arXiv subjects

He Zhao

Publications and source records attributed to He Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2026

Multi-Scale Wavelet Transformers for Operator Learning of Dynamical Systems

Recent years have seen a surge in data-driven surrogates for dynamical systems that can be orders of magnitude faster than numerical solvers. However, many machine learning-based models such as neural operators exhibit spectral bias, attenuating high-frequency components that often encode small-scale structure. This limitation is particularly damaging in applications such as weather forecasting, where misrepresented high frequencies can induce long-horizon instability. To address this issue, we propose multi-scale wavelet transformers (MSWTs), which learn system dynamics in a tokenized wavelet domain. The wavelet transform explicitly separates low- and high-frequency content across scales. MSWTs leverage a wavelet-preserving downsampling scheme that retains high-frequency features and employ wavelet-based attention to capture dependencies across scales and frequency bands. Experiments on chaotic dynamical systems show substantial error reductions and improved long horizon spectral fidelity. On the ERA5 climate reanalysis, MSWTs further reduce climatological bias, demonstrating their effectiveness in a real-world forecasting setting.

cs.LG↗

arXiv2026

Calibrating Tabular Anomaly Detection via Optimal Transport

Tabular anomaly detection (TAD) remains challenging due to the heterogeneity of tabular data: features lack natural relationships, vary widely in distribution and scale, and exhibit diverse types. Consequently, each TAD method makes implicit assumptions about anomaly patterns that work well on some datasets but fail on others, and no method consistently outperforms across diverse scenarios. We present CTAD (Calibrating Tabular Anomaly Detection), a model-agnostic post-processing framework that enhances any existing TAD detector through sample-specific calibration. Our approach characterizes normal data via two complementary distributions, i.e., an empirical distribution from random sampling and a structural distribution from K-means centroids, and measures how adding a test sample disrupts their compatibility using Optimal Transport (OT) distance. Normal samples maintain low disruption while anomalies cause high disruption, providing a calibration signal to amplify detection. We prove that OT distance has a lower bound proportional to the test sample's distance from centroids, and establish that anomalies systematically receive higher calibration scores than normals in expectation, explaining why the method generalizes across datasets. Extensive experiments on 34 diverse tabular datasets with 7 representative detectors spanning all major TAD categories (density estimation, classification, reconstruction, and isolation-based methods) demonstrate that CTAD consistently improves performance with statistical significance. Remarkably, CTAD enhances even state-of-the-art deep learning methods and shows robust performance across diverse hyperparameter settings, requiring no additional tuning for practical deployment.

cs.LG↗

arXiv2026

Interpreting V1 Population Activity via Image-Neural Latent Representation Alignment

Understanding the neural mechanisms underlying visual computation has long been a central challenge in neuroscience. Recent alignment based approaches have improved the accuracy of decoding visual stimuli from brain activity, yet they provide limited insight into the neural computations that give rise to these improvements. To address this gap, we propose Dual-Tower Image-Neural Alignment (DINA), an interpretable contrastive framework for analyzing population level visual computations in primary visual cortex (V1). DINA jointly trains a biologically motivated dual-tower architecture that aligns visual stimuli and corresponding V1 population responses in a shared latent space at the level of intermediate feature maps, enabling both accurate decoding and direct access to interpretable feature maps. Evaluated on large-scale two-photon calcium imaging data from mouse V1, DINA achieves accurate neural-based decoding while revealing that decoding performance is primarily supported by coarse, low-level visual structure, rather than semantic category information or fine-grained details. Further analysis reveals that alignable feature maps emerge from multiple spatially distributed image regions, capturing both shape and texture cues, and are predominantly reconstructed by sparse subsets of strongly responsive neurons and their functional interactions. Together, these results confirm that, beyond enabling accurate decoding, DINA provides a principled framework for probing the computational mechanisms underlying visual processing in V1.

cs.NE↗

arXiv2026

Structure-Augmented Standard Plane Detection with Temporal Aggregation in Blind-Sweep Fetal Ultrasound

In low-resource settings, blind-sweep ultrasound provides a practical and accessible method for identifying fetal growth restriction. However, unlike freehand ultrasound which is subjectively controlled, detection of biometry plane in blind-sweep ultrasound is more challenging due to the uncontrolled fetal structure to be observed and the variaties of oblique planes in the scan. In this work, we propose a structure-augmented system to detect fetal abdomen plane, where the abdominal structure is highlighted using a segmentation prior. Since standard planes are emerging gradually, the decision boundary of the keyframes is unstable to predict. We thus aggregated the structure-augmented planes with a temporal sliding window to help stabilise keyframe localisation. Extensive results indicate that the structure-augmented temporal sliding strategy significantly improves and stabilises the detection of anatomically meaningful planes, which enables more reliable biometric measurements in blind-sweep ultrasound.

cs.CV↗

arXiv2026

Near OOD Detection for Vision-Language Prompt Learning with Contrastive Logit Score

Prompt learning has emerged as an efficient and effective method for fine-tuning vision-language models such as CLIP. While many studies have explored generalisation abilities of these models in few-shot classification tasks and a few studies have addressed far out-of-distribution (OOD) of the models, their potential for addressing near OOD detection remains underexplored. Existing methods either require training from scratch, need fine-tuning, or are not designed for vision-language prompt learning. To address this, we introduce the Contrastive Logit Score (CLS), a novel post-hoc, plug-and-play scoring function. CLS significantly improves near OOD detection of pre-trained vision-language prompt learning methods without modifying their model architectures or requiring retraining. Our method achieves up to an 11.67% improvement in AUROC for near OOD detection with minimal computational overhead. Extensive evaluations validate the effectiveness, efficiency, and generalisability of our approach. Our code is available at https://github.com/davidmcjung/near-OOD-prompt-learning.

cs.CV↗

arXiv2026

HiRO-Nav: Hybrid ReasOning Enables Efficient Embodied Navigation

Embodied navigation agents built upon large reasoning models (LRMs) can handle complex, multimodal environmental input and perform grounded reasoning per step to improve sequential decision-making for long-horizon tasks. However, a critical question remains: \textit{how can the reasoning capabilities of LRMs be harnessed intelligently and efficiently for long-horizon navigation tasks?} In simple scenes, agents are expected to act reflexively, while in complex ones they should engage in deliberate reasoning before acting.To achieve this, we introduce \textbf{H}ybr\textbf{i}d \textbf{R}eas\textbf{O}ning \textbf{Nav}igation (\textbf{HiRO-Nav}) agent, the first kind of agent capable of adaptively determining whether to perform thinking at every step based on its own action entropy. Specifically, by examining how the agent's action entropy evolves over the navigation trajectories, we observed that only a small fraction of actions exhibit high entropy, and these actions often steer the agent toward novel scenes or critical objects. Furthermore, studying the relationship between action entropy and task completion (i.e., Q-value) reveals that improving high-entropy actions contributes more positively to task success.Hence, we propose a tailored training pipeline comprising hybrid supervised fine-tuning as a cold start, followed by online reinforcement learning with the proposed hybrid reasoning strategy to explicitly activate reasoning only for high-entropy actions, significantly reducing computational overhead while improving decision quality. Extensive experiments on the \textsc{CHORES}-$\mathbb{S}$ ObjectNav benchmark showcases that HiRO-Nav achieves a better trade-off between success rates and token efficiency than both dense-thinking and no-thinking baselines.

cs.AI↗

arXiv2026

Graph Topology Information Enhanced Heterogeneous Graph Representation Learning

Real-world heterogeneous graphs are inherently noisy and usually not in the optimal graph structures for downstream tasks, which often adversely affects the performance of GRL models in downstream tasks. Although Graph Structure Learning (GSL) methods have been proposed to learn graph structures and downstream tasks simultaneously, existing methods are predominantly designed for homogeneous graphs, while GSL for heterogeneous graphs remains largely unexplored. Two challenges arise in this context. Firstly, the quality of the input graph structure has a more profound impact on GNN-based heterogeneous GRL models compared to their homogeneous counterparts. Secondly, most existing homogenous GRL models encounter memory consumption issues when applied directly to heterogeneous graphs. In this paper, we propose a novel Graph Topology learning Enhanced Heterogeneous Graph Representation Learning framework (ToGRL).ToGRL learns high-quality graph structures and representations for downstream tasks by incorporating task-relevant latent topology information. Specifically, a novel GSL module is first proposed to extract downstream task-related topology information from a raw graph structure and project it into topology embeddings. These embeddings are utilized to construct a new graph with smooth graph signals. This two-stage approach to GSL separates the optimization of the adjacency matrix from node representation learning to reduce memory consumption. Following this, a representation learning module takes the new graph as input to learn embeddings for downstream tasks. ToGRL also leverages prompt tuning to better utilize the knowledge embedded in learned representations, thus enhancing adaptability to downstream tasks. Extensive experiments on five real-world datasets show that our ToGRL outperforms state-of-the-art methods by a large margin.

cs.LG↗

arXiv2026

LLM as an Algorithmist: Enhancing Anomaly Detectors via Programmatic Synthesis

Existing anomaly detection (AD) methods for tabular data usually rely on some assumptions about anomaly patterns, leading to inconsistent performance in real-world scenarios. While Large Language Models (LLMs) show remarkable reasoning capabilities, their direct application to tabular AD is impeded by fundamental challenges, including difficulties in processing heterogeneous data and significant privacy risks. To address these limitations, we propose LLM-DAS, a novel framework that repositions the LLM from a ``data processor'' to an ``algorithmist''. Instead of being exposed to raw data, our framework leverages the LLM's ability to reason about algorithms. It analyzes a high-level description of a given detector to understand its intrinsic weaknesses and then generates detector-specific, data-agnostic Python code to synthesize ``hard-to-detect'' anomalies that exploit these vulnerabilities. This generated synthesis program, which is reusable across diverse datasets, is then instantiated to augment training data, systematically enhancing the detector's robustness by transforming the problem into a more discriminative two-class classification task. Extensive experiments on 36 TAD benchmarks show that LLM-DAS consistently boosts the performance of mainstream detectors. By bridging LLM reasoning with classic AD algorithms via programmatic synthesis, LLM-DAS offers a scalable, effective, and privacy-preserving approach to patching the logical blind spots of existing detectors.

cs.LG↗

arXiv2026

MiroEval: Benchmarking Multimodal Deep Research Agents in Process and Outcome

Recent progress in deep research systems has been impressive, but evaluation still lags behind real user needs. Existing benchmarks predominantly assess final reports using fixed rubrics, failing to evaluate the underlying research process. Most also offer limited multimodal coverage, rely on synthetic tasks that do not reflect real-world query complexity, and cannot be refreshed as knowledge evolves. To address these gaps, we introduce MiroEval, a benchmark and evaluation framework for deep research systems. The benchmark comprises 100 tasks (70 text-only, 30 multimodal), all grounded in real user needs and constructed via a dual-path pipeline that supports periodic updates, enabling a live and evolving setting. The proposed evaluation suite assesses deep research systems along three complementary dimensions: adaptive synthesis quality evaluation with task-specific rubrics, agentic factuality verification via active retrieval and reasoning over both web sources and multimodal attachments, and process-centric evaluation audits how the system searches, reasons, and refines throughout its investigation. Evaluation across 13 systems yields three principal findings: the three evaluation dimensions capture complementary aspects of system capability, with each revealing distinct strengths and weaknesses across systems; process quality serves as a reliable predictor of overall outcome while revealing weaknesses invisible to output-level metrics; and multimodal tasks pose substantially greater challenges, with most systems declining by 3 to 10 points. The MiroThinker series achieves the most balanced performance, with MiroThinker-H1 ranking the highest overall in both settings. Human verification and robustness results confirm the reliability of the benchmark and evaluation framework. MiroEval provides a holistic diagnostic tool for the next generation of deep research agents.

cs.AI↗

arXiv2026

Radiative Compression of Dense Cores in the Pillars of Creation as Revealed by JWST Extinction Mapping

The Pillars of Creation in M16 represent an iconic star-forming region where stellar feedback shapes molecular cloud evolution. We present a detailed investigation of dust extinction and density structure in the Pillars of Creation using multiband photometric observations from \emph{JWST} NIRCam. A high-resolution (2\arcsec) extinction map reaching depths of $A_V\sim 100$ mag has been constructed using NIRCam filters F090W, F200W, F335M, and F444W. This map clearly reveals the intricate structure of dense gas within the molecular cloud in the Pillars of Creation region. Analysis of the column density probability distribution function (N-PDF) exhibits a characteristic lognormal distribution at intermediate extinctions ($A_V\approx10-30$\,mag), which transitions to a power-law tail at high extinctions ($A_V\gtrsim$ 30\,mag) where star-forming cores reside. The power-law slope $α$ displays significant spatial variation, steepening from $α\approx 2.0$ at the pillar tips facing the NGC 6611 cluster to $α\approx$4.0 in regions distant from the cluster. This systematic gradient demonstrates that stellar feedback not only disperses molecular clouds but can also locally enhance the formation of dense, self-gravitating structures through radiative compression.

astro-ph.GA↗

arXiv2026

The Extinction Distance of Supernova Remnants in Combination with the CO Line Measurements

Accurate distance measurements to supernova remnants (SNRs) are crucial for understanding their physical properties and evolution. We present a novel method that combines CO line observations with three-dimensional (3D) extinction maps to determine distances to SNRs (G93.7$-$0.2, G109.1$-$1.0, G156.2+5.7, and G166.0+4.3) through their associated molecular clouds. For each SNR, candidate CO velocity components corresponding to interacting molecular clouds are identified based on previous observational evidence with refinements: [$-$19, $-$3] km s$^{-1}$ for G93.7$-$0.2, [$-$51, $-$46] km s$^{-1}$ for G109.1$-$1.0, [$-$10, 0] km s$^{-1}$ for G156.2+5.7, and [$-$27, $-$15] km s$^{-1}$ for G166.0+4.3. By examining extinction-distance profiles along the sightlines and identifying extinction jumps that spatially coincide with CO emission features, we derive distances of 1.82$\pm$0.13 kpc for G93.7$-$0.2, 3.05$\pm$0.15 kpc for G109.1$-$1.0, 0.60$\pm$0.15 kpc for G156.2+5.7, and 3.44$\pm$0.23 kpc for G166.0+4.3. Our extinction-based distances are largely consistent with previous estimates while with better accuracy and robustness.

astro-ph.GA↗

arXiv2026

Deep Tabular Representation Corrector

Tabular data have been playing a mostly important role in diverse real-world fields, such as healthcare, engineering, finance, etc. The recent success of deep learning has fostered many deep networks (e.g., Transformer, ResNet) based tabular learning methods. Generally, existing deep tabular machine learning methods are along with the two paradigms, i.e., in-learning and pre-learning. In-learning methods need to train networks from scratch or impose extra constraints to regulate the representations which nonetheless train multiple tasks simultaneously and make learning more difficult, while pre-learning methods design several pretext tasks for pre-training and then conduct task-specific fine-tuning, which however need much extra training effort with prior knowledge. In this paper, we introduce a novel deep Tabular Representation Corrector, TRC, to enhance any trained deep tabular model's representations without altering its parameters in a model-agnostic manner. Specifically, targeting the representation shift and representation redundancy that hinder prediction, we propose two tasks, i.e., (i) Tabular Representation Re-estimation, that involves training a shift estimator to calculate the inherent shift of tabular representations to subsequently mitigate it, thereby re-estimating the representations and (ii) Tabular Space Mapping, that transforms the above re-estimated representations into a light-embedding vector space via a coordinate estimator while preserves crucial predictive information to minimize redundancy. The two tasks jointly enhance the representations of deep tabular models without touching on the original models thus enjoying high efficiency. Finally, we conduct extensive experiments on state-of-the-art deep tabular machine learning models coupled with TRC on various tabular benchmarks which have shown consistent superiority.

cs.LG↗

arXiv2026

A cationic carrier for diffuse interstellar band at 862.1 nm: Evidence from the skin effect in nearby diffuse-to-translucent clouds

The tendency of some diffuse interstellar band (DIB) carriers to concentrate in the outer, UV-illuminated layers of molecular clouds (MCs)--the ``skin effect''--makes their spatial distribution a powerful probe of their physical nature. We leverage Gaia DR3 measurements of the DIB at 862.1 nm to investigate its behavior across 12 nearby MCs, spanning diffuse to translucent regimes ($A_{\rm V}\,{\sim}\,0.2{-}3.5$ mag). We find significant diversity in the DIB behavior, both between different clouds and within individual clouds from their outer to inner regions. To quantify these trends, we employed a piecewise linear model (PLM) to fit the average slope ($α$) between the normalized DIB strength, ${\rm log_{10}}(W_{8621}/A_{\rm V})$, and dust extinction, ${\rm log_{10}}(A_{\rm V})$. In general, ${\rm log_{10}}(W_{8621}/A_{\rm V})$ declines with ${\rm log_{10}}(A_{\rm V})$ with $α$ between 0 and --1, becoming progressively steeper at higher $A_{\rm V}$. These observed slopes and their variations are consistent with the photoionization equilibrium models, where the carrier abundance is governed by local conditions, particularly the UV radiation field and cloud structure (e.g., density profiles, clumpiness). Particularly, the Taurus cloud region uniquely displays an initial increase in ${\rm log_{10}}(W_{8621}/A_{\rm V})$ at low extinction, a signature predicted for a cationic carrier. By fitting the slope of this rising trend, we estimate an ionization potential of $E_{\rm IP}\,{=}\,12.40^{+1.90}_{-2.29}$ eV for the DIB$λ$8621 carrier, which aligns well with the secondary ionization energies of large carbonaceous molecules like polycyclic aromatic hydrocarbons (PAHs) or fullerenes.

astro-ph.GA↗

arXiv2026

Permutation-based Inference for Variational Learning of Directed Acyclic Graphs

Estimating the structure of Bayesian networks as directed acyclic graphs (DAGs) from observational data is a fundamental challenge, particularly in causal discovery. Bayesian approaches excel by quantifying uncertainty and addressing identifiability, but key obstacles remain: (i) representing distributions over DAGs and (ii) estimating a posterior in the underlying combinatorial space. We introduce PIVID, a method that jointly infers a distribution over permutations and DAGs using variational inference and continuous relaxations of discrete distributions. Through experiments on synthetic and real-world datasets, we show that PIVID can outperform deterministic and Bayesian approaches, achieving superior accuracy-uncertainty trade-offs while scaling efficiently with the number of variables.

cs.LG↗

arXiv2026

Unifying Stable Optimization and Reference Regularization in RLHF

Reinforcement Learning from Human Feedback (RLHF) has advanced alignment capabilities significantly but remains hindered by two core challenges: \textbf{reward hacking} and \textbf{stable optimization}. Current solutions independently address these issues through separate regularization strategies, specifically a KL-divergence penalty against a supervised fine-tuned model ($π_0$) to mitigate reward hacking, and policy ratio clipping towards the current policy ($π_t$) to promote stable alignment. However, the implicit trade-off arising from simultaneously regularizing towards both $π_0$ and $π_t$ remains under-explored. In this paper, we introduce a unified regularization approach that explicitly balances the objectives of preventing reward hacking and maintaining stable policy updates. Our simple yet principled alignment objective yields a weighted supervised fine-tuning loss with a superior trade-off, which demonstrably improves both alignment results and implementation complexity. Extensive experiments across diverse benchmarks validate that our method consistently outperforms RLHF and online preference learning methods, achieving enhanced alignment performance and stability.

cs.LG↗

arXiv2026

Design and fabrication of guiding patterns for topography-based searching of 2D devices for scanning tunneling microscopy measurements

We report the design and fabrication of guiding patterns for topography-based searching of two-dimensional (2D) devices for scanning tunneling microscopy (STM) measurements. Sub-micron geometric coordinate markers were etched into SiO2/Si wafers, serving as both substrates for 2D device integration and guiding maps for sample navigation. Here, we used a monolayer graphene/h-BN device with an active area of smaller than 20 um x 20 um as a model system and demonstrated that the device could be reliably located in STM solely through topographic imaging of the guiding patterns and in situ stage calibration, without reliance on optical viewports or capacitive sensing. Atomically resolved topographic imaging and tunneling spectroscopy were also obtained. Our proposed navigation strategy is fully compatible with standard device fabrication procedures and requires no hardware modification to existing STM setups. As such, it offers a practical alternative for locating miniaturized devices in STM, shedding light on studying emerging quantum phenomena in 2D systems.

cond-mat.str-el↗

arXiv2026

Leveraging Persistence Image to Enhance Robustness and Performance in Curvilinear Structure Segmentation

Segmenting curvilinear structures in medical images is essential for analyzing morphological patterns in clinical applications. Integrating topological properties, such as connectivity, improves segmentation accuracy and consistency. However, extracting and embedding such properties - especially from Persistence Diagrams (PD) - is challenging due to their non-differentiability and computational cost. Existing approaches mostly encode topology through handcrafted loss functions, which generalize poorly across tasks. In this paper, we propose PIs-Regressor, a simple yet effective module that learns persistence image (PI) - finite, differentiable representations of topological features - directly from data. Together with Topology SegNet, which fuses these features in both downsampling and upsampling stages, our framework integrates topology into the network architecture itself rather than auxiliary losses. Unlike existing methods that depend heavily on handcrafted loss functions, our approach directly incorporates topological information into the network structure, leading to more robust segmentation. Our design is flexible and can be seamlessly combined with other topology-based methods to further enhance segmentation performance. Experimental results show that integrating topological features enhances model robustness, effectively handling challenges like overexposure and blurring in medical imaging. Our approach on three curvilinear benchmarks demonstrate state-of-the-art performance in both pixel-level accuracy and topological fidelity.

cs.CV↗

arXiv2026

Safeguarding LLM Fine-tuning via Push-Pull Distributional Alignment

The inherent safety alignment of Large Language Models (LLMs) is prone to erosion during fine-tuning, even when using seemingly innocuous datasets. While existing defenses attempt to mitigate this via data selection, they typically rely on heuristic, instance-level assessments that neglect the global geometry of the data distribution and fail to explicitly repel harmful patterns. To address this, we introduce Safety Optimal Transport (SOT), a novel framework that reframes safe fine-tuning from an instance-level filtering challenge to a distribution-level alignment task grounded in Optimal Transport (OT). At its core is a dual-reference ``push-pull'' weight-learning mechanism: SOT optimizes sample importance by actively pulling the downstream distribution towards a trusted safe anchor while simultaneously pushing it away from a general harmful reference. This establishes a robust geometric safety boundary that effectively purifies the training data. Extensive experiments across diverse model families and domains demonstrate that SOT significantly enhances model safety while maintaining competitive downstream performance, achieving a superior safety-utility trade-off compared to baselines.

cs.LG↗