Search

Low-rank Adaptation (LoRA) has demonstrated remarkable capabilities for task specific fine-tuning. However, in scenarios that involve multiple tasks, training a separate LoRA model for each one results in considerable inefficiency in terms of storage and inference. Moreover, existing parameter generation methods fail to capture the correlations among these tasks, making multi-task LoRA parameter generation challenging. To address these limitations, we propose In-Context Meta LoRA (ICM-LoRA), a novel approach that efficiently achieves task-specific customization of large language models (LLMs). Specifically, we use training data from all tasks to train a tailored generator, Conditional Variational Autoencoder (CVAE). CVAE takes task descriptions as inputs and produces task-aware LoRA weights as outputs. These LoRA weights are then merged with LLMs to create task-specialized models without the need for additional fine-tuning. Furthermore, we utilize in-context meta-learning for knowledge enhancement and task mapping, to capture the relationship between tasks and parameter distributions. As a result, our method achieves more accurate LoRA parameter generation for diverse tasks using CVAE. ICM-LoRA enables more accurate LoRA parameter reconstruction than current parameter reconstruction methods and is useful for implementing task-specific enhancements of LoRA parameters. At the same time, our method occupies 283MB, only 1\% storage compared with the original LoRA.

We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as a preliminary step, demonstrates remarkable reasoning capabilities. Through RL, DeepSeek-R1-Zero naturally emerges with numerous powerful and intriguing reasoning behaviors. However, it encounters challenges such as poor readability, and language mixing. To address these issues and further enhance reasoning performance, we introduce DeepSeek-R1, which incorporates multi-stage training and cold-start data before RL. DeepSeek-R1 achieves performance comparable to OpenAI-o1-1217 on reasoning tasks. To support the research community, we open-source DeepSeek-R1-Zero, DeepSeek-R1, and six dense models (1.5B, 7B, 8B, 14B, 32B, 70B) distilled from DeepSeek-R1 based on Qwen and Llama.

Precise alignment of the electron beam is critical for successful application of scanning transmission electron microscopes (STEM) to understanding materials at atomic level. Despite the success of aberration correctors, aberration correction is still a complex process. Here we approach aberration correction from the perspective of accelerator physics and show it is equivalent to minimizing the emittance growth of the beam, the span of the phase space distribution of the probe. We train a deep learning model to predict emittance growth from experimentally accessible Ronchigrams. Both simulation and experimental results show the model can capture the emittance variation with aberration coefficients accurately. We further demonstrate the model can act as a fast-executing function for the global optimization of the lens parameters. Our approach enables new ways to quickly quantify and automate aberration correction that takes advantage of the rapid measurements possible with high-speed electron cameras. In part II of the paper, we demonstrate how the emittance metric enables rapid online tuning of the aberration corrector using Bayesian optimization., Comment: 27 pages, 6 figures

Aberration-corrected Scanning Transmission Electron Microscopy (STEM) has become an essential tool in understanding materials at the atomic scale. However, tuning the aberration corrector to produce a sub-{\AA}ngstr\"om probe is a complex and time-costly procedure, largely due to the difficulty of precisely measuring the optical state of the system. When measurements are both costly and noisy, Bayesian methods provide rapid and efficient optimization. To this end, we develop a Bayesian approach to fully automate the process by minimizing a new quality metric, beam emittance, which is shown to be equivalent to performing aberration correction. In part I, we derived several important properties of the beam emittance metric and trained a deep neural network to predict beam emittance growth from a single Ronchigram. Here we use this as the black box function for Bayesian Optimization and demonstrate automated tuning of simulated and real electron microscopes. We explore different surrogate functions for the Bayesian optimizer and implement a deep neural network kernel to effectively learn the interactions between different control channels without the need to explicitly measure a full set of aberration coefficients. Both simulation and experimental results show the proposed method outperforms conventional approaches by achieving a better optical state with a higher convergence rate., Comment: 24 pages, 12 figures

We present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks. The model checkpoints are available at https://github.com/deepseek-ai/DeepSeek-V3.

The rapid advance of Large Language Models (LLMs) has catalyzed the development of Vision-Language Models (VLMs). Monolithic VLMs, which avoid modality-specific encoders, offer a promising alternative to the compositional ones but face the challenge of inferior performance. Most existing monolithic VLMs require tuning pre-trained LLMs to acquire vision abilities, which may degrade their language capabilities. To address this dilemma, this paper presents a novel high-performance monolithic VLM named HoVLE. We note that LLMs have been shown capable of interpreting images, when image embeddings are aligned with text embeddings. The challenge for current monolithic VLMs actually lies in the lack of a holistic embedding module for both vision and language inputs. Therefore, HoVLE introduces a holistic embedding module that converts visual and textual inputs into a shared space, allowing LLMs to process images in the same way as texts. Furthermore, a multi-stage training strategy is carefully designed to empower the holistic embedding module. It is first trained to distill visual features from a pre-trained vision encoder and text embeddings from the LLM, enabling large-scale training with unpaired random images and text tokens. The whole model further undergoes next-token prediction on multi-modal data to align the embeddings. Finally, an instruction-tuning stage is incorporated. Our experiments show that HoVLE achieves performance close to leading compositional models on various benchmarks, outperforming previous monolithic models by a large margin. Model available at https://huggingface.co/OpenGVLab/HoVLE.

Accurate 3D mapping in endoscopy enables quantitative, holistic lesion characterization within the gastrointestinal (GI) tract, requiring reliable depth and pose estimation. However, endoscopy systems are monocular, and existing methods relying on synthetic datasets or complex models often lack generalizability in challenging endoscopic conditions. We propose a robust self-supervised monocular depth and pose estimation framework that incorporates a Generative Latent Bank and a Variational Autoencoder (VAE). The Generative Latent Bank leverages extensive depth scenes from natural images to condition the depth network, enhancing realism and robustness of depth predictions through latent feature priors. For pose estimation, we reformulate it within a VAE framework, treating pose transitions as latent variables to regularize scale, stabilize z-axis prominence, and improve x-y sensitivity. This dual refinement pipeline enables accurate depth and pose predictions, effectively addressing the GI tract's complex textures and lighting. Extensive evaluations on SimCol and EndoSLAM datasets confirm our framework's superior performance over published self-supervised methods in endoscopic depth and pose estimation.

The hybrid skin-topological effect (HSTE) has recently been proposed as a mechanism where topological edge states collapse into corner states under the influence of the non-Hermitian skin effect (NHSE). However, directly observing this effect is challenging due to the complex frequencies of eigenmodes. In this study, we experimentally observe HSTE corner states using synthetic complex frequency excitations in a transmission line network. We demonstrate that HSTE induces asymmetric transmission along a specific direction within the topological band gap. Besides HSTE, we identify corner states originating from non-chiral edge states, which are caused by the unbalanced effective onsite energy shifts at the boundaries of the network. Furthermore, our results suggest that whether the bulk interior is Hermitian or non-Hermitian is not a key factor for HSTE. Instead, the HSTE states can be realized and relocated simply by adjusting the non-Hermitian distribution at the boundaries. Our research has deepened the understanding of a range of issues regarding HSTE, paving the way for advancements in the design of non-Hermitian topological devices.

End-to-end style autonomous driving models have been developed recently. These models lack interpretability of decision-making process from perception to control of the ego vehicle, resulting in anxiety for passengers. To alleviate it, it is effective to build a model which outputs captions describing future behaviors of the ego vehicle and their reason. However, the existing approaches generate reasoning text that inadequately reflects the future plans of the ego vehicle, because they train models to output captions using momentary control signals as inputs. In this study, we propose a reasoning model that takes future planning trajectories of the ego vehicle as inputs to solve this limitation with the dataset newly collected., Comment: Accepted and presented at ECCV 2024 2nd Workshop on Vision-Centric Autonomous Driving (VCAD) on September 30, 2024. 13 pages, 5 figures

Large language models (LLMs) show impressive performance in solving complex language tasks. However, its large number of parameters present significant challenges for the deployment and application of the model on edge devices. Compressing large language models to low bits can enable them to run on resource-constrained devices, often leading to performance degradation. To address this problem, we propose gradient-aware weight quantization (GWQ), the first quantization approach for low-bit weight quantization that leverages gradients to localize outliers, requiring only a minimal amount of calibration data for outlier detection. GWQ retains the weights corresponding to the top 1% outliers preferentially at FP16 precision, while the remaining non-outlier weights are stored in a low-bit format. GWQ found experimentally that utilizing the sensitive weights in the gradient localization model is more scientific compared to utilizing the sensitive weights in the Hessian matrix localization model. Compared to current quantization methods, GWQ can be applied to multiple language models and achieves lower PPL on the WikiText2 and C4 dataset. In the zero-shot task, GWQ quantized models have higher accuracy compared to other quantization methods. GWQ is also suitable for multimodal model quantization, and the quantized Qwen-VL family model is more accurate than other methods. Zero-shot target detection task dataset RefCOCO outperforms the current stat-of-the-arts method SPQR. GWQ achieves 1.2 times inference speedup in comparison to the original model, and effectively reduces the inference memory.

Whole heart segmentation (WHS) supports cardiovascular disease (CVD) diagnosis, disease monitoring, treatment planning, and prognosis. Deep learning has become the most widely used method for WHS applications in recent years. However, segmentation of whole-heart structures faces numerous challenges including heart shape variability during the cardiac cycle, clinical artifacts like motion and poor contrast-to-noise ratio, domain shifts in multi-center data, and the distinct modalities of CT and MRI. To address these limitations and improve segmentation quality, this paper introduces a new topology-preserving module that is integrated into deep neural networks. The implementation achieves anatomically plausible segmentation by using learned topology-preserving fields, which are based entirely on 3D convolution and are therefore very effective for 3D voxel data. We incorporate natural constraints between structures into the end-to-end training and enrich the feature representation of the neural network. The effectiveness of the proposed method is validated on an open-source medical heart dataset, specifically using the WHS++ data. The results demonstrate that the architecture performs exceptionally well, achieving a Dice coefficient of 0.939 during testing. This indicates full topology preservation for individual structures and significantly outperforms other baselines in preserving the overall scene topology.

Vision Transformers with various attention modules have demonstrated superior performance on vision tasks. While using sparsity-adaptive attention, such as in DAT, has yielded strong results in image classification, the key-value pairs selected by deformable points lack semantic relevance when fine-tuning for semantic segmentation tasks. The query-aware sparsity attention in BiFormer seeks to focus each query on top-k routed regions. However, during attention calculation, the selected key-value pairs are influenced by too many irrelevant queries, reducing attention on the more important ones. To address these issues, we propose the Deformable Bi-level Routing Attention (DBRA) module, which optimizes the selection of key-value pairs using agent queries and enhances the interpretability of queries in attention maps. Based on this, we introduce the Deformable Bi-level Routing Attention Transformer (DeBiFormer), a novel general-purpose vision transformer built with the DBRA module. DeBiFormer has been validated on various computer vision tasks, including image classification, object detection, and semantic segmentation, providing strong evidence of its effectiveness.Code is available at {https://github.com/maclong01/DeBiFormer}, Comment: 20 pages, 7 figures. arXiv admin note: text overlap with arXiv:2303.08810 by other authors

Recently, NVS in human-object interaction scenes has received increasing attention. Existing human-object interaction datasets mainly consist of static data with limited views, offering only RGB images or videos, mostly containing interactions between a single person and objects. Moreover, these datasets exhibit complexities in lighting environments, poor synchronization, and low resolution, hindering high-quality human-object interaction studies. In this paper, we introduce a new people-object interaction dataset that comprises 38 series of 30-view multi-person or single-person RGB-D video sequences, accompanied by camera parameters, foreground masks, SMPL models, some point clouds, and mesh files. Video sequences are captured by 30 Kinect Azures, uniformly surrounding the scene, each in 4K resolution 25 FPS, and lasting for 1$\sim$19 seconds. Meanwhile, we evaluate some SOTA NVS models on our dataset to establish the NVS benchmarks. We hope our work can inspire further research in humanobject interaction.

Sodium niobate, a lead-free ferroic material, hosts delicately-balanced, competing order parameters, including ferroelectric states that can be stabilized by epitaxial strain. Here, we show that the resulting macroscopic ferroelectricity exhibits an unconventional microscopic structure using multislice electron ptychography. This technique overcomes multiple scattering artifacts limiting conventional electron microscopy, enabling both lateral spatial resolution beyond the diffraction limit and recovery of three-dimensional structural information. These imaging capabilities allow us to separate the ferroelectric interior of the sample from the relaxed surface structure and identify the soft phonon mode and related structural distortions with picometer precision. Unlike conventional ferroelectric perovskites, we find that the polar distortion in this material involves minimal distortions of the cation sublattices and is instead dominated by anion displacements relative to the niobium sublattice. We establish limits on film thickness for interfacial octahedral rotation engineering and directly visualize a random octahedral rotation pattern, arising from the flat dispersion of the associated phonon mode., Comment: 69 pages, 5 figures, 10 Extended Data figures, 10 Supplementary Figures

Transferable adversarial attacks pose significant threats to deep neural networks, particularly in black-box scenarios where internal model information is inaccessible. Studying adversarial attack methods helps advance the performance of defense mechanisms and explore model vulnerabilities. These methods can uncover and exploit weaknesses in models, promoting the development of more robust architectures. However, current methods for transferable attacks often come with substantial computational costs, limiting their deployment and application, especially in edge computing scenarios. Adversarial generative models, such as Generative Adversarial Networks (GANs), are characterized by their ability to generate samples without the need for retraining after an initial training phase. GE-AdvGAN, a recent method for transferable adversarial attacks, is based on this principle. In this paper, we propose a novel general framework for gradient editing-based transferable attacks, named GE-AdvGAN+, which integrates nearly all mainstream attack methods to enhance transferability while significantly reducing computational resource consumption. Our experiments demonstrate the compatibility and effectiveness of our framework. Compared to the baseline AdvGAN, our best-performing method, GE-AdvGAN++, achieves an average ASR improvement of 47.8. Additionally, it surpasses the latest competing algorithm, GE-AdvGAN, with an average ASR increase of 5.9. The framework also exhibits enhanced computational efficiency, achieving 2217.7 FPS, outperforming traditional methods such as BIM and MI-FGSM. The implementation code for our GE-AdvGAN+ framework is available at https://github.com/GEAdvGANP

Mean field games (MFGs) model interactions in large-population multi-agent systems through population distributions. Traditional learning methods for MFGs are based on fixed-point iteration (FPI), where policy updates and induced population distributions are computed separately and sequentially. However, FPI-type methods may suffer from inefficiency and instability due to potential oscillations caused by this forward-backward procedure. In this work, we propose a novel perspective that treats the policy and population as a unified parameter controlling the game dynamics. By applying stochastic parameter approximation to this unified parameter, we develop SemiSGD, a simple stochastic gradient descent (SGD)-type method, where an agent updates its policy and population estimates simultaneously and fully asynchronously. Building on this perspective, we further apply linear function approximation (LFA) to the unified parameter, resulting in the first population-aware LFA (PA-LFA) for learning MFGs on continuous state-action spaces. A comprehensive finite-time convergence analysis is provided for SemiSGD with PA-LFA, including its convergence to the equilibrium for linear MFGs -- a class of MFGs with a linear structure concerning the population -- under the standard contractivity condition, and to a neighborhood of the equilibrium under a more practical condition. We also characterize the approximation error for non-linear MFGs. We validate our theoretical findings with six experiments on three MFGs., Comment: Published as a conference paper at ICLR 2025

In recent times, the swift evolution of adversarial attacks has captured widespread attention, particularly concerning their transferability and other performance attributes. These techniques are primarily executed at the sample level, frequently overlooking the intrinsic parameters of models. Such neglect suggests that the perturbations introduced in adversarial samples might have the potential for further reduction. Given the essence of adversarial attacks is to impair model integrity with minimal noise on original samples, exploring avenues to maximize the utility of such perturbations is imperative. Against this backdrop, we have delved into the complexities of adversarial attack algorithms, dissecting the adversarial process into two critical phases: the Directional Supervision Process (DSP) and the Directional Optimization Process (DOP). While DSP determines the direction of updates based on the current samples and model parameters, it has been observed that existing model parameters may not always be conducive to adversarial attacks. The impact of models on adversarial efficacy is often overlooked in current research, leading to the neglect of DSP. We propose that under certain conditions, fine-tuning model parameters can significantly enhance the quality of DSP. For the first time, we propose that under certain conditions, fine-tuning model parameters can significantly improve the quality of the DSP. We provide, for the first time, rigorous mathematical definitions and proofs for these conditions, and introduce multiple methods for fine-tuning model parameters within DSP. Our extensive experiments substantiate the effectiveness of the proposed P3A method. Our code is accessible at: https://anonymous.4open.science/r/P3A-A12C/

Recently, the text-to-image diffusion model has gained considerable attention from the community due to its exceptional image generation capability. A representative model, Stable Diffusion, amassed more than 10 million users within just two months of its release. This surge in popularity has facilitated studies on the robustness and safety of the model, leading to the proposal of various adversarial attack methods. Simultaneously, there has been a marked increase in research focused on defense methods to improve the robustness and safety of these models. In this survey, we provide a comprehensive review of the literature on adversarial attacks and defenses targeting text-to-image diffusion models. We begin with an overview of text-to-image diffusion models, followed by an introduction to a taxonomy of adversarial attacks and an in-depth review of existing attack methods. We then present a detailed analysis of current defense methods that improve model robustness and safety. Finally, we discuss ongoing challenges and explore promising future research directions. For a complete list of the adversarial attack and defense methods covered in this survey, please refer to our curated repository at https://github.com/datar001/Awesome-AD-on-T2IDM., Comment: Accepted for Information Fusion. Related benchmarks and codes are available at \url{https://github.com/datar001/Awesome-AD-on-T2IDM}

Electromagnetic wave propagation in three-dimensional space typically suffers omnidirectional scattering when encountering obstacles. In this study, we employed Chern vectors to construct a topological heterostructure, where large-volume non-reciprocal topological transport in three-dimension is achieved. The shape of the cross-section in the heterostructure can be arbitrary designed, and we experimentally observed the distinctive cross-shaped field pattern transport, non-reciprocal energy harvesting, and most importantly, the remarkable ability of electromagnetic wave to traverse obstacles and abrupt structure changes without encountering reflections in 3D space., Comment: 17 pages, 3 figures

We propose a discrete time graphon game formulation on continuous state and action spaces using a representative player to study stochastic games with heterogeneous interaction among agents. This formulation admits both philosophical and mathematical advantages, compared to a widely adopted formulation using a continuum of players. We prove the existence and uniqueness of the graphon equilibrium with mild assumptions, and show that this equilibrium can be used to construct an approximate solution for finite player game on networks, which is challenging to analyze and solve due to curse of dimensionality. An online oracle-free learning algorithm is developed to solve the equilibrium numerically, and sample complexity analysis is provided for its convergence., Comment: Published as a conference paper at ICML 2024

Mean field games (MFGs) are a promising framework for modeling the behavior of large-population systems. However, solving MFGs can be challenging due to the coupling of forward population evolution and backward agent dynamics. Typically, obtaining mean field Nash equilibria (MFNE) involves an iterative approach where the forward and backward processes are solved alternately, known as fixed-point iteration (FPI). This method requires fully observed population propagation and agent dynamics over the entire spatial domain, which could be impractical in some real-world scenarios. To overcome this limitation, this paper introduces a novel online single-agent model-free learning scheme, which enables a single agent to learn MFNE using online samples, without prior knowledge of the state-action space, reward function, or transition dynamics. Specifically, the agent updates its policy through the value function (Q), while simultaneously evaluating the mean field state (M), using the same batch of observations. We develop two variants of this learning scheme: off-policy and on-policy QM iteration. We prove that they efficiently approximate FPI, and a sample complexity guarantee is provided. The efficacy of our methods is confirmed by numerical experiments., Comment: Published as a conference paper at ECAI 2024

With the development of mobile communication and industrial internet technologies, the demand for robust absolute time synchronization based on network for diverse scenarios is significantly growing. TAP is a novel network timing method that aims to achieve sub-microsecond synchronization over air interface. This paper investigates the improvement and end-to-end realization of TAP. This paper first analyzes the effectiveness and deficiencies of TAP by establishing an equivalent clock model which evaluates TAP from timing error composition and allan variance. Second, this paper proposes a detailed base station and terminal design and the corresponding improvement of TAP. Both hardware compensation and protocol software design are taken into account so as to minimize timing error and system cost while maximizing compatibility with 3GPP. Finally, this paper presents a TAP end-to-end 5G prototype system developed based on software defined radio base station and COTS baseband module. The field test results show that the proposed scheme effectively solves the problems of TAP in application and robustly achieves 200ns level timing accuracy in various situations. The average accuracy with long observations can reach 1 nanosecond. It is 2$\sim$3 orders of magnitude better than common network timing methods, including NTP, PTP and the original TAP.

Federated reinforcement learning (FRL) has emerged as a promising paradigm for reducing the sample complexity of reinforcement learning tasks by exploiting information from different agents. However, when each agent interacts with a potentially different environment, little to nothing is known theoretically about the non-asymptotic performance of FRL algorithms. The lack of such results can be attributed to various technical challenges and their intricate interplay: Markovian sampling, linear function approximation, multiple local updates to save communication, heterogeneity in the reward functions and transition kernels of the agents' MDPs, and continuous state-action spaces. Moreover, in the on-policy setting, the behavior policies vary with time, further complicating the analysis. In response, we introduce FedSARSA, a novel federated on-policy reinforcement learning scheme, equipped with linear function approximation, to address these challenges and provide a comprehensive finite-time error analysis. Notably, we establish that FedSARSA converges to a policy that is near-optimal for all agents, with the extent of near-optimality proportional to the level of heterogeneity. Furthermore, we prove that FedSARSA leverages agent collaboration to enable linear speedups as the number of agents increases, which holds for both fixed and adaptive step-size configurations., Comment: Published as a conference paper at ICLR 2024

Two-dimensional (2D) magnets exhibit unique physical properties for potential applications in spintronics. To date, most 2D ferromagnets are obtained by mechanical exfoliation of bulk materials with van der Waals interlayer interactions, and the synthesis of single or few-layer 2D ferromagnets with strong interlayer coupling remains experimentally challenging. Here, we report the epitaxial growth of 2D non-van der Waals ferromagnetic bilayer FeSb on SrTiO3(001) substrates stabilized by strong coupling to the substrate, which exhibits in-plane magnetic anisotropy and a Curie temperature above 300 K. In-situ low-temperature scanning tunneling microscopy/spectroscopy and density-functional theory calculations further reveal that a Fe Kagome layer terminates the bilayer FeSb. Our results open a new avenue for further exploring emergent quantum phenomena from the interplay of ferromagnetism and topology for application in spintronics.

All-around, real-time navigation and sensing across the water environments by miniature soft robotics are promising, for their merits of small size, high agility and good compliance to the unstructured surroundings. In this paper, we propose and demonstrate a mantas-like soft aquatic robot which propels itself by flapping-fins using rolled dielectric elastomer actuators (DEAs) with bending motions. This robot exhibits fast-moving capabilities of swimming at 57mm/s or 1.25 body length per second (BL/s), skating on water surface at 64 mm/s (1.36 BL/s) and vertical ascending at 38mm/s (0.82 BL/s) at 1300 V, 17 Hz of the power supply. These results show the feasibility of adopting rolled DEAs for mesoscale aquatic robots with high motion performance in various water-related scenarios., Comment: 6 Pages, 12 Figures, Published at IROS 2023

This paper presents strong worst-case iteration and operation complexity guarantees for Riemannian adaptive regularized Newton methods, a unified framework encompassing both Riemannian adaptive regularization (RAR) methods and Riemannian trust region (RTR) methods. We comprehensively characterize the sources of approximation in second-order manifold optimization methods: the objective function's smoothness, retraction's smoothness, and subproblem solver's inexactness. Specifically, for a function with a $\mu$-H\"older continuous Hessian, when equipped with a retraction featuring a $\nu$-H\"older continuous differential and a $\theta$-inexact subproblem solver, both RTR and RAR with $2+\alpha$ regularization (where $\alpha=\min\{\mu,\nu,\theta\}$) locate an $(\epsilon,\epsilon^{\alpha/(1+\alpha)})$-approximate second-order stationary point within at most $O(\epsilon^{-(2+\alpha)/(1+\alpha)})$ iterations and at most $\tilde{O}(\epsilon^{-(4+3\alpha)/(2(1+\alpha))})$ Hessian-vector products. These complexity results are novel and sharp, and reduce to an iteration complexity of $O(\epsilon^{-3/2})$ and an operation complexity of $\tilde{O}(\epsilon^{-7/4})$ when $\alpha=1$.

Efficiently monitoring the condition of civil infrastructure requires automating the structural condition assessment in visual inspection. This paper proposes an Attention-Enhanced Co-Interactive Fusion Network (AECIF-Net) for automatic structural condition assessment in visual bridge inspection. AECIF-Net can simultaneously parse structural elements and segment surface defects on the elements in inspection images. It integrates two task-specific relearning subnets to extract task-specific features from an overall feature embedding. A co-interactive feature fusion module further captures the spatial correlation and facilitates information sharing between tasks. Experimental results demonstrate that the proposed AECIF-Net outperforms the current state-of-the-art approaches, achieving promising performance with 92.11% mIoU for element segmentation and 87.16% mIoU for corrosion segmentation on the test set of the new benchmark dataset Steel Bridge Condition Inspection Visual (SBCIV). An ablation study verifies the merits of the designs for AECIF-Net, and a case study demonstrates its capability to automate structural condition assessment., Comment: Submitted to Automation in Construction