arXiv subjects

Hao Sun

Publications and source records attributed to Hao Sun.

At least 163 records · Page 9Linked to original sources

arXiv2025

Mixed WIMP-FIMP scenario in a two-component dark matter model

We consider the mixed WIMP-FIMP scenario in a two-component dark matter model with $Z_2 \times Z_4$ symmetry, where a singlet scalar $S$ and a Majarano fermion $χ$ are introduced as dark matter candidates. We also introduce another singlet scalar $S_0$ with a non-zero vacuum expectation value to the SM so that the fermion dark matter can obtain mass after spontaneous symmetry breaking. Either $S$ or $χ$ relic density can be generated via the "Freeze-out" mechanism. In contrast, the other DM candidate relic density is obtained by the "Freeze-in" mechanism, and we therefore have two different cases. In the case of $χ$ as WIMP and $S$ as FIMP, we perform random scans to estimate the allowed parameter space consistent with the dark matter constraint. The results show that this case is viable over a wide range of dark matter masses with the Yukawa coupling of $S_0$ and $χ$ should be larger than 1. Instead, for the case of $S$ as WIMP and $χ$ as FIMP, the viable parameter space is more constrained by the direct detection experiements, and we have two regions with $m_S \approx 62.5$ GeV and $m_S>400$ GeV under the constraints, which is consistent with the singlet scalar DM result but the scalar DM mass can be as low as a few hundred GeV for the heavy mass region in the model.

hep-ph↗

arXiv2025

Spatiotemporal Graph Learning with Direct Volumetric Information Passing and Feature Enhancement

Data-driven learning of physical systems has kindled significant attention, where many neural models have been developed. In particular, mesh-based graph neural networks (GNNs) have demonstrated significant potential in modeling spatiotemporal dynamics across arbitrary geometric domains. However, the existing node-edge message-passing and aggregation mechanism in GNNs limits the representation learning ability. In this paper, we proposed a dual-module framework, Cell-embedded and Feature-enhanced Graph Neural Network (aka, CeFeGNN), for learning spatiotemporal dynamics. Specifically, we embed learnable cell attributions to the common node-edge message passing process, which better captures the spatial dependency of regional features. Such a strategy essentially upgrades the local aggregation scheme from first order (e.g., from edge to node) to a higher order (e.g., from volume and edge to node), which takes advantage of volumetric information in message passing. Meanwhile, a novel feature-enhanced block is designed to further improve the model's performance and alleviate the over-smoothness problem. Extensive experiments on various PDE systems and one real-world dataset demonstrate that CeFeGNN achieves superior performance compared with other baselines.

cs.LG↗

arXiv2025

A Set of Generalized Components to Achieve Effective Poison-only Clean-label Backdoor Attacks with Collaborative Sample Selection and Triggers

Poison-only Clean-label Backdoor Attacks aim to covertly inject attacker-desired behavior into DNNs by merely poisoning the dataset without changing the labels. To effectively implant a backdoor, multiple \textbf{triggers} are proposed for various attack requirements of Attack Success Rate (ASR) and stealthiness. Additionally, sample selection enhances clean-label backdoor attacks' ASR by meticulously selecting ``hard'' samples instead of random samples to poison. Current methods 1) usually handle the sample selection and triggers in isolation, leading to severely limited improvements on both ASR and stealthiness. Consequently, attacks exhibit unsatisfactory performance on evaluation metrics when converted to PCBAs via a mere stacking of methods. Therefore, we seek to explore the bidirectional collaborative relations between the sample selection and triggers to address the above dilemma. 2) Since the strong specificity within triggers, the simple combination of sample selection and triggers fails to substantially enhance both evaluation metrics, with generalization preserved among various attacks. Therefore, we seek to propose a set of components to significantly improve both stealthiness and ASR based on the commonalities of attacks. Specifically, Component A ascertains two critical selection factors, and then makes them an appropriate combination based on the trigger scale to select more reasonable ``hard'' samples for improving ASR. Component B is proposed to select samples with similarities to relevant trigger implanted samples to promote stealthiness. Component C reassigns trigger poisoning intensity on RGB colors through distinct sensitivity of the human visual system to RGB for higher ASR, with stealthiness ensured by sample selection, including Component B. Furthermore, all components can be strategically integrated into diverse PCBAs.

cs.CR↗

arXiv2025

Learnable cut flow for high energy physics

Neural networks have emerged as a powerful paradigm for tasks in high energy physics, yet their opaque training process renders them as a black box. In contrast, the traditional cut flow method offers simplicity and interpretability but requires extensive manual tuning to identify optimal cut boundaries. To merge the strengths of both approaches, we propose the Learnable Cut Flow (LCF), a neural network that transforms the traditional cut selection into a fully differentiable, data-driven process. LCF implements two cut strategies-parallel, where observable distributions are treated independently, and sequential, where prior cuts shape subsequent ones-to flexibly determine optimal boundaries. Building on this strategy, we introduce the Learnable Importance, a metric that quantifies feature importance and adjusts their contributions to the loss accordingly, offering model-driven insights unlike ad-hoc metrics. To ensure differentiability, a modified loss function replaces hard cuts with mask operations, preserving data shape throughout the training process. LCF is tested on six varied mock datasets and a realistic diboson vs. QCD dataset. Results demonstrate that LCF 1. accurately learns cut boundaries across typical feature distributions in both parallel and sequential strategies, 2. assigns higher importance to discriminative features with minimal overlap, 3. handles redundant or correlated features robustly, and 4. performs effectively in real-world scenarios. In the diboson dataset, LCF initially underperforms boosted decision trees and multiplayer perceptrons when using all observables. LCF bridges the gap between traditional cut flow method and modern black-box neural networks, delivering actionable insights into the training process and feature importance. Source code and experimental data are available at https://github.com/Star9daisy/learnable-cut-flow.

cs.LG↗

arXiv2025

AnchDrive: Bootstrapping Diffusion Policies with Hybrid Trajectory Anchors for End-to-End Driving

End-to-end multi-modal planning has become a transformative paradigm in autonomous driving, effectively addressing behavioral multi-modality and the generalization challenge in long-tail scenarios. We propose AnchDrive, a framework for end-to-end driving that effectively bootstraps a diffusion policy to mitigate the high computational cost of traditional generative models. Rather than denoising from pure noise, AnchDrive initializes its planner with a rich set of hybrid trajectory anchors. These anchors are derived from two complementary sources: a static vocabulary of general driving priors and a set of dynamic, context-aware trajectories. The dynamic trajectories are decoded in real-time by a Transformer that processes dense and sparse perceptual features. The diffusion model then learns to refine these anchors by predicting a distribution of trajectory offsets, enabling fine-grained refinement. This anchor-based bootstrapping design allows for efficient generation of diverse, high-quality trajectories. Experiments on the NAVSIM benchmark confirm that AnchDrive sets a new state-of-the-art and shows strong generalizability

cs.RO↗

arXiv2025

MAIN: Mutual Alignment Is Necessary for instruction tuning

Instruction tuning has empowered large language models (LLMs) to achieve remarkable performance, yet its success heavily depends on the availability of large-scale, high-quality instruction-response pairs. To meet this demand, various methods have been developed to synthesize data at scale. However, current methods for scaling up data generation often overlook a crucial aspect: the alignment between instructions and responses. We hypothesize that the quality of instruction-response pairs is determined not by the individual quality of each component, but by the degree of mutual alignment. To address this, we propose a Mutual Alignment Framework (MAIN) which enforces coherence between instructions and responses through mutual constraints. We demonstrate that MAIN generalizes well across model architectures and sizes, achieving state-of-the-art performance on LLaMA, Mistral, and Qwen models across diverse benchmarks. This work underscores the critical role of instruction-response alignment in enabling generalizable and high-quality instruction tuning for LLMs. All code is available from our repository.

cs.CL↗

arXiv2025

Flow Matching-Based Active Learning for Radio Map Construction with Low-Altitude UAVs

The employment of unmanned aerial vehicles (UAVs) in the lowaltitude economy necessitates precise and real-time radio maps for reliable communication and safe navigation. However, constructing such maps is hindered by the infeasibility of exhaustive measurements due to UAVs' limited flight endurance. To address this, we propose a novel active learning framework for low-altitude radio map construction based on limited measurements. First, a Plug-and-Play (PnP)-refined flow matching algorithm is introduced, which leverages flow matching as a powerful generative prior within a PnP scheme to reconstruct high-fidelity radio maps. Second, the generative nature of flow matching is exploited to quantify uncertainty by generating an ensemble of radio maps and computing the location-wise variance. The resulting uncertainty map guides a multi-objective candidate selection and then a trajectory is planned via utility-aware path search (UAPS), directing the UAV to the most informative locations while taking travel costs into account. Simulation results demonstrate that our method significantly outperforms the baselines, achieving more than a 70% reduction in normalized mean squared error (NMSE).

eess.SP↗

arXiv2025

FlowDrive: Energy Flow Field for End-to-End Autonomous Driving

Recent advances in end-to-end autonomous driving leverage multi-view images to construct BEV representations for motion planning. In motion planning, autonomous vehicles need considering both hard constraints imposed by geometrically occupied obstacles (e.g., vehicles, pedestrians) and soft, rule-based semantics with no explicit geometry (e.g., lane boundaries, traffic priors). However, existing end-to-end frameworks typically rely on BEV features learned in an implicit manner, lacking explicit modeling of risk and guidance priors for safe and interpretable planning. To address this, we propose FlowDrive, a novel framework that introduces physically interpretable energy-based flow fields-including risk potential and lane attraction fields-to encode semantic priors and safety cues into the BEV space. These flow-aware features enable adaptive refinement of anchor trajectories and serve as interpretable guidance for trajectory generation. Moreover, FlowDrive decouples motion intent prediction from trajectory denoising via a conditional diffusion planner with feature-level gating, alleviating task interference and enhancing multimodal diversity. Experiments on the NAVSIM v2 benchmark demonstrate that FlowDrive achieves state-of-the-art performance with an EPDMS of 86.3, surpassing prior baselines in both safety and planning quality. The project is available at https://astrixdrive.github.io/FlowDrive.github.io/.

cs.RO↗

arXiv2025

Discovering physical laws with parallel symbolic enumeration

Symbolic regression plays a crucial role in modern scientific research thanks to its capability of discovering concise and interpretable mathematical expressions from data. A key challenge lies in the search for parsimonious and generalizable mathematical formulas, in an infinite search space, while intending to fit the training data. Existing algorithms have faced a critical bottleneck of accuracy and efficiency over a decade when handling problems of complexity, which essentially hinders the pace of applying symbolic regression for scientific exploration across interdisciplinary domains. To this end, we introduce parallel symbolic enumeration (PSE) to efficiently distill generic mathematical expressions from limited data. Experiments show that PSE achieves higher accuracy and faster computation compared to the state-of-the-art baseline algorithms across over 200 synthetic and experimental problem sets (e.g., improving the recovery accuracy by up to 99% and reducing runtime by an order of magnitude). PSE represents an advance in accurate and efficient data-driven discovery of symbolic, interpretable models (e.g., underlying physical laws), and improves the scalability of symbolic learning.

cs.LG↗

arXiv2025

Large Language Models Might Not Care What You Are Saying: Prompt Format Beats Descriptions

With the help of in-context learning (ICL), large language models (LLMs) have achieved impressive performance across various tasks. However, the function of descriptive instructions during ICL remains under-explored. In this work, we propose an ensemble prompt framework to describe the selection criteria of multiple in-context examples, and preliminary experiments on machine translation (MT) across six translation directions confirm that this framework boosts ICL performance. But to our surprise, LLMs might not care what the descriptions actually say, and the performance gain is primarily caused by the ensemble format, since it could lead to improvement even with random descriptive nouns. We further apply this new ensemble framework on a range of commonsense, math, logical reasoning and hallucination tasks with three LLMs and achieve promising results, suggesting again that designing a proper prompt format would be much more effective and efficient than paying effort into specific descriptions.

cs.CL↗

arXiv2025

Almost Tight Approximation Hardness and Online Algorithms for Resource Scheduling

We study the precedence-constrained resource scheduling problem [SICOMP'75]. There are $n$ jobs where each job takes a certain time to finish and has a resource requirement throughout the execution time. There are precedence among the jobs. The problem asks that given a resource budget, schedule the jobs obeying the precedence constraints to minimize makespan (maximum completion time of a job) such that at any point in time, the total resource being used by all the jobs is at most the given resource budget. In the offline setting, an important open question is whether a polynomial-time $O(1)$-factor approximation algorithm can be found. We prove almost tight hardness of approximation: For some constant $α> 0$, there is no $o((\log t_{\max})^α)$-factor ( or $o( ( \log n )^α)$-factor ) approximation algorithm with $n$ jobs of maximum job length $t_{\max}$, unless P = NP ( or NP $\subset$ DTIME$(O( 2^{\text{polylog}(n)}))$ ). We further show a connection between this scheduling problem and a seemingly unrelated problem called the shortest common super-sequence (SCS) problem, which has wide application in Biology and Genomics. We prove that an $o(\log t_{\max})$-factor approximation of the scheduling problem would imply the existence of an $o(|Σ|)$-approximation algorithm for SCS with alphabet $Σ$. We then consider the online setting. We present $Ω(\log n)$ and $Ω(\log t_{\max})$ lower bounds of the competitive ratio of any randomized online algorithm. Moreover, we present a matching $O(\min\{\log n, \log t_{\max}\})$-competitive deterministic online algorithm.

cs.DS↗

arXiv2025

One Framework to Rule Them All: Unifying Multimodal Tasks with LLM Neural-Tuning

Large-scale models have exhibited remarkable capabilities across diverse domains, including automated medical services and intelligent customer support. However, as most large models are trained on single-modality corpora, enabling them to effectively process and understand multimodal signals remains a significant challenge. Current research often focuses on designing task-specific or scenario-specific tuning strategies, which limits the scalability and versatility. To address this limitation, we propose a unified framework that concurrently handles multiple tasks and modalities. In this framework, all modalities and tasks are represented as unified tokens and trained using a single, consistent approach. To enable efficient multitask processing, we introduce a novel tuning strategy termed neural tuning, inspired by the concept of sparse distributed representation in the human brain, where only specific subsets of neurons are activated for each task. Furthermore, to advance research in multimodal and multitask learning, we present a new benchmark, MMUD, which includes samples annotated with multiple task labels spanning reasoning segmentation, referring segmentation, image captioning, and text-to-image generation. By applying neural tuning to pretrained large models on the MMUD benchmark, we demonstrate the ability to handle multiple tasks simultaneously in a streamlined and efficient manner. All models, code, and datasets will be released publicly upon publication, fostering further research and innovation in this field.

cs.CV↗

arXiv2025

Prospects for $P$ and $C\!P$ violation in $Λ_{c}^{+}$ decays with polarized beam at Super Tau-Charm Facility

Weak decays of the charmed baryons offer an ideal platform to study parity ($P$) and charge conjugation-parity ($C\!P$) violation in quark sector, to stringently test the Standard Model and search for new physics. It is a key goal in the next-generation positron-electron collider, such as the Super Tau-Charm Facility (STCF). Thanks to the quantum-entangled pair production with super high luminosity and the possibility of beam polarization, STCF provides a unique environment to probe such symmetry violations with unprecedented sensitivity. In this paper, we evaluate the precisions of the $P$-violating parameters and subsequently obtain the expected sensitivity of the $C\!P$-violating parameters in charmed baryon decays of $Λ_{c}^{+}\to pπ^0$, $pη$, $ΛK^+$, $Σ^0 K^+$, and $Σ^+ K^0_S$, regarding to different polarization setups in STCF. The study suggests that the implementation of longitudinal beam polarization in STCF would greatly enhance the experimental capability in studying $P$ and $C\!P$ violation.

hep-ph↗

arXiv2025

IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model

Vision-Language-Action (VLA) models have demonstrated potential in autonomous driving. However, two critical challenges hinder their development: (1) Existing VLA architectures are typically based on imitation learning in open-loop setup which tends to capture the recorded behaviors in the dataset, leading to suboptimal and constrained performance, (2) Close-loop training relies heavily on high-fidelity sensor simulation, where domain gaps and computational inefficiencies pose significant barriers. In this paper, we introduce IRL-VLA, a novel close-loop Reinforcement Learning via \textbf{I}nverse \textbf{R}einforcement \textbf{L}earning reward world model with a self-built VLA approach. Our framework proceeds in a three-stage paradigm: In the first stage, we propose a VLA architecture and pretrain the VLA policy via imitation learning. In the second stage, we construct a lightweight reward world model via inverse reinforcement learning to enable efficient close-loop reward computation. To further enhance planning performance, finally, we design specialized reward world model guidence reinforcement learning via PPO(Proximal Policy Optimization) to effectively balance the safety incidents, comfortable driving, and traffic efficiency. Our approach achieves state-of-the-art performance in NAVSIM v2 end-to-end driving benchmark, 1st runner up in CVPR2025 Autonomous Grand Challenge. We hope that our framework will accelerate VLA research in close-loop autonomous driving.

cs.AI↗

arXiv2025

OmniFluids: Physics Pre-trained Modeling of Fluid Dynamics

Computational fluid dynamics (CFD) drives progress in numerous scientific and engineering fields, yet high-fidelity simulations remain computationally prohibitive. While machine learning approaches offer computing acceleration, they typically specialize in single physical systems or require extensive training data, hindering their applicability in highly nonlinear and 3D flow scenarios. To overcome these limitations, we propose OmniFluids, a pure physics pre-trained model that captures fundamental fluid dynamics laws and adapts efficiently to diverse downstream tasks with minimal data. We develop a training framework combining physics-only pre-training, coarse-grid operator distillation, and few-shot fine-tuning. This enables OmniFluids to retain broad physics knowledge while delivering fast and accurate predictions. Architecturally, OmniFluids integrates a mixture of operators, a multi-frame decoder, and factorized Fourier layers, seamlessly incorporating physics-based supervision while allowing efficient and scalable modeling of diverse tasks. Extensive tests on a broad range of 2D and 3D benchmarks show that OmniFluids outperforms state-of-the-art AI-driven methods in terms of flow field prediction and turbulence statistics. It delivers 10--100$\times$ speedups over traditional solvers while maintaining a comparable accuracy and accurately identifies unknown physical parameters from sparse, noisy data. This work demonstrates the potential of training a unified CFD solver exclusively from physics knowledge, offering a new approach for efficient and generalizable modeling across complex fluid systems.

physics.flu-dyn↗

arXiv2025

Physical Layer Authentication Based on Hierarchical Variational Auto-Encoder for Industrial Internet of Things

Recently, Physical Layer Authentication (PLA) has attracted much attention since it takes advantage of the channel randomness nature of transmission media to achieve communication confidentiality and authentication. In the complex environment, such as the Industrial Internet of Things (IIoT), machine learning (ML) is widely employed with PLA to extract and analyze complex channel characteristics for identity authentication. However, most PLA schemes for IIoT require attackers' prior channel information, leading to severe performance degradation when the source of the received signals is unknown in the training stage. Thus, a channel impulse response (CIR)-based PLA scheme named "Hierarchical Variational Auto-Encoder (HVAE)" for IIoT is proposed in this article, aiming at achieving high authentication performance without knowing attackers' prior channel information even when trained on a few data in the complex environment. HVAE consists of an Auto-Encoder (AE) module for CIR characteristics extraction and a Variational Auto-Encoder (VAE) module for improving the representation ability of the CIR characteristic and outputting the authentication results. Besides, a new objective function is constructed in which both the single-peak and the double-peak Gaussian distribution are taken into consideration in the VAE module. Moreover, the simulations are conducted under the static and mobile IIoT scenario, which verify the superiority of the proposed HVAE over three comparison PLA schemes even with a few training data.

eess.SP↗

arXiv2025

SVGen: Interpretable Vector Graphics Generation with Large Language Models

Scalable Vector Graphics (SVG) is widely used in front-end development and UI/UX design due to its scalability, editability, and rendering efficiency. However, turning creative ideas into precise vector graphics remains a time-consuming challenge. To address this, we introduce SVG-1M, a large-scale dataset of high-quality SVGs paired with natural language descriptions. Through advanced data augmentation and annotation, we create well-aligned Text to SVG training pairs, including a subset with Chain of Thought annotations for enhanced semantic guidance. Based on this dataset, we propose SVGen, an end-to-end model that generates SVG code from natural language inputs. Our approach ensures semantic accuracy and structural completeness, supported by curriculum learning and reinforcement learning optimization. Experiments show that SVGen outperforms general large models and traditional rendering methods in both effectiveness and efficiency. Code, model, and dataset are available on GitHub.

cs.LG↗

arXiv2025

DiffSemanticFusion: Semantic Raster BEV Fusion for Autonomous Driving via Online HD Map Diffusion

Autonomous driving requires accurate scene understanding, including road geometry, traffic agents, and their semantic relationships. In online HD map generation scenarios, raster-based representations are well-suited to vision models but lack geometric precision, while graph-based representations retain structural detail but become unstable without precise maps. To harness the complementary strengths of both, we propose DiffSemanticFusion -- a fusion framework for multimodal trajectory prediction and planning. Our approach reasons over a semantic raster-fused BEV space, enhanced by a map diffusion module that improves both the stability and expressiveness of online HD map representations. We validate our framework on two downstream tasks: trajectory prediction and planning-oriented end-to-end autonomous driving. Experiments on real-world autonomous driving benchmarks, nuScenes and NAVSIM, demonstrate improved performance over several state-of-the-art methods. For the prediction task on nuScenes, we integrate DiffSemanticFusion with the online HD map informed QCNet, achieving a 5.1\% performance improvement. For end-to-end autonomous driving in NAVSIM, DiffSemanticFusion achieves state-of-the-art results, with a 15\% performance gain in NavHard scenarios. In addition, extensive ablation and sensitivity studies show that our map diffusion module can be seamlessly integrated into other vector-based approaches to enhance performance. All artifacts are available at https://github.com/SunZhigang7/DiffSemanticFusion.

cs.CV↗