arXiv subjects

Jiaxuan Li

Publications and source records attributed to Jiaxuan Li.

At least 55 records · Page 3Linked to original sources

arXiv2025

The DREAMS Project: Disentangling the Impact of Halo-to-Halo Variance and Baryonic Feedback on Milky Way Satellite Galaxies

We analyze the properties of satellite galaxies around 1,024 Milky Way-mass hosts from the DREAMS Project, simulated within a $Λ$CDM cosmology. Utilizing the TNG galaxy-formation model, the DREAMS simulations incorporate both baryonic physics and cosmological uncertainties for a large sample of galaxies with diverse environments and formation histories. We investigate the relative impact of the physical uncertainty from the galaxy-formation model on predicted satellite properties using four metrics: the satellite stellar mass function, radial distribution, inner slope of dark matter density profile, and stellar half-light radius. We compare these predictions to observations from the SAGA Survey and the DREAMS N-body simulations and find that uncertainties from baryonic physics modeling are subdominant to the scatter arising from halo-to-halo variance. Where baryonic modeling does affect satellites, the supernova wind energy has the largest effect on the satellite properties that we investigate. Specifically, increased supernova wind energy suppresses the stellar mass of satellites and results in more extended stellar half-light radii. The adopted wind speed has only a minor impact, and other astrophysical and cosmological parameters show no measurable effect. Our findings highlight the robustness of satellite properties against uncertainties in baryonic physics modeling.

astro-ph.GA↗

arXiv2025

CoMA: Complementary Masking and Hierarchical Dynamic Multi-Window Self-Attention in a Unified Pre-training Framework

Masked Autoencoders (MAE) achieve self-supervised learning of image representations by randomly removing a portion of visual tokens and reconstructing the original image as a pretext task, thereby significantly enhancing pretraining efficiency and yielding excellent adaptability across downstream tasks. However, MAE and other MAE-style paradigms that adopt random masking generally require more pre-training epochs to maintain adaptability. Meanwhile, ViT in MAE suffers from inefficient parameter use due to fixed spatial resolution across layers. To overcome these limitations, we propose the Complementary Masked Autoencoders (CoMA), which employ a complementary masking strategy to ensure uniform sampling across all pixels, thereby improving effective learning of all features and enhancing the model's adaptability. Furthermore, we introduce DyViT, a hierarchical vision transformer that employs a Dynamic Multi-Window Self-Attention (DM-MSA), significantly reducing the parameters and FLOPs while improving fine-grained feature learning. Pre-trained on ImageNet-1K with CoMA, DyViT matches the downstream performance of MAE using only 12% of the pre-training epochs, demonstrating more effective learning. It also attains a 10% reduction in pre-training time per epoch, further underscoring its superior pre-training efficiency.

cs.CV↗

arXiv2025

Testing the Stellar Feedback-driven Breathing Mode in Low-mass Galaxies with Gas Kinematics

Hydrodynamic simulations have proposed that stellar feedback and bursty star-formation can produce dark matter cores in low-mass galaxies. A key prediction is that feedback-driven gas outflow and inflow cycles can lead to ``breathing modes'' (rapid fluctuations in the global gravitational potential) which drive correlated variations in galaxy size, kinematics, and star-formation rate. In this paper, we test the dynamical effects of feedback-driven breathing modes using a sample of 103 star-forming low-mass galaxies with stellar masses between $7.9<\rm \log M_*/M_\odot<9.6$ and $0.02<z<0.19$. We measure ionized gas velocity dispersions from H$α$ emission lines and compare them to mock observations from the FIRE-2 simulations. We compare gas velocity dispersions ($\rm σ_{gas}$), stellar masses, and specific star-formation rates (sSFR). We find a positive correlation between gas velocity dispersion residuals at fixed stellar masses ($\rm Δσ_{gas}$) and sSFR in both data and simulations. However, the relation is tighter in FIRE-2 compared to the data. FIRE-2 produces more low-sSFR galaxies compared to our observational sample, however, the sSFR distributions agree after limiting both samples to a minimum sSFR. A deeper and more complete photometric sample further indicates that observed low-mass galaxies could span the full range of sSFR predicted in the FIRE-2 simulations. Our results support the existence of short-timescale dynamical effects driven by gas outflow and inflow cycles in low-mass galaxies and motivate additional tests of the breathing mode.

astro-ph.GA↗

arXiv2025

Diversity and universality: evolution of dwarf galaxies with self-interacting dark matter

Dark matter halos with self-interacting dark matter (SIDM) experience a unique evolutionary phenomenon, in that their central regions eventually collapse to high density through the runaway gravothermal process after initially forming a large and low-density core. When coupled with orbital evolution, this is expected to naturally produce a large diversity in dark-matter halos' inner mass distribution, potentially explaining the diversity problem of dwarf galaxies. However, it remains unknown how the diversity in SIDM dark-matter halos propagates to the more easily observed luminous matter at the center of the halo, especially the stellar component. In this work, we use idealized N-body simulations with two species of particles (dark matter and stars) to study the response of the stellar properties of field and satellite dwarf galaxies to SIDM evolution and orbital effects on their halos. Galaxies' stellar components, including galaxy size, mass-to-light ratio, and stellar velocity dispersion, display a much larger scatter in SIDM than the standard cold dark matter (CDM) model. Importantly, we find signs of universality in the evolution pathways, or "tidal tracks", of SIDM dwarf satellites, which are physically interpretable and potentially parameterizable. This type of tidal-track model can be layered onto larger-scale, cosmological simulations to reconstruct the evolution of populations of SIDM dwarfs in cases where high-resolution simulations of galaxies are otherwise prohibitively expensive.

astro-ph.GA↗

arXiv2025

A Universal Banach--Bregman Framework for Stochastic Iterations: Unifying Stochastic Mirror Descent, Learning and LLM Training

Stochastic optimization powers the scalability of modern artificial intelligence, spanning machine learning, deep learning, reinforcement learning, and large language model training. Yet, existing theory remains largely confined to Hilbert spaces, relying on inner-product frameworks and orthogonality. This paradigm fails to capture non-Euclidean settings, such as mirror descent on simplices, Bregman proximal methods for sparse learning, natural gradient descent in information geometry, or Kullback--Leibler-regularized language model training. Unlike Euclidean-based Hilbert-space methods, this approach embraces general Banach spaces. This work introduces a pioneering Banach--Bregman framework for stochastic iterations, establishing Bregman geometry as a foundation for next-generation optimization. It (i) provides a unified template via Bregman projections and Bregman--Fejer monotonicity, encompassing stochastic approximation, mirror descent, natural gradient, adaptive methods, and mirror-prox; (ii) establishes super-relaxations ($λ> 2$) in non-Hilbert settings, enabling flexible geometries and elucidating their acceleration effect; and (iii) delivers convergence theorems spanning almost-sure boundedness to geometric rates, validated on synthetic and real-world tasks. Empirical studies across machine learning (UCI benchmarks), deep learning (e.g., Transformer training), reinforcement learning (actor--critic), and large language models (WikiText-2 with distilGPT-2) show up to 20% faster convergence, reduced variance, and enhanced accuracy over classical baselines. These results position Banach--Bregman geometry as a cornerstone unifying optimization theory and practice across core AI paradigms.

cs.LG↗

arXiv2025

Six-DoF Hand-Based Teleoperation for Omnidirectional Aerial Robots

Omnidirectional aerial robots offer full 6-DoF independent control over position and orientation, making them popular for aerial manipulation. Although advancements in robotic autonomy, human operation remains essential in complex aerial environments. Existing teleoperation approaches for multirotors fail to fully leverage the additional DoFs provided by omnidirectional rotation. Additionally, the dexterity of human fingers should be exploited for more engaged interaction. In this work, we propose an aerial teleoperation system that brings the rotational flexibility of human hands into the unbounded aerial workspace. Our system includes two motion-tracking marker sets--one on the shoulder and one on the hand--along with a data glove to capture hand gestures. Using these inputs, we design four interaction modes for different tasks, including Spherical Mode and Cartesian Mode for long-range moving, Operation Mode for precise manipulation, as well as Locking Mode for temporary pauses, where the hand gestures are utilized for seamless mode switching. We evaluate our system on a vertically mounted valve-turning task in the real world, demonstrating how each mode contributes to effective aerial manipulation. This interaction framework bridges human dexterity with aerial robotics, paving the way for enhanced aerial teleoperation in unstructured environments.

cs.RO↗

arXiv2025

CFFormer: Cross CNN-Transformer Channel Attention and Spatial Feature Fusion for Improved Segmentation of Heterogeneous Medical Images

Medical image segmentation plays an important role in computer-aided diagnosis. Existing methods mainly utilize spatial attention to highlight the region of interest. However, due to limitations of medical imaging devices, medical images exhibit significant heterogeneity, posing challenges for segmentation. Ultrasound images, for instance, often suffer from speckle noise, low resolution, and poor contrast between target tissues and background, which may lead to inaccurate boundary delineation. To address these challenges caused by heterogeneous image quality, we propose a hybrid CNN-Transformer model,called CFFormer, which leverages effective channel feature extraction to enhance the model' s ability to accurately identify tissue regions by capturing rich contextual information. The proposed architecture contains two key components: the Cross Feature Channel Attention (CFCA) module and the X-Spatial Feature Fusion (XFF) module. The model incorporates dual encoders, with the CNN encoder focusing on capturing local features and the Transformer encoder modeling global features. The CFCA module filters and facilitates interactions between the channel features from the two encoders, while the XFF module effectively reduces the significant semantic information differences in spatial features, enabling a smooth and cohesive spatial feature fusion. We evaluate our model across eight datasets covering five modalities to test its generalization capability. Experimental results demonstrate that our model outperforms current state-of-the-art methods and maintains accurate tissue region segmentation across heterogeneous medical image datasets. The code is available at https://github.com/JiaxuanFelix/CFFormer.

cs.CV↗

arXiv2025

EP250108a/SN 2025kg: A Jet-Driven Stellar Explosion Interacting With Circumstellar Material

We present optical, radio, and X-ray observations of EP250108a/SN 2025kg, a broad-line Type Ic supernova (SN Ic-BL) accompanying an Einstein Probe (EP) fast X-ray transient (FXT) at $z=0.176$. EP250108a/SN 2025kg possesses a double-peaked optical light curve and its spectrum transitions from a blue underlying continuum to a typical SN Ic-BL spectrum over time. We fit a radioactive decay model to the second peak of the optical light curve and find SN parameters that are consistent with the SNe Ic-BL population, while its X-ray and radio properties are consistent with those of low-luminosity GRB (LLGRB) 060218/SN 2006aj. We explore three scenarios to understand the system's multi-wavelength emission -- (a) SN ejecta interacting with an extended circumstellar medium (CSM), (b) the shocked cocoon of a collapsar-driven jet choked in its stellar envelope, and (c) the shocked cocoon of a collapsar-driven jet choked in an extended CSM. Models (b) and (c) can explain the optical light curve and are also consistent with the radio and X-ray observations. We favor model (c) because it can self-consistently explain both the X-ray prompt emission and first optical peak, but we do not rule out model (b). From the properties of the first peak in model (c), we find evidence that EP250108a/SN 2025kg interacts with an extended CSM, and infer an envelope mass $M_{\rm e} \sim 0.1\,\rm M_\odot$ and radius $R_{\rm e} \sim 4 \times 10^{13}$ cm. EP250108a/SN 2025kg's multi-wavelength properties make it a close analog to LLGRB 060218/SN 2006aj, and highlight the power of early follow-up observations in mapping the environments of massive stars prior to core collapse.

astro-ph.HE↗

arXiv2025

De-LightSAM: Modality-Decoupled Lightweight SAM for Generalizable Medical Segmentation

The universality of deep neural networks across different modalities and their generalization capabilities to unseen domains play an essential role in medical image segmentation. The recent segment anything model (SAM) has demonstrated strong adaptability across diverse natural scenarios. However, the huge computational costs, demand for manual annotations as prompts and conflict-prone decoding process of SAM degrade its generalization capabilities in medical scenarios. To address these limitations, we propose a modality-decoupled lightweight SAM for domain-generalized medical image segmentation, named De-LightSAM. Specifically, we first devise a lightweight domain-controllable image encoder (DC-Encoder) that produces discriminative visual features for diverse modalities. Further, we introduce the self-patch prompt generator (SP-Generator) to automatically generate high-quality dense prompt embeddings for guiding segmentation decoding. Finally, we design the query-decoupled modality decoder (QM-Decoder) that leverages a one-to-one strategy to provide an independent decoding channel for every modality, preventing mutual knowledge interference of different modalities. Moreover, we design a multi-modal decoupled knowledge distillation (MDKD) strategy to leverage robust common knowledge to complement domain-specific medical feature representations. Extensive experiments indicate that De-LightSAM outperforms state-of-the-arts in diverse medical imaging segmentation tasks, displaying superior modality universality and generalization capabilities. Especially, De-LightSAM uses only 2.0% parameters compared to SAM-H. The source code is available at https://github.com/xq141839/De-LightSAM.

eess.IV↗

arXiv2025

Exposure measurement error correction in longitudinal studies with discrete outcomes

Environmental epidemiologists are often interested in estimating the effect of time-varying functions of the exposure history on health outcomes. However, the individual exposure measurements that constitute the history upon which an exposure history function is constructed are usually subject to measurement errors. To obtain unbiased estimates of the effects of such mismeasured functions in longitudinal studies with discrete outcomes, a method applicable to the main study/validation study design is developed. Various estimation procedures are explored. Simulation studies were conducted to assess its performance compared to standard analysis, and we found that the proposed method had good performance in terms of finite sample bias reduction and nominal coverage probability improvement. As an illustrative example, we applied the new method to a study of long-term exposure to PM2.5, in relation to the occurrence of anxiety disorders in the Nurses Health Study II. Failing to correct the error-prone exposure can lead to an underestimation of the chronic exposure effect of PM2.5.

stat.ME↗

arXiv2025

CSPENet: Contour-Aware and Saliency Priors Embedding Network for Infrared Small Target Detection

Infrared small target detection (ISTD) plays a critical role in a wide range of civilian and military applications. Existing methods suffer from deficiencies in the localization of dim targets and the perception of contour information under dense clutter environments, severely limiting their detection performance. To tackle these issues, we propose a contour-aware and saliency priors embedding network (CSPENet) for ISTD. We first design a surround-convergent prior extraction module (SCPEM) that effectively captures the intrinsic characteristic of target contour pixel gradients converging toward their center. This module concurrently extracts two collaborative priors: a boosted saliency prior for accurate target localization and multi-scale structural priors for comprehensively enriching contour detail representation. Building upon this, we propose a dual-branch priors embedding architecture (DBPEA) that establishes differentiated feature fusion pathways, embedding these two priors at optimal network positions to achieve performance enhancement. Finally, we develop an attention-guided feature enhancement module (AGFEM) to refine feature representations and improve saliency estimation accuracy. Experimental results on public datasets NUDT-SIRST, IRSTD-1k, and NUAA-SIRST demonstrate that our CSPENet outperforms other state-of-the-art methods in detection performance. The code is available at https://github.com/IDIP2025/CSPENet.

cs.CV↗

arXiv2025

ELVES-Dwarf I: Satellites Systems of Eight Isolated Dwarf Galaxies in the Local Volume

The satellite populations of Milky Way--mass systems have been extensively studied, significantly advancing our understanding of galaxy formation and dark matter physics. In contrast, the satellites of lower-mass dwarf galaxies remain largely unexplored, despite hierarchical structure formation predicting that dwarf galaxies should host their own satellites. We present the first results of the ELVES-Dwarf survey, which aims to statistically characterize the satellite populations of isolated dwarf galaxies in the Local Volume ($4<D<10$~Mpc). We identify satellite candidates in integrated light using the Legacy Surveys data and are complete down to $M_g\approx -9$ mag. We then confirm the association of satellite candidates with host galaxies using surface brightness fluctuation distances measured from the Hyper Suprime-Cam data. We surveyed 8 isolated dwarf galaxies with stellar masses ranging from sub-Small Magellanic Cloud to Large Magellanic Cloud scales ($10^{7.8} < M_\star^{\rm host}<10^{9.5}\, M_\odot$) and confirmed 6 satellites with stellar masses between $10^{5.6}$ and $10^{8} \, M_\odot$. Most confirmed satellites are star-forming, contrasting with the primarily quiescent satellites observed around Milky Way--mass hosts. By comparing observed satellite abundances and stellar mass functions with theoretical predictions, we find no evidence of a "missing satellite problem" in the dwarf galaxy regime.

astro-ph.GA↗

arXiv2025

Counting Little Red Dots at $z<4$ with Ground-based Surveys and Spectroscopic Follow-up

Little red dots (LRDs) are a population of red, compact objects discovered by JWST at $z>4$. At $4 4$ to $\sim10^{-5.3}\,\mathrm{cMpc^{-3}}$ at $2.7<z<3.7$ and $\sim10^{-5.7}\,\mathrm{cMpc^{-3}}$ at $1.7<z<2.7$. We also present the Magellan/FIRE spectrum of our first followed-up candidate, DEEP23-z2LRD1 at $z_\mathrm{spec}=2.26$, as a proof of concept for our sample selection. Similar to high-redshift LRDs, the spectrum of DEEP23-z2LRD1 exhibits broad H$α$ emission with $\mathrm{FWHM}\approx2400\,\mathrm{km\,s^{-1}}$ and with nearly symmetric narrow H$α$ absorption. Additionally, DEEP23-z2LRD1 has extremely narrow [OIII] lines with $\mathrm{FWHM}\approx140\,\mathrm{km\,s^{-1}}$, suggesting the presence of an accreting black hole in a low-mass host galaxy. Limited by the angular resolution of ground-based surveys, we emphasize that spectroscopic follow-ups are required to characterize the contamination fraction of this sample and pin down LRD number density at $z<4$.

astro-ph.GA↗

arXiv2025

SPACER: A Parallel Dataset of Speech Production And Comprehension of Error Repairs

Speech errors are a natural part of communication, yet they rarely lead to complete communicative failure because both speakers and comprehenders can detect and correct errors. Although prior research has examined error monitoring and correction in production and comprehension separately, integrated investigation of both systems has been impeded by the scarcity of parallel data. In this study, we present SPACER, a parallel dataset that captures how naturalistic speech errors are corrected by both speakers and comprehenders. We focus on single-word substitution errors extracted from the Switchboard corpus, accompanied by speaker's self-repairs and comprehenders' responses from an offline text-editing experiment. Our exploratory analysis suggests asymmetries in error correction strategies: speakers are more likely to repair errors that introduce greater semantic and phonemic deviations, whereas comprehenders tend to correct errors that are phonemically similar to more plausible alternatives or do not fit into prior contexts. Our dataset enables future research on integrated approaches toward studying language production and comprehension.

cs.CL↗

arXiv2025

Hedgehog: An Isolated Quiescent Dwarf Galaxy at 2.4 Mpc

It is well-known that almost all isolated dwarf galaxies are actively forming stars. We report the discovery of dw1322m2053 (nicknamed Hedgehog), an isolated quiescent dwarf galaxy at a distance of $2.40\pm0.15$ Mpc with a stellar mass of $M_\star \approx 10^{5.8}\, M_\odot$. The distance is measured using surface brightness fluctuations with both Legacy Surveys and deep Magellan/IMACS imaging data. Hedgehog is 1.7~Mpc from the nearest galaxy group, Centaurus A, and has no neighboring galaxies within 1~Mpc, making it one of the most isolated quiescent dwarf galaxies at this stellar mass. It has a red optical color and early-type morphology and shows no UV emission. This indicates that Hedgehog has an old stellar population and no ongoing star formation. Compared with other quiescent dwarfs in the Local Group and Local Volume, Hedgehog appears smaller in size for its luminosity but is consistent with the mass--size relations. Hedgehog might be a backsplash galaxy from the Centaurus A group, but it could also have been quenched in the field by ram pressure stripping in the cosmic web, reionization, or internal processes such as supernova and stellar feedback. Future observations are needed to fully unveil its formation, history, and quenching mechanisms.

astro-ph.GA↗

arXiv2024

NEMO: Can Multimodal LLMs Identify Attribute-Modified Objects?

Multimodal Large Language Models (MLLMs) have made notable advances in visual understanding, yet their abilities to recognize objects modified by specific attributes remain an open question. To address this, we explore MLLMs' reasoning capabilities in object recognition, ranging from commonsense to beyond-commonsense scenarios. We introduce a novel benchmark, NEMO, which comprises 900 images of origiNal fruits and their corresponding attributE-MOdified ones; along with a set of 2,700 questions including open-, multiple-choice-, unsolvable types. We assess 26 recent open-sourced and commercial models using our benchmark. The findings highlight pronounced performance gaps in recognizing objects in NEMO and reveal distinct answer preferences across different models. Although stronger vision encoders improve performance, MLLMs still lag behind standalone vision encoders. Interestingly, scaling up the model size does not consistently yield better outcomes, as deeper analysis reveals that larger LLMs can weaken vision encoders during fine-tuning. These insights shed light on critical limitations in current MLLMs and suggest potential pathways toward developing more versatile and resilient multimodal models.

cs.CV↗

arXiv2024

Decomposition of surprisal: Unified computational model of ERP components in language processing

The functional interpretation of language-related ERP components has been a central debate in psycholinguistics for decades. We advance an information-theoretic model of human language processing in the brain in which incoming linguistic input is processed at first shallowly and later with more depth, with these two kinds of information processing corresponding to distinct electroencephalographic signatures. Formally, we show that the information content (surprisal) of a word in context can be decomposed into two quantities: (A) shallow surprisal, which signals shallow processing difficulty for a word, and corresponds with the N400 signal; and (B) deep surprisal, which reflects the discrepancy between shallow and deep representations, and corresponds to the P600 signal and other late positivities. Both of these quantities can be estimated straightforwardly using modern NLP models. We validate our theory by successfully simulating ERP patterns elicited by a variety of linguistic manipulations in previously-reported experimental data from six experiments, with successful novel qualitative and quantitative predictions. Our theory is compatible with traditional cognitive theories assuming a `good-enough' shallow representation stage, but with a precise information-theoretic formulation. The model provides an information-theoretic model of ERP components grounded on cognitive processes, and brings us closer to a fully-specified neuro-computational model of language processing.

cs.CL↗

arXiv2024

Measurement Error Correction for Spatially Defined Environmental Exposures in Survival Analysis

Environmental exposures are often defined using buffer zones around geocoded home addresses, but these static boundaries can miss dynamic daily activity patterns, leading to biased results. This paper presents a novel measurement error correction method for spatially defined environmental exposures within a survival analysis framework using the Cox proportional hazards model. The method corrects high-dimensional surrogate exposures from geocoded residential data at multiple buffer radii by applying principal component analysis for dimension reduction and leveraging external GPS-tracked validation datasets containing true exposure measurements. It also derives the asymptotic properties and variances of the proposed estimators. Extensive simulations are conducted to evaluate the performance of the proposed estimators, demonstrating its ability to improve accuracy in estimated exposure effects. An illustrative application assesses the impact of greenness exposure on depression incidence in the Nurses' Health Study (NHS). The results demonstrate that correcting for measurement error significantly enhances the accuracy of exposure estimates. This method offers a critical advancement for accurately assessing the health impacts of environmental exposures, outperforming traditional static buffer approaches.

stat.ME↗