredients for tracking MPC with a continuous control set, we design and compute terminal ingredients using a finite control set. The developed methodology is validated on switched systems and power electronics benchmark examples.

math.OC↗

arXiv2024

Surveying Image Segmentation Approaches in Astronomy

Image segmentation plays a critical role in unlocking the mysteries of the universe, providing astronomers with a clearer perspective on celestial objects within complex astronomical images and data cubes. Manual segmentation, while traditional, is not only time-consuming but also susceptible to biases introduced by human intervention. As a result, automated segmentation methods have become essential for achieving robust and consistent results in astronomical studies. This review begins by summarizing traditional and classical segmentation methods widely used in astronomical tasks. Despite the significant improvements these methods have brought to segmentation outcomes, they fail to meet astronomers' expectations, requiring additional human correction, further intensifying the labor-intensive nature of the segmentation process. The review then focuses on the transformative impact of machine learning, particularly deep learning, on segmentation tasks in astronomy. It introduces state-of-the-art machine learning approaches, highlighting their applications and the remarkable advancements they bring to segmentation accuracy in both astronomical images and data cubes. As the field of machine learning continues to evolve rapidly, it is anticipated that astronomers will increasingly leverage these sophisticated techniques to enhance segmentation tasks in their research projects. In essence, this review serves as a comprehensive guide to the evolution of segmentation methods in astronomy, emphasizing the transition from classical approaches to cutting-edge machine learning methodologies. We encourage astronomers to embrace these advancements, fostering a more streamlined and accurate segmentation process that aligns with the ever-expanding frontiers of astronomical exploration.

astro-ph.IM↗

arXiv2024

CityLLaVA: Efficient Fine-Tuning for VLMs in City Scenario

In the vast and dynamic landscape of urban settings, Traffic Safety Description and Analysis plays a pivotal role in applications ranging from insurance inspection to accident prevention. This paper introduces CityLLaVA, a novel fine-tuning framework for Visual Language Models (VLMs) designed for urban scenarios. CityLLaVA enhances model comprehension and prediction accuracy through (1) employing bounding boxes for optimal visual data preprocessing, including video best-view selection and visual prompt engineering during both training and testing phases; (2) constructing concise Question-Answer sequences and designing textual prompts to refine instruction comprehension; (3) implementing block expansion to fine-tune large VLMs efficiently; and (4) advancing prediction accuracy via a unique sequential questioning-based prediction augmentation. Demonstrating top-tier performance, our method achieved a benchmark score of 33.4308, securing the leading position on the leaderboard. The code can be found: https://github.com/alibaba/AICITY2024_Track2_AliOpenTrek_CityLLaVA

cs.CV↗

arXiv2024

Disk Wind Feedback from High-mass Protostars. III. Synthetic CO Line Emission

To test theoretical models of massive star formation it is important to compare their predictions with observed systems. To this end, we conduct CO molecular line radiative transfer post-processing of 3D magneto-hydrodynamic (MHD) simulations of various stages in the evolutionary sequence of a massive protostellar core, including its infall envelope and disk wind outflow. Synthetic position-position-velocity (PPV) cubes of various transitions of CO, 13CO, and C18O emission are generated. We also carry out simulated Atacama Large Millimeter/submillimeter Array (ALMA) observations of this emission. We compare the mass, momentum and kinetic energy estimates obtained from molecular lines to the true values, finding that the mass and momentum estimates can have uncertainties of up to a factor of four. However, the kinetic energy estimated from molecular lines is more significantly underestimated. Additionally, we compare the mass outflow rate and momentum outflow rate obtained from the synthetic spectra with the true values. Finally, we compare the synthetic spectra with real examples of ALMA-observed protostars and determine the best fitting protostellar masses and outflow inclination angles. We then calculate the mass outflow rate and momentum outflow rate for these sources, finding that both rates agree with theoretical protostellar evolutionary tracks.

astro-ph.GA↗

arXiv2024

An Effective Learning Management System for Revealing Student Performance Attributes

A learning management system streamlines the management of the teaching process in a centralized place, recording, tracking, and reporting the delivery of educational courses and student performance. Educational knowledge discovery from such an e-learning system plays a crucial role in rule regulation, policy establishment, and system development. However, existing LMSs do not have embedded mining modules to directly extract knowledge. As educational modes become more complex, educational data mining efficiency from those heterogeneous student learning behaviours is gradually degraded. Therefore, an LMS incorporated with an advanced educational mining module is proposed in this study, as a means to mine efficiently from student performance records to provide valuable insights for educators in helping plan effective learning pedagogies, improve curriculum design, and guarantee quality of teaching. Through two illustrative case studies, experimental results demonstrate increased mining efficiency of the proposed mining module without information loss compared to classic educational mining algorithms. The mined knowledge reveals a set of attributes that significantly impact student academic performance, and further classification evaluation validates the identified attributes. The design and application of such an effective LMS can enable educators to learn from past student performance experiences, empowering them to guide and intervene with students in time, and eventually improve their academic success.

cs.CY↗

arXiv2024

An Order-Complexity Aesthetic Assessment Model for Aesthetic-aware Music Recommendation

Computational aesthetic evaluation has made remarkable contribution to visual art works, but its application to music is still rare. Currently, subjective evaluation is still the most effective form of evaluating artistic works. However, subjective evaluation of artistic works will consume a lot of human and material resources. The popular AI generated content (AIGC) tasks nowadays have flooded all industries, and music is no exception. While compared to music produced by humans, AI generated music still sounds mechanical, monotonous, and lacks aesthetic appeal. Due to the lack of music datasets with rating annotations, we have to choose traditional aesthetic equations to objectively measure the beauty of music. In order to improve the quality of AI music generation and further guide computer music production, synthesis, recommendation and other tasks, we use Birkhoff's aesthetic measure to design a aesthetic model, objectively measuring the aesthetic beauty of music, and form a recommendation list according to the aesthetic feeling of music. Experiments show that our objective aesthetic model and recommendation method are effective.

cs.CV↗

arXiv2024

Polarized Light from Massive Protoclusters (POLIMAP). I. Dissecting the role of magnetic fields in the massive infrared dark cloud G28.37+0.07

Magnetic fields may play a crucial role in setting the initial conditions of massive star and star cluster formation. To investigate this, we report SOFIA-HAWC+ $214\:μ$m observations of polarized thermal dust emission and high-resolution GBT-Argus C$^{18}$O(1-0) observations toward the massive Infrared Dark Cloud (IRDC) G28.37+0.07. Considering the local dispersion of $B$-field orientations, we produce a map of $B$-field strength of the IRDC, which exhibits values between $\sim0.03 - 1\:$mG based on a refined Davis-Chandrasekhar-Fermi (r-DCF) method proposed by Skalidis \& Tassis. Comparing to a map of inferred density, the IRDC exhibits a $B-n$ relation with a power law index of $0.51\pm0.02$, which is consistent with a scenario of magnetically-regulated anisotropic collapse. Consideration of the mass-to-flux ratio map indicates that magnetic fields are dynamically important in most regions of the IRDC. A virial analysis of a sample of massive, dense cores in the IRDC, including evaluation of magnetic and kinetic internal and surface terms, indicates consistency with virial equilibrium, sub-Alfvénic conditions and a dominant role for $B-$fields in regulating collapse. A clear alignment of magnetic field morphology with direction of steepest column density gradient is also detected. However, there is no preferred orientation of protostellar outflow directions with the $B-$field. Overall, these results indicate that magnetic fields play a crucial role in regulating massive star and star cluster formation and so need to be accounted for in theoretical models of these processes.

astro-ph.GA↗