arXiv subjects

Bo Wang

Publications and source records attributed to Bo Wang.

At least 217 records · Page 12Linked to original sources

arXiv2025

DualRAG: A Dual-Process Approach to Integrate Reasoning and Retrieval for Multi-Hop Question Answering

Multi-Hop Question Answering (MHQA) tasks permeate real-world applications, posing challenges in orchestrating multi-step reasoning across diverse knowledge domains. While existing approaches have been improved with iterative retrieval, they still struggle to identify and organize dynamic knowledge. To address this, we propose DualRAG, a synergistic dual-process framework that seamlessly integrates reasoning and retrieval. DualRAG operates through two tightly coupled processes: Reasoning-augmented Querying (RaQ) and progressive Knowledge Aggregation (pKA). They work in concert: as RaQ navigates the reasoning path and generates targeted queries, pKA ensures that newly acquired knowledge is systematically integrated to support coherent reasoning. This creates a virtuous cycle of knowledge enrichment and reasoning refinement. Through targeted fine-tuning, DualRAG preserves its sophisticated reasoning and retrieval capabilities even in smaller-scale models, demonstrating its versatility and core advantages across different scales. Extensive experiments demonstrate that this dual-process approach substantially improves answer accuracy and coherence, approaching, and in some cases surpassing, the performance achieved with oracle knowledge access. These results establish DualRAG as a robust and efficient solution for complex multi-hop reasoning tasks.

cs.LG↗

arXiv2025

Erdos-Turan photonic Ising machines with record-high coupling resolution

Ising machines have emerged as promising platforms for efficiently tackling a wide range of combinatorial optimization problems relevant to resource allocation, statistical inference and deep learning, yet their practical utility is fundamentally constrained by the coarse resolution of spin-spin couplings (Jij). Current implementations, relying on direct modulation of physical parameters, achieve at most 256 discrete coupling levels, which severely hinder the faithfully modeling of arbitrary real-valued interactions in realistic applications. Here we present a novel photonic Ising machine that encodes spins in random lattices while programming couplings in the momentum space of light. By introducing the Sidon set-a mathematical structure ensuring pairwise difference uniqueness - and employing the Erdos-Turan bound, we establish an optical framework in which each spin pair can be assigned a unique Jij. This approach decouples the resolution limit from hardware modulation to the spatial precision in the momentum space of light. Experimentally, we demonstrate a record-high coupling resolution of 7,038 on a simple photonic platform, surpassing previous Ising machines. Our results highlight the power of uniting discrete mathematics with momentum-space photonics, paving the way toward scalable Ising machines capable of faithfully modeling real-world optimization problems.

physics.optics↗

arXiv2025

Revisiting Visual Understanding in Multimodal Reasoning through a Lens of Image Perturbation

Despite the rapid progress of multimodal large language models (MLLMs), they have largely overlooked the importance of visual processing. In a simple yet revealing experiment, we interestingly find that language-only models, when provided with image captions, can achieve comparable or even better performance than MLLMs that consume raw visual inputs. This suggests that current MLLMs may generate accurate visual descriptions but fail to effectively integrate them during reasoning. Motivated by this, we propose a simple visual perturbation framework that enhances perceptual robustness without requiring algorithmic modifications or additional training data. Our approach introduces three targeted perturbations: distractor concatenation, dominance-preserving mixup, and random rotation, that can be easily integrated into existing post-training pipelines including SFT, DPO, and GRPO. Through extensive experiments across multiple datasets, we demonstrate consistent improvements in mathematical reasoning performance, with gains comparable to those achieved through algorithmic changes. Additionally, we achieve competitive performance among open-source 7B RL-tuned models by training Qwen2.5-VL-7B with visual perturbation. Through comprehensive ablation studies, we analyze the effectiveness of different perturbation strategies, revealing that each perturbation type contributes uniquely to different aspects of visual reasoning. Our findings highlight the critical role of visual perturbation in multimodal mathematical reasoning: better reasoning begins with better seeing. Our code is available at https://github.com/YutingLi0606/Vision-Matters.

cs.CV↗

arXiv2025

Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling

We introduce InternVL 2.5, an advanced multimodal large language model (MLLM) series that builds upon InternVL 2.0, maintaining its core model architecture while introducing significant enhancements in training and testing strategies as well as data quality. In this work, we delve into the relationship between model scaling and performance, systematically exploring the performance trends in vision encoders, language models, dataset sizes, and test-time configurations. Through extensive evaluations on a wide range of benchmarks, including multi-discipline reasoning, document understanding, multi-image / video understanding, real-world comprehension, multimodal hallucination detection, visual grounding, multilingual capabilities, and pure language processing, InternVL 2.5 exhibits competitive performance, rivaling leading commercial models such as GPT-4o and Claude-3.5-Sonnet. Notably, our model is the first open-source MLLMs to surpass 70% on the MMMU benchmark, achieving a 3.7-point improvement through Chain-of-Thought (CoT) reasoning and showcasing strong potential for test-time scaling. We hope this model contributes to the open-source community by setting new standards for developing and applying multimodal AI systems. HuggingFace demo see https://huggingface.co/spaces/OpenGVLab/InternVL

cs.CV↗

arXiv2025

Helium-burning blue large-amplitude pulsators: A Population Study with BPASS

Blue Large-Amplitude Pulsators (BLAPs) are a class of radially pulsating stars with effective temperatures ranging from 20,000 to 35,000 K and pulsation periods between 7 and 75 minutes. This study utilizes the Binary Population and Spectral Synthesis (BPASS) code to investigate helium-burning stars as a formation channel for BLAPs in the Milky Way. The progenitor stars have initial masses of 3-6 $M_{\odot}$, resulting in BLAPs with final masses of 0.5-1.2 $M_{\odot}$. Based on a constant star formation rate of 3 $ M_{\odot}\text{yr}^{-1}$ and solar metallicity (Z = 0.020), population synthesis predicts approximately 14,351 helium-burning BLAPs in the Milky Way: 12,799 with Main Sequence (MS) companions and 1,551 with evolved/compact-object companions. Helium-burning BLAPs show prolonged lifetimes in the pulsation region and a narrow stellar age range for entering this regime (log(t/yr) = 8.0-8.6), unlike pre-white dwarf models. BLAPs with MS companions typically form via Roche lobe overflow, leading to longer orbital periods ($\sim$100 days). Those with evolved/compact-object companions form through common envelope evolution, resulting in shorter periods. While Galactic extinction makes most BLAPs faint (apparent magnitudes $>$ 25), future surveys like WFST and VRO LSST are expected to detect approximately 500-900. This research establishes helium-burning stars as a significant BLAP contributor and offers testable predictions regarding their binary properties and Galactic distribution.

astro-ph.SR↗

arXiv2025

IDEATOR: Jailbreaking and Benchmarking Large Vision-Language Models Using Themselves

As large Vision-Language Models (VLMs) gain prominence, ensuring their safe deployment has become critical. Recent studies have explored VLM robustness against jailbreak attacks-techniques that exploit model vulnerabilities to elicit harmful outputs. However, the limited availability of diverse multimodal data has constrained current approaches to rely heavily on adversarial or manually crafted images derived from harmful text datasets, which often lack effectiveness and diversity across different contexts. In this paper, we propose IDEATOR, a novel jailbreak method that autonomously generates malicious image-text pairs for black-box jailbreak attacks. IDEATOR is grounded in the insight that VLMs themselves could serve as powerful red team models for generating multimodal jailbreak prompts. Specifically, IDEATOR leverages a VLM to create targeted jailbreak texts and pairs them with jailbreak images generated by a state-of-the-art diffusion model. Extensive experiments demonstrate IDEATOR's high effectiveness and transferability, achieving a 94% attack success rate (ASR) in jailbreaking MiniGPT-4 with an average of only 5.34 queries, and high ASRs of 82%, 88%, and 75% when transferred to LLaVA, InstructBLIP, and Chameleon, respectively. Building on IDEATOR's strong transferability and automated process, we introduce the VLJailbreakBench, a safety benchmark comprising 3,654 multimodal jailbreak samples. Our benchmark results on 11 recently released VLMs reveal significant gaps in safety alignment. For instance, our challenge set achieves ASRs of 46.31% on GPT-4o and 19.65% on Claude-3.5-Sonnet, underscoring the urgent need for stronger defenses. VLJailbreakBench is publicly available at https://roywang021.github.io/VLJailbreakBench.

cs.CV↗

arXiv2025

Exploring the Design Space of 3D MLLMs for CT Report Generation

Multimodal Large Language Models (MLLMs) have emerged as a promising way to automate Radiology Report Generation (RRG). In this work, we systematically investigate the design space of 3D MLLMs, including visual input representation, projectors, Large Language Models (LLMs), and fine-tuning techniques for 3D CT report generation. We also introduce two knowledge-based report augmentation methods that improve performance on the GREEN score by up to 10%, achieving the 2nd place on the MICCAI 2024 AMOS-MM challenge. Our results on the 1,687 cases from the AMOS-MM dataset show that RRG is largely independent of the size of LLM under the same training protocol. We also show that larger volume size does not always improve performance if the original ViT was pre-trained on a smaller volume size. Lastly, we show that using a segmentation mask along with the CT volume improves performance. The code is publicly available at https://github.com/bowang-lab/AMOS-MM-Solution

eess.IV↗

arXiv2025

Dynamic Expert Specialization: Towards Catastrophic Forgetting-Free Multi-Domain MoE Adaptation

Mixture-of-Experts (MoE) models offer immense capacity via sparsely gated expert subnetworks, yet adapting them to multiple domains without catastrophic forgetting remains an open challenge. Existing approaches either incur prohibitive computation, suffer cross-domain interference, or require separate runs per domain. We propose DES-MoE, a dynamic expert specialization framework for multi-domain adaptation of Mixture-of-Experts models. DES-MoE addresses catastrophic forgetting through three innovations: (1) an adaptive router balancing pre-trained knowledge retention and task-specific updates via distillation, (2) real-time expert-domain correlation mapping to isolate domain-specific gradients, and (3) a three-phase adaptive fine-tuning schedule that progressively freezes non-specialized parameters. Evaluated on six domains (math, code, law, etc.), DES-MoE matches single-domain ESFT performance while training one unified model, reduces forgetting by 89% compared to full fine-tuning as domains scale from 2 to 6, and achieves 68% faster convergence than conventional methods. Our work establishes dynamic expert isolation as a scalable paradigm for multi-task MoE adaptation.

cs.LG↗

arXiv2025

The Impact of ${}^{12} \mathrm{C}(α, γ)^{16} \mathrm{O}$ Reaction on the Evolution of He Stars

The ${}^{12} \mathrm{C}(α, γ)^{16} \mathrm{O}$ reaction is one of the most important reactions in the evolution of massive stars, yet its rate is still highly uncertain. In this work, we investigated how variations in the ${}^{12} \mathrm{C}(α, γ)^{16} \mathrm{O}$ reaction rate affect the evolution of a 14 $\rm M_{\odot}$ He star using the MESA code. Our simulations indicate that the ${}^{12} \mathrm{C}(α, γ)^{16} \mathrm{O}$ reaction rate determines the conditions for C burning, affecting its explodability. As the reaction rate increases, central C-burning becomes neutrino-dominated, transitioning from the convective to the radiative regime. This leads to higher compactness and a larger iron core, indicating a more compact pre-SN core structure that is difficult to explode. Conversely, lower reaction rates shorten the C-burning lifetime and trigger earlier central Ne ignition, which counteracts core contraction. This results in reduced compactness and iron core mass. We also found that variations in reaction rates shift the location of the last C-burning shell. When this shell exceeds the mass coordinate used for compactness evaluation, the overall compactness increases significantly. Although the Si- and O-burning convective shells decrease compactness, the overall increase remains unaffected. This work suggests that the ${}^{12} \mathrm{C}(α, γ)^{16} \mathrm{O}$ reaction play an important role in the pre-SN core structure and potentially impact the explodability of massive He stars.

astro-ph.SR↗

arXiv2025

ScenGAN: Attention-Intensive Generative Model for Uncertainty-Aware Renewable Scenario Forecasting

To address the intermittency of renewable energy source (RES) generation, scenario forecasting offers a series of stochastic realizations for predictive objects with superior flexibility and direct views. Based on a long time-series perspective, this paper explores uncertainties in the realms of renewable power and deep learning. Then, an uncertainty-aware model is meticulously designed for renewable scenario forecasting, which leverages an attention mechanism and generative adversarial networks (GANs) to precisely capture complex spatial-temporal dynamics. To improve the interpretability of uncertain behavior in RES generation, Bayesian deep learning and adaptive instance normalization (AdaIN) are incorporated to simulate typical patterns and variations. Additionally, the integration of meteorological information, forecasts, and historical trajectories in the processing layer improves the synergistic forecasting capability for multiscale periodic regularities. Numerical experiments and case analyses demonstrate that the proposed approach provides an appropriate interpretation for renewable uncertainty representation, including both aleatoric and epistemic uncertainties, and shows superior performance over state-of-the-art methods.

cs.LG↗

arXiv2025

Introduction to the Chinese Space Station Survey Telescope (CSST)

The Chinese Space Station Survey Telescope (CSST) is an upcoming Stage-IV sky survey telescope, distinguished by its large field of view (FoV), high image quality, and multi-band observation capabilities. It can simultaneously conduct precise measurements of the Universe by performing multi-color photometric imaging and slitless spectroscopic surveys. The CSST is equipped with five scientific instruments, i.e. Multi-band Imaging and Slitless Spectroscopy Survey Camera (SC), Multi-Channel Imager (MCI), Integral Field Spectrograph (IFS), Cool Planet Imaging Coronagraph (CPI-C), and THz Spectrometer (TS). Using these instruments, CSST is expected to make significant contributions and discoveries across various astronomical fields, including cosmology, galaxies and active galactic nuclei (AGN), the Milky Way and nearby galaxies, stars, exoplanets, Solar System objects, astrometry, and transients and variable sources. This review aims to provide a comprehensive overview of the CSST instruments, observational capabilities, data products, and scientific potential.

astro-ph.IM↗

arXiv2025

LongCat-Flash Technical Report

We introduce LongCat-Flash, a 560-billion-parameter Mixture-of-Experts (MoE) language model designed for both computational efficiency and advanced agentic capabilities. Stemming from the need for scalable efficiency, LongCat-Flash adopts two novel designs: (a) Zero-computation Experts, which enables dynamic computational budget allocation and activates 18.6B-31.3B (27B on average) per token depending on contextual demands, optimizing resource usage. (b) Shortcut-connected MoE, which enlarges the computation-communication overlap window, demonstrating notable gains in inference efficiency and throughput compared to models of a comparable scale. We develop a comprehensive scaling framework for large models that combines hyperparameter transfer, model-growth initialization, a multi-pronged stability suite, and deterministic computation to achieve stable and reproducible training. Notably, leveraging the synergy among scalable architectural design and infrastructure efforts, we complete model training on more than 20 trillion tokens within 30 days, while achieving over 100 tokens per second (TPS) for inference at a cost of \$0.70 per million output tokens. To cultivate LongCat-Flash towards agentic intelligence, we conduct a large-scale pre-training on optimized mixtures, followed by targeted mid- and post-training on reasoning, code, and instructions, with further augmentation from synthetic data and tool use tasks. Comprehensive evaluations demonstrate that, as a non-thinking foundation model, LongCat-Flash delivers highly competitive performance among other leading models, with exceptional strengths in agentic tasks. The model checkpoint of LongCat-Flash is open-sourced to foster community research. LongCat Chat: https://longcat.ai Hugging Face: https://huggingface.co/meituan-longcat GitHub: https://github.com/meituan-longcat

cs.CL↗

arXiv2025

End4: End-to-end Denoising Diffusion for Diffusion-Based Inpainting Detection

The powerful generative capabilities of diffusion models have significantly advanced the field of image synthesis, enhancing both full image generation and inpainting-based image editing. Despite their remarkable advancements, diffusion models also raise concerns about potential misuse for malicious purposes. However, existing approaches struggle to identify images generated by diffusion-based inpainting models, even when similar inpainted images are included in their training data. To address this challenge, we propose a novel detection method based on End-to-end denoising diffusion (End4). Specifically, End4 designs a denoising reconstruction model to improve the alignment degree between the latent spaces of the reconstruction and detection processes, thus reconstructing features that are more conducive to detection. Meanwhile, it leverages a Scale-aware Pyramid-like Fusion Module (SPFM) that refines local image features under the guidance of attention pyramid layers at different scales, enhancing feature discriminability. Additionally, to evaluate detection performance on inpainted images, we establish a comprehensive benchmark comprising images generated from five distinct masked regions. Extensive experiments demonstrate that our End4 effectively generalizes to unseen masking patterns and remains robust under various perturbations. Our code and dataset will be released soon.

cs.CV↗

arXiv2025

The enhanced X-ray Timing and Polarimetry mission -- eXTP for launch in 2030

In this paper we present the current status of the enhanced X-ray Timing and Polarimetry mission, which has been fully approved for launch in 2030. eXTP is a space science mission designed to study fundamental physics under extreme conditions of matter density, gravity, and magnetism. The mission aims at determining the equation of state of matter at supra-nuclear density, measuring the effects of quantum electro-dynamics, and understanding the dynamics of matter in strong-field gravity. In addition to investigating fundamental physics, the eXTP mission is poised to become a leading observatory for time-domain and multi-messenger astronomy in the 2030's, as well as providing observations of unprecedented quality on a variety of galactic and extragalactic objects. After briefly introducing the history and a summary of the scientific objectives of the eXTP mission, this paper presents a comprehensive overview of: 1) the cutting-edge technology, technical specifications, and anticipated performance of the mission's scientific instruments; 2) the full mission profile, encompassing spacecraft design, operational capabilities, and ground segment infrastructure.

astro-ph.HE↗

arXiv2025

BadPromptFL: A Novel Backdoor Threat to Prompt-based Federated Learning in Multimodal Models

Prompt-based tuning has emerged as a lightweight alternative to full fine-tuning in large vision-language models, enabling efficient adaptation via learned contextual prompts. This paradigm has recently been extended to federated learning settings (e.g., PromptFL), where clients collaboratively train prompts under data privacy constraints. However, the security implications of prompt-based aggregation in federated multimodal learning remain largely unexplored, leaving a critical attack surface unaddressed. In this paper, we introduce \textbf{BadPromptFL}, the first backdoor attack targeting prompt-based federated learning in multimodal contrastive models. In BadPromptFL, compromised clients jointly optimize local backdoor triggers and prompt embeddings, injecting poisoned prompts into the global aggregation process. These prompts are then propagated to benign clients, enabling universal backdoor activation at inference without modifying model parameters. Leveraging the contextual learning behavior of CLIP-style architectures, BadPromptFL achieves high attack success rates (e.g., $>90\%$) with minimal visibility and limited client participation. Extensive experiments across multiple datasets and aggregation protocols validate the effectiveness, stealth, and generalizability of our attack, raising critical concerns about the robustness of prompt-based federated learning in real-world deployments.

cs.LG↗

arXiv2025

A Type 2 Fuzzy Set Approach for Building Linear Linguistic Regression Analysis under Multi Uncertainty

In this paper, we propose a novel heuristic algorithm for constructing a Type-2 Fuzzy Set of the Linear Linguistic Regression (T2F-LLR) model, designed to address uncertainty and vagueness in real-world decision-making. We consider a practical scenario involving a cosmetic company's promotional planning across four product categories: Basic Face Care, Face Cleaning, Cosmetics, and Body Care, aimed at both male and female consumers. Data are collected using fuzzy linguistic questionnaires from customers and expert managers, with responses expressed using qualitative terms such as 'always', 'frequently', 'Often', 'Sometimes', and 'frequently'. These linguistic evaluations are modeled as Type-2 Fuzzy Set of Linear Linguistic regression (T2F-LLR) variables to capture both randomness and higher-order fuzziness. We rigorously develop a solution framework based on a one-sigma confidence interval using the credibility measure to calculate the expected values and variances of the model output. To improve computational efficiency and usability of decisions, we introduce a heuristic algorithm tailored for non-meta datasets, significantly reducing the complexity of the model solving process. The experimental results demonstrate the effectiveness of our approach, which yields a mean absolute percentage error (MAPE) of weight equals 7.97\% with all variables statistically significant. We also provide the significant results for each product using the one-way analysis of variance test (one-way ANOVA test) ($p$-value = 0.15) and the paired $t$ test ($p$-value = 0.16). The results show that there is no significant difference between observed and predicted weights overall. This paper provides a robust and interpretable methodology for decision makers dealing with imprecise data and time-sensitive planning.

math.GM↗

arXiv2025

FraGNNet: A Deep Probabilistic Model for Tandem Mass Spectrum Prediction

Compound identification from tandem mass spectrometry (MS/MS) data is a critical step in the analysis of complex mixtures. Typical solutions for the MS/MS spectrum to compound (MS2C) problem involve comparing the unknown spectrum against a library of known spectrum-molecule pairs, an approach that is limited by incomplete library coverage. Compound to MS/MS spectrum (C2MS) models can improve retrieval rates by augmenting real libraries with predicted MS/MS spectra. Unfortunately, many existing C2MS models suffer from problems with mass accuracy, generalization, or interpretability. We develop a new probabilistic method for C2MS prediction, FraGNNet, that can efficiently and accurately simulate MS/MS spectra with high mass accuracy. Our approach formulates the C2MS problem as learning a distribution over molecule fragments. FraGNNet achieves state-of-the-art performance in terms of prediction error and surpasses existing C2MS models as a tool for retrieval-based MS2C.

cs.LG↗

arXiv2025

Boosting Redundancy-based Automated Program Repair by Fine-grained Pattern Mining

Redundancy-based automated program repair (APR), which generates patches by referencing existing source code, has gained much attention since they are effective in repairing real-world bugs with good interpretability. However, since existing approaches either demand the existence of multi-line similar code or randomly reference existing code, they can only repair a small number of bugs with many incorrect patches, hindering their wide application in practice. In this work, we aim to improve the effectiveness of redundancy-based APRs by exploring more effective source code reuse methods for improving the number of correct patches and reducing incorrect patches. Specifically, we have proposed a new repair technique named Repatt, which incorporates a two-level pattern mining process for guiding effective patch generation (i.e., token and expression levels). We have conducted an extensive experiment on the widely-used Defects4J benchmark and compared Repatt with ten state-of-the-art APR approaches. The results show that it complements existing approaches by repairing 9 unique bugs compared with the latest Large Language Model (LLM)-based and deep learning-based methods and 19 unique bugs compared with traditional repair methods when providing the perfect fault localization. In addition, when the perfect fault localization is unknown in real practice, Repatt significantly outperforms the baseline approaches by achieving much higher patch precision, i.e., 83.8\%, although it repairs fewer bugs. Moreover, we further proposed an effective patch ranking strategy for combining the strength of Repatt and the baseline methods. The result shows that it repairs 124 bugs when only considering the Top-1 patches and improves the best-performing repair method by repairing 39 more bugs. The results demonstrate the effectiveness of our approach for practical use.

cs.SE↗