arXiv subjects

Zhenyang Huang

Publications and source records attributed to Zhenyang Huang.

11 recordsLinked to original sources

arXiv2026

SMR: Scheduler with Multi-Channel Map-Encoded Reinforcement Learning for Radio Telescopes

Observation scheduling for large single-dish radio telescopes is a multi-objective optimization problem: schedulers must maximize on-source scientific return under strict mechanical and environmental constraints. Previous dynamic scheduling relies on expert-designed heuristics, while existing reinforcement-learning (RL) approaches often struggle with variable-length target lists and lack an intrinsic representation of sky geometry. We present SMR (Scheduler with Map-encoded Reinforcement Learning), which projects discrete targets onto an azimuth--elevation (Az--El) grid in the local horizon frame. The resulting aligned multi-channel sky maps encode target attributes together with direction-dependent cues such as satellite-interference risk and elevation-dependent receiver gain. This representation provides an explicit spatial inductive bias and enables SMR to learn directly from the sky state. Simulations based on real catalogs and site parameters show that, compared with a tuned look-ahead greedy baseline, SMR achieves about a 10\% relative improvement in time utilization by learning non-myopic scheduling strategies. In the full three-channel setting, SMR further achieves joint trade-off among efficiency, interference avoidance, and observation quality, with up to 17\% higher LIER and 54\% higher HGOR relative to an MLP baseline while maintaining higher utilization across both 12 h and 24 h horizons. Overall, SMR provides a simple and extensible way for data-driven single-dish scheduling.

astro-ph.IM↗

arXiv2026

Inversion of Hydrogen-rich Atmosphere and Water Content for GJ 486b

GJ~486b is a close-in planet orbiting an M dwarf and is therefore expected to have undergone strong atmospheric escape. Motivated by theoretical and observational studies on the constraints of its water and atmosphere, we investigate which combinations of an primordial hydrogen-rich atmosphere and water inventory could fit the current water content implied by bulk density measurements. We model the atmosphere escape using VPLanet, following the loss of an initial hydrogen-rich atmosphere and the subsequent escape of a water-dominated atmosphere. By scanning a broad parameter space across different stellar ages, we invert for the initial hydrogen-rich atmospheric mass and water inventory consistent with the current constraints. Our results reveal a strong degeneracy between the water reservoir and the initial hydrogen-rich atmosphere. Even a modest hydrogen-rich atmosphere can significantly delay early escape of the water and reduce the water inventory required to reproduce the current water content. We also find that the inferred initial conditions are also strongly age dependent. Incorporating a planet formation dataset as a prior, we derive a probabilistic constraint on the host star age, yielding an expected age of $2.90^{+2.47}_{-2.27}$~Gyr, which is consistent with the results obtained from other methods to determine M dwarf ages.

astro-ph.EP↗

arXiv2025

ReasonCD: A Multimodal Reasoning Large Model for Implicit Change-of-Interest Semantic Mining

Remote sensing image change detection is one of the fundamental tasks in remote sensing intelligent interpretation. Its core objective is to identify changes within change regions of interest (CRoI). Current multimodal large models encode rich human semantic knowledge, which is utilized for guidance in tasks such as remote sensing change detection. However, existing methods that use semantic guidance for detecting users' CRoI overly rely on explicit textual descriptions of CRoI, leading to the problem of near-complete performance failure when presented with implicit CRoI textual descriptions. This paper proposes a multimodal reasoning change detection model named ReasonCD, capable of mining users' implicit task intent. The model leverages the powerful reasoning capabilities of pre-trained large language models to mine users' implicit task intents and subsequently obtains different change detection results based on these intents. Experiments on public datasets demonstrate that the model achieves excellent change detection performance, with an F1 score of 92.1\% on the BCDD dataset. Furthermore, to validate its superior reasoning functionality, this paper annotates a subset of reasoning data based on the SECOND dataset. Experimental results show that the model not only excels at basic reasoning-based change detection tasks but can also explain the reasoning process to aid human decision-making.

cs.CV↗

arXiv2025

A Core-Collapse Supernova Neutrino Parameterization with Enhanced Physical Interpretability

We introduce a novel parameterization of supernova neutrino energy spectra with a clear physical motivation. Its central parameter, $\tau(t)$, quantifies the characteristic thermal-diffusion area during the explosion. When applied to the historic SN1987A data, this parameterization yields statistically significant fits and provides robust constraints on the unobserved low-energy portion of the spectrum. Beyond this specific application, we demonstrate the model's power on a suite of 3D core-collapse supernova simulations, finding that the temporal evolution of $\tau(t)$ distinctly separates successful from failed explosions. Furthermore, we constrain the progenitor mass of SN 1987A to approximately 19 solar masses by applying Smoothed Isotonic Regression, while noting the sensitivity of this estimate to observational uncertainties. Moreover, in these simulations, $\tau(t)$ and the gravitational-wave strain amplitude display a strong, synergistic co-evolution, directly linking the engine's energetic evolution to its geometric asymmetry. This implies that the thermodynamic state of the explosion is imprinted not only on the escaping neutrino flux, but also recorded in the shape of the energy spectrum. Our framework therefore offers a valuable tool for decoding the detailed core dynamics and multi-messenger processes of future galactic supernovae.

astro-ph.HE↗

arXiv2025

An Interpretable AI Framework to Disentangle Self-Interacting and Cold Dark Matter in Galaxy Clusters: The CKAN Approach

Convolutional neural networks have shown their ability to differentiate between self-interacting dark matter (SIDM) and cold dark matter (CDM) on galaxy cluster scales. However, their large parameter counts and ''black-box'' nature make it difficult to assess whether their decisions adhere to physical principles. To address this issue, we have built a Convolutional Kolmogorov-Arnold Network (CKAN) that reduces parameter count and enhances interpretability, and propose a novel analytical framework to understand the network's decision-making process. With this framework, we leverage our network to qualitatively assess the offset between the dark matter distribution center and the galaxy cluster center, as well as the size of heating regions in different models. These findings are consistent with current theoretical predictions and show the reliability and interpretability of our network. By combining network interpretability with unseen test results, we also estimate that for SIDM in galaxy clusters, the minimum cross-section $(\sigma/m)_{\mathrm{th}}$ required to reliably identify its collisional nature falls between $0.1\,\mathrm{cm}^2/\mathrm{g}$ and $0.3\,\mathrm{cm}^2/\mathrm{g}$. Moreover, CKAN maintains robust performance under simulated JWST and Euclid noise, highlighting its promise for application to forthcoming observational surveys.

astro-ph.IM↗

arXiv2025

Application of interpretable data-driven methods for the reconstruction of supernova neutrino energy spectra following fast neutrino flavor conversions

Neutrinos can experience fast flavor conversions (FFCs) in highly dense astrophysical environments, such as core-collapse supernovae and neutron star mergers, potentially affecting energy transport and other processes. Simulating fast flavor conversions under realistic astrophysical conditions requires substantial computational resources and poses significant analytical challenges. While machine learning methods such as multilayer perceptrons have been used to accurately predict the asymptotic outcomes of FFCs, their "black-box" nature limits the extraction of direct physical insight. To mitigate this limitation, we employ two distinct interpretable machine learning frameworks, Kolmogorov-Arnold Networks (KANs) and Sparse Identification of Nonlinear Dynamics (SINDy), to learn interpretable surrogates for the asymptotic input-output mapping from an FFC simulation dataset. Our analysis reveals a fundamental trade-off between predictive accuracy and model simplicity. KANs demonstrate high fidelity in reconstructing post-conversion neutrino energy spectra, achieving accuracies of up to 90%. In contrast, SINDy yields a low-rank, compact closed-form approximation of the input-output mapping, at the expense of some predictive accuracy. Critically, using these structured and sparse surrogates as diagnostic tools, we identify that the system's evolution is most sensitive to the initial number density of heavy-lepton neutrinos when FFCs are triggered, compared with other physical quantities. Ultimately, this work provides a methodological framework for interpretable machine learning that supports genuine data-driven scientific discovery in astronomy and astrophysics, going beyond prediction alone.

astro-ph.HE↗

arXiv2025

Hunting Hidden Axion Signals in Pulsar Dispersion Measurements with Machine Learning

In axion models, interactions between axions and electromagnetic waves induce frequency-dependent time delays determined by the axion mass and decay constant. These small delays are difficult to detect, limiting the effectiveness of traditional methods. We compute such delays under realistic radio telescope conditions and identify a prominent dispersive feature near half the axion mass, which appears non-divergent within the limits of observational resolution. Based on this, we develop a machine learning method that achieves 90\% classification accuracy and demonstrates well performance in low signal-to-noise regimes. The method's robustness is confirmed against false positives using both simulated noisy data and real-world, known-null observations. Future improvements in optical clock precision and telescope bandwidth, particularly with instruments such as the Qitai Radio Telescope, may enhance constraints on the axion decay constant by up to four orders of magnitude in the $10^{-6} \sim 10^{-4}$ eV mass range.

astro-ph.HE↗

arXiv2025

Dark Matter (S)pins the Planet

Dark matter heating in planets has been proposed as a potential probe for dark matter detection. Assuming near-equilibrium conditions, we find that the energy input from dark matter raises planetary temperatures and accelerates rotation. The distribution of energy between heating and rotational acceleration depends on both planetary properties and external inputs, suggesting that previous studies may have overestimated the heating contribution. At high dark matter densities, planetary rotation stabilizes earlier and becomes primarily governed by dark matter effects.

astro-ph.HE↗

arXiv2024

SeFi-CD: A Semantic First Change Detection Paradigm That Can Detect Any Change You Want

The existing change detection(CD) methods can be summarized as the visual-first change detection (ViFi-CD) paradigm, which first extracts change features from visual differences and then assigns them specific semantic information. However, CD is essentially dependent on change regions of interest (CRoIs), meaning that the CD results are directly determined by the semantics changes of interest, making its primary image factor semantic of interest rather than visual. The ViFi-CD paradigm can only assign specific semantics of interest to specific change features extracted from visual differences, leading to the inevitable omission of potential CRoIs and the inability to adapt to different CRoI CD tasks. In other words, changes in other CRoIs cannot be detected by the ViFi-CD method without retraining the model or significantly modifying the method. This paper introduces a new CD paradigm, the semantic-first CD (SeFi-CD) paradigm. The core idea of SeFi-CD is to first perceive the dynamic semantics of interest and then visually search for change features related to the semantics. Based on the SeFi-CD paradigm, we designed Anything You Want Change Detection (AUWCD). Experiments on public datasets demonstrate that the AUWCD outperforms the current state-of-the-art CD methods, achieving an average F1 score 5.01\% higher than that of these advanced supervised baselines on the SECOND dataset, with a maximum increase of 13.17\%. The proposed SeFi-CD offers a novel CD perspective and approach.

cs.CV↗

arXiv2024

MFDS-Net: Multi-Scale Feature Depth-Supervised Network for Remote Sensing Change Detection with Global Semantic and Detail Information

Change detection as an interdisciplinary discipline in the field of computer vision and remote sensing at present has been receiving extensive attention and research. Due to the rapid development of society, the geographic information captured by remote sensing satellites is changing faster and more complex, which undoubtedly poses a higher challenge and highlights the value of change detection tasks. We propose MFDS-Net: Multi-Scale Feature Depth-Supervised Network for Remote Sensing Change Detection with Global Semantic and Detail Information (MFDS-Net) with the aim of achieving a more refined description of changing buildings as well as geographic information, enhancing the localisation of changing targets and the acquisition of weak features. To achieve the research objectives, we use a modified ResNet_34 as backbone network to perform feature extraction and DO-Conv as an alternative to traditional convolution to better focus on the association between feature information and to obtain better training results. We propose the Global Semantic Enhancement Module (GSEM) to enhance the processing of high-level semantic information from a global perspective. The Differential Feature Integration Module (DFIM) is proposed to strengthen the fusion of different depth feature information, achieving learning and extraction of differential features. The entire network is trained and optimized using a deep supervision mechanism. The experimental outcomes of MFDS-Net surpass those of current mainstream change detection networks. On the LEVIR dataset, it achieved an F1 score of 91.589 and IoU of 84.483, on the WHU dataset, the scores were F1: 92.384 and IoU: 86.807, and on the GZ-CD dataset, the scores were F1: 86.377 and IoU: 76.021. The code is available at https://github.com/AOZAKIiii/MFDS-Net

cs.CV↗

arXiv2024

HSONet:A Siamese foreground association-driven hard case sample optimization network for high-resolution remote sensing image change detection

In the later training stages, further improvement of the models ability to determine changes relies on how well the change detection (CD) model learns hard cases; however, there are two additional challenges to learning hard case samples: (1) change labels are limited and tend to pointer only to foreground targets, yet hard case samples are prevalent in the background, which leads to optimizing the loss function focusing on the foreground targets and ignoring the background hard cases, which we call imbalance. (2) Complex situations, such as light shadows, target occlusion, and seasonal changes, induce hard case samples, and in the absence of both supervisory and scene information, it is difficult for the model to learn hard case samples directly to accurately obtain the feature representations of the change information, which we call missingness. We propose a Siamese foreground association-driven hard case sample optimization network (HSONet). To deal with this imbalance, we propose an equilibrium optimization loss function to regulate the optimization focus of the foreground and background, determine the hard case samples through the distribution of the loss values, and introduce dynamic weights in the loss term to gradually shift the optimization focus of the loss from the foreground to the background hard cases as the training progresses. To address this missingness, we understand hard case samples with the help of the scene context, propose the scene-foreground association module, use potential remote sensing spatial scene information to model the association between the target of interest in the foreground and the related context to obtain scene embedding, and apply this information to the feature reinforcement of hard cases. Experiments on four public datasets show that HSONet outperforms current state-of-the-art CD methods, particularly in detecting hard case samples.

cs.CV↗