Search

Despite their remarkable capabilities, Large Language Models (LLMs) are prone to generate responses that contradict verifiable facts, i.e., unfaithful hallucination content. Existing efforts generally focus on optimizing model parameters or editing semantic representations, which compromise the internal factual knowledge of target LLMs. In addition, hallucinations typically exhibit multifaceted patterns in downstream tasks, limiting the model's holistic performance across tasks. In this paper, we propose a Comparator-driven Decoding-Time (CDT) framework to alleviate the response hallucination. Firstly, we construct hallucinatory and truthful comparators with multi-task fine-tuning samples. In this case, we present an instruction prototype-guided mixture of experts strategy to enhance the ability of the corresponding comparators to capture different hallucination or truthfulness patterns in distinct task instructions. CDT constrains next-token predictions to factuality-robust distributions by contrasting the logit differences between the target LLMs and these comparators. Systematic experiments on multiple downstream tasks show that our framework can significantly improve the model performance and response factuality., Comment: Hallucination Mitigation in LLMs

Backdoor attacks pose a significant threat to the training process of deep neural networks (DNNs). As a widely-used DNN-based application in real-world scenarios, face recognition systems once implanted into the backdoor, may cause serious consequences. Backdoor research on face recognition is still in its early stages, and the existing backdoor triggers are relatively simple and visible. Furthermore, due to the perceptibility, diversity, and similarity of facial datasets, many state-of-the-art backdoor attacks lose effectiveness on face recognition tasks. In this work, we propose a novel feature space backdoor attack against face recognition via makeup transfer, dubbed MakeupAttack. In contrast to many feature space attacks that demand full access to target models, our method only requires model queries, adhering to black-box attack principles. In our attack, we design an iterative training paradigm to learn the subtle features of the proposed makeup-style trigger. Additionally, MakeupAttack promotes trigger diversity using the adaptive selection method, dispersing the feature distribution of malicious samples to bypass existing defense methods. Extensive experiments were conducted on two widely-used facial datasets targeting multiple models. The results demonstrate that our proposed attack method can bypass existing state-of-the-art defenses while maintaining effectiveness, robustness, naturalness, and stealthiness, without compromising model performance.

Multiple instance learning (MIL) has become a standard paradigm for weakly supervised classification of whole slide images (WSI). However, this paradigm relies on the use of a large number of labelled WSIs for training. The lack of training data and the presence of rare diseases present significant challenges for these methods. Prompt tuning combined with the pre-trained Vision-Language models (VLMs) is an effective solution to the Few-shot Weakly Supervised WSI classification (FSWC) tasks. Nevertheless, applying prompt tuning methods designed for natural images to WSIs presents three significant challenges: 1) These methods fail to fully leverage the prior knowledge from the VLM's text modality; 2) They overlook the essential multi-scale and contextual information in WSIs, leading to suboptimal results; and 3) They lack exploration of instance aggregation methods. To address these problems, we propose a Multi-Scale and Context-focused Prompt Tuning (MSCPT) method for FSWC tasks. Specifically, MSCPT employs the frozen large language model to generate pathological visual language prior knowledge at multi-scale, guiding hierarchical prompt tuning. Additionally, we design a graph prompt tuning module to learn essential contextual information within WSI, and finally, a non-parametric cross-guided instance aggregation module has been introduced to get the WSI-level features. Based on two VLMs, extensive experiments and visualizations on three datasets demonstrated the powerful performance of our MSCPT., Comment: 11 pages, 5 figures, 5tables

Accurate and robust multimodal multi-task perception is crucial for modern autonomous driving systems. However, current multimodal perception research follows independent paradigms designed for specific perception tasks, leading to a lack of complementary learning among tasks and decreased performance in multi-task learning (MTL) due to joint training. In this paper, we propose MaskBEV, a masked attention-based MTL paradigm that unifies 3D object detection and bird's eye view (BEV) map segmentation. MaskBEV introduces a task-agnostic Transformer decoder to process these diverse tasks, enabling MTL to be completed in a unified decoder without requiring additional design of specific task heads. To fully exploit the complementary information between BEV map segmentation and 3D object detection tasks in BEV space, we propose spatial modulation and scene-level context aggregation strategies. These strategies consider the inherent dependencies between BEV segmentation and 3D detection, naturally boosting MTL performance. Extensive experiments on nuScenes dataset show that compared with previous state-of-the-art MTL methods, MaskBEV achieves 1.3 NDS improvement in 3D object detection and 2.7 mIoU improvement in BEV map segmentation, while also demonstrating slightly leading inference speed., Comment: Accepted to ACM MM 2024

Vision-based 3D semantic scene completion (SSC) describes autonomous driving scenes through 3D volume representations. However, the occlusion of invisible voxels by scene surfaces poses challenges to current SSC methods in hallucinating refined 3D geometry. This paper proposes HybridOcc, a hybrid 3D volume query proposal method generated by Transformer framework and NeRF representation and refined in a coarse-to-fine SSC prediction framework. HybridOcc aggregates contextual features through the Transformer paradigm based on hybrid query proposals while combining it with NeRF representation to obtain depth supervision. The Transformer branch contains multiple scales and uses spatial cross-attention for 2D to 3D transformation. The newly designed NeRF branch implicitly infers scene occupancy through volume rendering, including visible and invisible voxels, and explicitly captures scene depth rather than generating RGB color. Furthermore, we present an innovative occupancy-aware ray sampling method to orient the SSC task instead of focusing on the scene surface, further improving the overall performance. Extensive experiments on nuScenes and SemanticKITTI datasets demonstrate the effectiveness of our HybridOcc on the SSC task., Comment: Accepted to IEEE RAL

Previous works have shown that reducing parameter overhead and computations for transformer-based single image super-resolution (SISR) models (e.g., SwinIR) usually leads to a reduction of performance. In this paper, we present GRFormer, an efficient and lightweight method, which not only reduces the parameter overhead and computations, but also greatly improves performance. The core of GRFormer is Grouped Residual Self-Attention (GRSA), which is specifically oriented towards two fundamental components. Firstly, it introduces a novel grouped residual layer (GRL) to replace the Query, Key, Value (QKV) linear layer in self-attention, aimed at efficiently reducing parameter overhead, computations, and performance loss at the same time. Secondly, it integrates a compact Exponential-Space Relative Position Bias (ES-RPB) as a substitute for the original relative position bias to improve the ability to represent position information while further minimizing the parameter count. Extensive experimental results demonstrate that GRFormer outperforms state-of-the-art transformer-based methods for $\times$2, $\times$3 and $\times$4 SISR tasks, notably outperforming SOTA by a maximum PSNR of 0.23dB when trained on the DIV2K dataset, while reducing the number of parameter and MACs by about \textbf{60\%} and \textbf{49\% } in only self-attention module respectively. We hope that our simple and effective method that can easily applied to SR models based on window-division self-attention can serve as a useful tool for further research in image super-resolution. The code is available at \url{https://github.com/sisrformer/GRFormer}., Comment: Accepted for ACM MM 2024

We introduce a generic estimator for the false discovery rate of any model selection procedure, in common statistical modeling settings including the Gaussian linear model, Gaussian graphical model, and model-X setting. We prove that our method has a conservative (non-negative) bias in finite samples under standard statistical assumptions, and provide a bootstrap method for assessing its standard error. For methods like the Lasso, forward-stepwise regression, and the graphical Lasso, our estimator serves as a valuable companion to cross-validation, illuminating the tradeoff between prediction error and variable selection accuracy as a function of the model complexity parameter.

Let $T$ be a tournament with $n$ vertices $v_1,\ldots,v_n$. The skew-adjacency matrix of $T$ is the $n\times n$ zero-diagonal matrix $S_T = [s_{ij}]$ in which $s_{ij}=-s_{ji}=1$ if $ v_i $ dominates $ v_j $. We define the determinant $\det(T)$ of $ T $ as the determinant of $ S_T $. It is well-known that $\det(T)=0$ if $n$ is odd and $\det(T)$ is the square of an odd integer if $n$ is even. Let $\mathcal{D}_k$ be the set of tournaments whose all subtournaments have determinant at most $ k^{2} $, where $k$ is a positive odd integer. The necessary and sufficient condition for $T\in \mathcal{D}_1$ or $T\in \mathcal{D}_3$ has been characterized in $2023$. In this paper, we characterize the set $\mathcal{D}_5$, obtain some properties of $\mathcal{D}_k$. Moreover, for any positive odd integer $k$, we give a construction of a tournament $T$ satisfying that $\det(T)=k^2$, and $T\in \mathcal{D}_k\backslash\mathcal{D}_{k-2}$ if $k\geq 3$, which implies $\mathcal{D}_k\backslash\mathcal{D}_{k-2}$ is not an empty set for $k\geq 3$., Comment: 28 pages, 1 figure

In this paper, we present an efficient visual SLAM system designed to tackle both short-term and long-term illumination challenges. Our system adopts a hybrid approach that combines deep learning techniques for feature detection and matching with traditional backend optimization methods. Specifically, we propose a unified convolutional neural network (CNN) that simultaneously extracts keypoints and structural lines. These features are then associated, matched, triangulated, and optimized in a coupled manner. Additionally, we introduce a lightweight relocalization pipeline that reuses the built map, where keypoints, lines, and a structure graph are used to match the query frame with the map. To enhance the applicability of the proposed system to real-world robots, we deploy and accelerate the feature detection and matching networks using C++ and NVIDIA TensorRT. Extensive experiments conducted on various datasets demonstrate that our system outperforms other state-of-the-art visual SLAM systems in illumination-challenging environments. Efficiency evaluations show that our system can run at a rate of 73Hz on a PC and 40Hz on an embedded platform., Comment: 19 pages, 14 figures

Temporal Action Segmentation (TAS) is an essential task in video analysis, aiming to segment and classify continuous frames into distinct action segments. However, the ambiguous boundaries between actions pose a significant challenge for high-precision segmentation. Recent advances in diffusion models have demonstrated substantial success in TAS tasks due to their stable training process and high-quality generation capabilities. However, the heavy sampling steps required by diffusion models pose a substantial computational burden, limiting their practicality in real-time applications. Additionally, most related works utilize Transformer-based encoder architectures. Although these architectures excel at capturing long-range dependencies, they incur high computational costs and face feature-smoothing issues when processing long video sequences. To address these challenges, we propose EffiDiffAct, an efficient and high-performance TAS algorithm. Specifically, we develop a lightweight temporal feature encoder that reduces computational overhead and mitigates the rank collapse phenomenon associated with traditional self-attention mechanisms. Furthermore, we introduce an adaptive skip strategy that allows for dynamic adjustment of timestep lengths based on computed similarity metrics during inference, thereby further enhancing computational efficiency. Comprehensive experiments on the 50Salads, Breakfast, and GTEA datasets demonstrated the effectiveness of the proposed algorithm., Comment: 25 pages, 6 figures

A fundamental result in extremal graph theory attributes to Mantel's theorem, which states that every graph on $n$ vertices with more than $\lfloor n^2/4 \rfloor$ edges contains a triangle. About half of a century ago, Lov\'{a}sz and Simonovits (1975) provided a supersaturation phenomenon, which asserts that for $q< n/2$, every graph with $\lfloor n^2/4 \rfloor +q$ edges contains at least $q\lfloor n/2 \rfloor$ triangles. This result solved a conjecture proposed by Erd\H{o}s in 1962. In this paper, we establish a spectral version of the result of Lov\'{a}sz and Simonovits. Let $Y_{n,2,q}$ be the graph obtained from the bipartite Tur\'{a}n graph $T_{n,2}$ by embedding a matching with $q$ edges into the vertex part of size $\lceil n/2\rceil$. Using the supersaturation-stability method and the classical spectral techniques, we firstly prove that for $n\ge 300q^2$, each graph $G$ on $n$ vertices with $\lambda (G) \ge \lambda (Y_{n,2,q})$ contains at least $q\lfloor n/2 \rfloor$ triangles. Moreover, let $T_{n,2,q}$ be the graph obtained from $T_{n,2}$ by embedding a star with $q$ edges into the vertex part of size $\lceil n/2\rceil$. Secondly, we show further that $T_{n,2,q}$ is the unique spectral extremal graph that contains at most $q\lfloor n/2 \rfloor$ triangles and attains the maximum of the spectral radius. This result answers a spectral triangle counting problem due to Ning and Zhai (2023). Thirdly, we present an asymptotically spectral stability result under a specific constraint on the triangle covering number. The third result could be regarded as a spectral extension of a recent result proved by Balogh and Clemen (2023), and independently by Liu and Mubayi (2022)., Comment: 28 pages, 1 figure. We proposed some spectral extremal graph problems for readers. Any suggestions are welcome

For most existing prescribed performance formation control methods, performance requirements are not directly imposed on the relative states between agents but on the consensus error, which lacks a clear physical interpretation of their solution. In this paper, we propose a novel adaptive prescribed performance formation control strategy, capable of guaranteeing prescribed performance on the relative errors, for uncertain high-order multi-agent systems under a class of directed graphs. Due to the consideration of performance constraints for relative errors, a coupled nonlinear interaction term that contains global graphic information among agents is involved in the error dynamics, leading to a fully distributed control design more difficult and challenging. Here by proposing a series of nonlinear mappings and utilizing the edge Laplacian along with Lyapunov stability theory, the presented formation control scheme exhibits the following appealing features when compared to existing results: 1) different performance requirements can be guaranteed in a unified way by solely tuning the design parameters a priori, without the need for control redesign and stability reanalysis under the proposed fixed control protocol, making the design more user-friendly and the implementation less demanding; 2) the complex and burdensome verification process for the initial constraint, often encountered in existing prescribed performance controls, is completely obviated if the performance requirements are global; and 3) nonlinear interaction is completely decoupled and the asymptotic stability of the formation manifold is ensured via using the adaptive parameter estimate technique. Finally, simulations of various performance behaviors are performed to show the efficiency of the theoretical results., Comment: 9 pages, 3 figures, submitted to IEEE

A differential dynamic programming (DDP)-based framework for inverse reinforcement learning (IRL) is introduced to recover the parameters in the cost function, system dynamics, and constraints from demonstrations. Different from existing work, where DDP was used for the inner forward problem with inequality constraints, our proposed framework uses it for efficient computation of the gradient required in the outer inverse problem with equality and inequality constraints. The equivalence between the proposed method and existing methods based on Pontryagin's Maximum Principle (PMP) is established. More importantly, using this DDP-based IRL with an open-loop loss function, a closed-loop IRL framework is presented. In this framework, a loss function is proposed to capture the closed-loop nature of demonstrations. It is shown to be better than the commonly used open-loop loss function. We show that the closed-loop IRL framework reduces to a constrained inverse optimal control problem under certain assumptions. Under these assumptions and a rank condition, it is proven that the learning parameters can be recovered from the demonstration data. The proposed framework is extensively evaluated through four numerical robot examples and one real-world quadrotor system. The experiments validate the theoretical results and illustrate the practical relevance of the approach., Comment: 20 pages, 15 figures; submitted to IEEE for potential publication

Semantic pattern of an object point cloud is determined by its topological configuration of local geometries. Learning discriminative representations can be challenging due to large shape variations of point sets in local regions and incomplete surface in a global perspective, which can be made even more severe in the context of unsupervised domain adaptation (UDA). In specific, traditional 3D networks mainly focus on local geometric details and ignore the topological structure between local geometries, which greatly limits their cross-domain generalization. Recently, the transformer-based models have achieved impressive performance gain in a range of image-based tasks, benefiting from its strong generalization capability and scalability stemming from capturing long range correlation across local patches. Inspired by such successes of visual transformers, we propose a novel Relational Priors Distillation (RPD) method to extract relational priors from the well-trained transformers on massive images, which can significantly empower cross-domain representations with consistent topological priors of objects. To this end, we establish a parameter-frozen pre-trained transformer module shared between 2D teacher and 3D student models, complemented by an online knowledge distillation strategy for semantically regularizing the 3D student model. Furthermore, we introduce a novel self-supervised task centered on reconstructing masked point cloud patches using corresponding masked multi-view image features, thereby empowering the model with incorporating 3D geometric information. Experiments on the PointDA-10 and the Sim-to-Real datasets verify that the proposed method consistently achieves the state-of-the-art performance of UDA for point cloud classification. The source code of this work is available at https://github.com/zou-longkun/RPD.git.

This paper proposes a new class of distributional causal quantities, referred to as the \textit{outcome conditioned partial policy effects} (OCPPEs), to measure the \textit{average} effect of a general counterfactual intervention of a target covariate on the individuals in different quantile ranges of the outcome distribution. The OCPPE approach is valuable in several aspects: (i) Unlike the unconditional quantile partial effect (UQPE) that is not $\sqrt{n}$-estimable, an OCPPE is $\sqrt{n}$-estimable. Analysts can use it to capture heterogeneity across the unconditional distribution of $Y$ as well as obtain accurate estimation of the aggregated effect at the upper and lower tails of $Y$. (ii) The semiparametric efficiency bound for an OCPPE is explicitly derived. (iii) We propose an efficient debiased estimator for OCPPE, and provide feasible uniform inference procedures for the OCPPE process. (iv) The efficient doubly robust score for an OCPPE can be used to optimize infinitesimal nudges to a continuous treatment by maximizing a quantile specific Empirical Welfare function. We illustrate the method by analyzing how anti-smoking policies impact low percentiles of live infants' birthweights.

High-precision control for nonlinear systems is impeded by the low-fidelity dynamical model and external disturbance. Especially, the intricate coupling between internal uncertainty and external disturbance is usually difficult to be modeled explicitly. Here we show an effective and convergent algorithm enabling accurate estimation of the coupled disturbance via combining control and learning philosophies. Specifically, by resorting to Chebyshev series expansion, the coupled disturbance is firstly decomposed into an unknown parameter matrix and two known structures depending on system state and external disturbance respectively. A Regularized Least Squares (RLS) algorithm is subsequently formalized to learn the parameter matrix by using historical time-series data. Finally, a higher-order disturbance observer (HODO) is developed to achieve a high-precision estimation of the coupled disturbance by utilizing the learned portion. The efficiency of the proposed algorithm is evaluated through extensive simulations. We believe this work can offer a new option to merge learning schemes into the control framework for addressing existing intractable control problems., Comment: 8 pages, 3 figures

Context-aware emotion recognition (CAER) is a complex and significant task that requires perceiving emotions from various contextual cues. Previous approaches primarily focus on designing sophisticated architectures to extract emotional cues from images. However, their knowledge is confined to specific training datasets and may reflect the subjective emotional biases of the annotators. Furthermore, acquiring large amounts of labeled data is often challenging in real-world applications. In this paper, we systematically explore the potential of leveraging Large Vision-Language Models (LVLMs) to empower the CAER task from three paradigms: 1) We fine-tune LVLMs on two CAER datasets, which is the most common way to transfer large models to downstream tasks. 2) We design zero-shot and few-shot patterns to evaluate the performance of LVLMs in scenarios with limited data or even completely unseen. In this case, a training-free framework is proposed to fully exploit the In-Context Learning (ICL) capabilities of LVLMs. Specifically, we develop an image similarity-based ranking algorithm to retrieve examples; subsequently, the instructions, retrieved examples, and the test example are combined to feed LVLMs to obtain the corresponding sentiment judgment. 3) To leverage the rich knowledge base of LVLMs, we incorporate Chain-of-Thought (CoT) into our framework to enhance the model's reasoning ability and provide interpretable results. Extensive experiments and analyses demonstrate that LVLMs achieve competitive performance in the CAER task across different paradigms. Notably, the superior performance in few-shot settings indicates the feasibility of LVLMs for accomplishing specific tasks without extensive training.

There is a significant demand for indoor localization technology in smart buildings, and the most promising solution in this field is using RF sensors and fingerprinting-based methods that employ machine learning models trained on crowd-sourced user data gathered from IoT devices. However, this raises security and privacy issues in practice. Some researchers propose to use federated learning to partially overcome privacy problems, but there still remain security concerns, e.g., single-point failure and malicious attacks. In this paper, we propose a framework named DFLoc to achieve precise 3D localization tasks while considering the following two security concerns. Particularly, we design a specialized blockchain to decentralize the framework by distributing the tasks such as model distribution and aggregation which are handled by a central server to all clients in most previous works, to address the issue of the single-point failure for a reliable and accurate indoor localization system. Moreover, we introduce an updated model verification mechanism within the blockchain to alleviate the concern of malicious node attacks. Experimental results substantiate the framework's capacity to deliver accurate 3D location predictions and its superior resistance to the impacts of single-point failure and malicious attacks when compared to conventional centralized federated learning systems.

Understanding emotions from diverse contexts has received widespread attention in computer vision communities. The core philosophy of Context-Aware Emotion Recognition (CAER) is to provide valuable semantic cues for recognizing the emotions of target persons by leveraging rich contextual information. Current approaches invariably focus on designing sophisticated structures to extract perceptually critical representations from contexts. Nevertheless, a long-neglected dilemma is that a severe context bias in existing datasets results in an unbalanced distribution of emotional states among different contexts, causing biased visual representation learning. From a causal demystification perspective, the harmful bias is identified as a confounder that misleads existing models to learn spurious correlations based on likelihood estimation, limiting the models' performance. To address the issue, we embrace causal inference to disentangle the models from the impact of such bias, and formulate the causalities among variables in the CAER task via a customized causal graph. Subsequently, we present a Contextual Causal Intervention Module (CCIM) to de-confound the confounder, which is built upon backdoor adjustment theory to facilitate seeking approximate causal effects during model training. As a plug-and-play component, CCIM can easily integrate with existing approaches and bring significant improvements. Systematic experiments on three datasets demonstrate the effectiveness of our CCIM., Comment: TPAMI 2024

Understanding human intentions (e.g., emotions) from videos has received considerable attention recently. Video streams generally constitute a blend of temporal data stemming from distinct modalities, including natural language, facial expressions, and auditory clues. Despite the impressive advancements of previous works via attention-based paradigms, the inherent temporal asynchrony and modality heterogeneity challenges remain in multimodal sequence fusion, causing adverse performance bottlenecks. To tackle these issues, we propose a Multimodal fusion approach for learning modality-Exclusive and modality-Agnostic representations (MEA) to refine multimodal features and leverage the complementarity across distinct modalities. On the one hand, MEA introduces a predictive self-attention module to capture reliable context dynamics within modalities and reinforce unique features over the modality-exclusive spaces. On the other hand, a hierarchical cross-modal attention module is designed to explore valuable element correlations among modalities over the modality-agnostic space. Meanwhile, a double-discriminator strategy is presented to ensure the production of distinct representations in an adversarial manner. Eventually, we propose a decoupled graph fusion mechanism to enhance knowledge exchange across heterogeneous modalities and learn robust multimodal representations for downstream tasks. Numerous experiments are implemented on three multimodal datasets with asynchronous sequences. Systematic analyses show the necessity of our approach., Comment: TCSVT 2024

Recently, Ning and Zhai (2023) proved that every $n$-vertex graph $G$ with $\lambda (G) \ge \sqrt{\lfloor n^2/4\rfloor}$ has at least $\lfloor n/2\rfloor -1$ triangles, unless $G=K_{\lceil \frac{n}{2} \rceil, \lfloor \frac{n}{2} \rfloor}$. The aim of this paper is two-fold. Using the supersaturation-stability method, we prove a stability variant of Ning-Zhai's result by showing that such a graph $G$ contains at least $n-3$ triangles if no vertex is in all triangles of $G$. This result could also be viewed as a spectral version of a result of Xiao and Katona (2021). The second part concerns with the spectral supersaturation for the bowtie, which consists of two triangles sharing a common vertex. A theorem of Erd\H{o}s, F\"{u}redi, Gould and Gunderson (1995) says that every $n$-vertex graph with more than $\lfloor n^2/4\rfloor +1$ edges contains a bowtie. For graphs of given order, the spectral supersaturation problem has not been considered for substructures that are not color-critical. In this paper, we give the first such theorem by counting the number of bowties. Let $K_{\lceil \frac{n}{2} \rceil, \lfloor \frac{n}{2} \rfloor}^{+2}$ be the graph obtained from $K_{\lceil \frac{n}{2} \rceil, \lfloor \frac{n}{2} \rfloor}$ by embedding two disjoint edges into the vertex part of size $\lceil \frac{n}{2} \rceil$. Our result shows that every graph $G$ with $n\ge 8.8 \times 10^6$ vertices and $\lambda (G)\ge \lambda (K_{\lceil \frac{n}{2} \rceil, \lfloor \frac{n}{2} \rfloor}^{+2})$ contains at least $\lfloor \frac{n}{2} \rfloor$ bowties, and $K_{\lceil \frac{n}{2} \rceil, \lfloor \frac{n}{2} \rfloor}^{+2}$ is the unique spectral extremal graph. This gives a spectral correspondence of a theorem of Kang, Makai and Pikhurko (2020). The method used in our paper provides a probable way to establish the spectral counting results for other graphs, even for non-color-critical graphs., Comment: 32 pages. Some spectral extremal graph problems are proposed in the end. Any suggestions are welcome. arXiv admin note: text overlap with arXiv:2406.13176

LiDAR odometry is a pivotal technology in the fields of autonomous driving and autonomous mobile robotics. However, most of the current works focus on nonlinear optimization methods, and still existing many challenges in using the traditional Iterative Extended Kalman Filter (IEKF) framework to tackle the problem: IEKF only iterates over the observation equation, relying on a rough estimate of the initial state, which is insufficient to fully eliminate motion distortion in the input point cloud; the system process noise is difficult to be determined during state estimation of the complex motions; and the varying motion models across different sensor carriers. To address these issues, we propose the Dual-Iteration Extended Kalman Filter (I2EKF) and the LiDAR odometry based on I2EKF (I2EKF-LO). This approach not only iterates over the observation equation but also leverages state updates to iteratively mitigate motion distortion in LiDAR point clouds. Moreover, it dynamically adjusts process noise based on the confidence level of prior predictions during state estimation and establishes motion models for different sensor carriers to achieve accurate and efficient state estimation. Comprehensive experiments demonstrate that I2EKF-LO achieves outstanding levels of accuracy and computational efficiency in the realm of LiDAR odometry. Additionally, to foster community development, our code is open-sourced.https://github.com/YWL0720/I2EKF-LO., Comment: Accepted by IROS 2024

This paper considers the collaborative graph exploration problem in GPS-denied environments, where a group of robots are required to cover a graph environment while maintaining reliable pose estimations in collaborative simultaneous localization and mapping (SLAM). Considering both objectives presents challenges for multi-robot pathfinding, as it involves the expensive covariance inference for SLAM uncertainty evaluation, especially considering various combinations of robots' paths. To reduce the computational complexity, we propose an efficient two-stage strategy where exploration paths are first generated for quick coverage, and then enhanced by adding informative and distance-efficient loop-closing actions, called loop edges, along the paths for reliable pose estimation. We formulate the latter problem as a non-monotone submodular maximization problem by relating SLAM uncertainty with pose graph topology, which (1) facilitates more efficient evaluation of SLAM uncertainty than covariance inference, and (2) allows the application of approximation algorithms in submodular optimization to provide optimality guarantees. We further introduce the ordering heuristics to improve objective values while preserving the optimality bound. Simulation experiments over randomly generated graph environments verify the efficiency of our methods in finding paths for quick coverage and enhanced pose graph reliability, and benchmark the performance of the approximation algorithms and the greedy-based algorithm in the loop edge selection problem. Our implementations will be open-source at https://github.com/bairuofei/CGE., Comment: 9 pages, 13 figures, accepted by IEEE/RSJ IROS(2024)

Most medical image lesion segmentation methods rely on hand-crafted accurate annotations of the original image for supervised learning. Recently, a series of weakly supervised or unsupervised methods have been proposed to reduce the dependence on pixel-level annotations. However, these methods are essentially based on pixel-level annotation, ignoring the image-level diagnostic results of the current massive medical images. In this paper, we propose a dual U-shaped two-stage framework that utilizes image-level labels to prompt the segmentation. In the first stage, we pre-train a classification network with image-level labels, which is used to obtain the hierarchical pyramid features and guide the learning of downstream branches. In the second stage, we feed the hierarchical features obtained from the classification branch into the downstream branch through short-skip and long-skip and get the lesion masks under the supervised learning of pixel-level labels. Experiments show that our framework achieves better results than networks simply using pixel-level annotations., Comment: Work in progress

A well-known theorem of Mantel states that every $n$-vertex graph with more than $\lfloor n^2/4\rfloor $ edges contains a triangle. An interesting problem in extremal graph theory studies the minimum number of edges contained in triangles among graphs with a prescribed number of vertices and edges. Erd\H{o}s, Faudree and Rousseau (1992) showed that a graph on $n$ vertices with more than $\lfloor n^2/4\rfloor $ edges contains at least $2\lfloor n/2\rfloor +1$ edges in triangles. Such edges are called triangular edges. In this paper, we present a spectral version of the result of Erd\H{o}s, Faudree and Rousseau. Using the supersaturation-stability and the spectral technique, we prove that every $n$-vertex graph $G$ with $\lambda (G) \ge \sqrt{\lfloor n^2/4\rfloor}$ contains at least $2 \lfloor {n}/{2} \rfloor -1$ triangular edges, unless $G$ is a balanced complete bipartite graph. The method in our paper has some interesting applications. Firstly, the supersaturation-stability can be used to revisit a conjecture of Erd\H{o}s concerning with the booksize of a graph, which was initially proved by Edwards (unpublished), and independently by Khad\v{z}iivanov and Nikiforov (1979). Secondly, our method can improve the bound on the order $n$ of a graph by dropping the condition on $n$ being sufficiently large, which is obtained from the triangle removal lemma. Thirdly, the supersaturation-stability can be applied to deal with the spectral extremal graph problems on counting triangles, which was recently studied by Ning and Zhai (2023)., Comment: 30 pages. At the end of the paper, we proposed many spectral extremal graph problems for readers. Any comments are welcome

When Large Vision Language Models (LVLMs) are applied to multimodal medical generative tasks, they suffer from significant model hallucination issues. This severely impairs the model's generative accuracy, making it challenging for LVLMs to be implemented in real-world medical scenarios to assist doctors in diagnosis. Enhancing the training data for downstream medical generative tasks is an effective way to address model hallucination. Moreover, the limited availability of training data in the medical field and privacy concerns greatly hinder the model's accuracy and generalization capabilities. In this paper, we introduce a method that mimics human cognitive processes to construct fine-grained instruction pairs and apply the concept of chain-of-thought (CoT) from inference scenarios to training scenarios, thereby proposing a method called MedThink. Our experiments on various LVLMs demonstrate that our novel data construction method tailored for the medical domain significantly improves the model's performance in medical image report generation tasks and substantially mitigates the hallucinations. All resources of this work will be released soon.

Large Vision Language Models (LVLMs) are increasingly integral to healthcare applications, including medical visual question answering and imaging report generation. While these models inherit the robust capabilities of foundational Large Language Models (LLMs), they also inherit susceptibility to hallucinations-a significant concern in high-stakes medical contexts where the margin for error is minimal. However, currently, there are no dedicated methods or benchmarks for hallucination detection and evaluation in the medical field. To bridge this gap, we introduce Med-HallMark, the first benchmark specifically designed for hallucination detection and evaluation within the medical multimodal domain. This benchmark provides multi-tasking hallucination support, multifaceted hallucination data, and hierarchical hallucination categorization. Furthermore, we propose the MediHall Score, a new medical evaluative metric designed to assess LVLMs' hallucinations through a hierarchical scoring system that considers the severity and type of hallucination, thereby enabling a granular assessment of potential clinical impacts. We also present MediHallDetector, a novel Medical LVLM engineered for precise hallucination detection, which employs multitask training for hallucination detection. Through extensive experimental evaluations, we establish baselines for popular LVLMs using our benchmark. The findings indicate that MediHall Score provides a more nuanced understanding of hallucination impacts compared to traditional metrics and demonstrate the enhanced performance of MediHallDetector. We hope this work can significantly improve the reliability of LVLMs in medical applications. All resources of this work will be released soon.

In studies of educational production functions or intergenerational mobility, it is common to transform the key variables into percentile ranks. Yet, it remains unclear what the regression coefficient estimates with ranks of the outcome or the treatment. In this paper, we derive effective causal estimands for a broad class of commonly-used regression methods, including the ordinary least squares (OLS), two-stage least squares (2SLS), difference-in-differences (DiD), and regression discontinuity designs (RDD). Specifically, we introduce a novel primitive causal estimand, the Rank Average Treatment Effect (rank-ATE), and prove that it serves as the building block of the effective estimands of all the aforementioned econometrics methods. For 2SLS, DiD, and RDD, we show that direct applications to outcome ranks identify parameters that are difficult to interpret. To address this issue, we develop alternative methods to identify more interpretable causal parameters.

To approximate sums of values in key-value data streams, sketches are widely used in databases and networking systems. They offer high-confidence approximations for any given key while ensuring low time and space overhead. While existing sketches are proficient in estimating individual keys, they struggle to maintain this high confidence across all keys collectively, an objective that is critically important in both algorithm theory and its practical applications. We propose ReliableSketch, the first to control the error of all keys to less than $\Lambda$ with a small failure probability $\Delta$, requiring only $O(1 + \Delta\ln\ln(\frac{N}{\Lambda}))$ amortized time and $O(\frac{N}{\Lambda} + \ln(\frac{1}{\Delta}))$ space. Furthermore, its simplicity makes it hardware-friendly, and we implement it on CPU servers, FPGAs, and programmable switches. Our experiments show that under the same small space, ReliableSketch not only keeps all keys' errors below $\Lambda$ but also achieves near-optimal throughput, outperforming competitors with thousands of uncontrolled estimations. We have made our source code publicly available.

Developing intelligent pediatric consultation systems offers promising prospects for improving diagnostic efficiency, especially in China, where healthcare resources are scarce. Despite recent advances in Large Language Models (LLMs) for Chinese medicine, their performance is sub-optimal in pediatric applications due to inadequate instruction data and vulnerable training procedures. To address the above issues, this paper builds PedCorpus, a high-quality dataset of over 300,000 multi-task instructions from pediatric textbooks, guidelines, and knowledge graph resources to fulfil diverse diagnostic demands. Upon well-designed PedCorpus, we propose PediatricsGPT, the first Chinese pediatric LLM assistant built on a systematic and robust training pipeline. In the continuous pre-training phase, we introduce a hybrid instruction pre-training mechanism to mitigate the internal-injected knowledge inconsistency of LLMs for medical domain adaptation. Immediately, the full-parameter Supervised Fine-Tuning (SFT) is utilized to incorporate the general medical knowledge schema into the models. After that, we devise a direct following preference optimization to enhance the generation of pediatrician-like humanistic responses. In the parameter-efficient secondary SFT phase, a mixture of universal-specific experts strategy is presented to resolve the competency conflict between medical generalist and pediatric expertise mastery. Extensive results based on the metrics, GPT-4, and doctor evaluations on distinct doctor downstream tasks show that PediatricsGPT consistently outperforms previous Chinese medical LLMs. Our model and dataset will be open-source for community development., Comment: A Technical Report on a Chinese Medical Large Language Model

In the process industry, optimizing production lines for long-term efficiency requires real-time monitoring and analysis of operation states to fine-tune production line parameters. However, the complexity in operational logic and the intricate coupling of production process parameters make it difficult to develop an accurate mathematical model for the entire process, thus hindering the deployment of efficient optimization mechanisms. In view of these difficulties, we propose to deploy a digital twin of the production line by digitally abstracting its physical layout and operational logic. By iteratively mapping the real-world data reflecting equipment operation status and product quality inspection in the digital twin, we adopt a quality prediction model for production process based on self-attention-enabled temporal convolutional neural networks. This model enables the data-driven state evolution of the digital twin. The digital twin takes a role of aggregating the information of actual operating conditions and the results of quality-sensitive analysis, which facilitates the optimization of process production quality with virtual-reality evolution under multi-dimensional constraints. Leveraging the digital twin model as an information-flow carrier, we extract temporal features from key process indicators and establish a production process quality prediction model based on the proposed composite neural network. Our operation experiments on a specific tobacco shredding line demonstrate that the proposed digital twin-based production process optimization method fosters seamless integration between virtual and real production lines. This integration achieves an average operating status prediction accuracy of over 98\% and near-optimal production process control.

Join size estimation on sensitive data poses a risk of privacy leakage. Local differential privacy (LDP) is a solution to preserve privacy while collecting sensitive data, but it introduces significant noise when dealing with sensitive join attributes that have large domains. Employing probabilistic structures such as sketches is a way to handle large domains, but it leads to hash-collision errors. To achieve accurate estimations, it is necessary to reduce both the noise error and hash-collision error. To tackle the noise error caused by protecting sensitive join values with large domains, we introduce a novel algorithm called LDPJoinSketch for sketch-based join size estimation under LDP. Additionally, to address the inherent hash-collision errors in sketches under LDP, we propose an enhanced method called LDPJoinSketch+. It utilizes a frequency-aware perturbation mechanism that effectively separates high-frequency and low-frequency items without compromising privacy. The proposed methods satisfy LDP, and the estimation error is bounded. Experimental results show that our method outperforms existing methods, effectively enhancing the accuracy of join size estimation under LDP.

Motion style transfer is a significant research direction in multimedia applications. It enables the rapid switching of different styles of the same motion for virtual digital humans, thus vastly increasing the diversity and realism of movements. It is widely applied in multimedia scenarios such as movies, games, and the Metaverse. However, most of the current work in this field adopts the GAN, which may lead to instability and convergence issues, making the final generated motion sequence somewhat chaotic and unable to reflect a highly realistic and natural style. To address these problems, we consider style motion as a condition and propose the Style Motion Conditioned Diffusion (SMCD) framework for the first time, which can more comprehensively learn the style features of motion. Moreover, we apply Mamba model for the first time in the motion style transfer field, introducing the Motion Style Mamba (MSM) module to handle longer motion sequences. Thirdly, aiming at the SMCD framework, we propose Diffusion-based Content Consistency Loss and Content Consistency Loss to assist the overall framework's training. Finally, we conduct extensive experiments. The results reveal that our method surpasses state-of-the-art methods in both qualitative and quantitative comparisons, capable of generating more realistic motion sequences.

Survival prediction is a complex ordinal regression task that aims to predict the survival coefficient ranking among a cohort of patients, typically achieved by analyzing patients' whole slide images. Existing deep learning approaches mainly adopt multiple instance learning or graph neural networks under weak supervision. Most of them are unable to uncover the diverse interactions between different types of biological entities(\textit{e.g.}, cell cluster and tissue block) across multiple scales, while such interactions are crucial for patient survival prediction. In light of this, we propose a novel multi-scale heterogeneity-aware hypergraph representation framework. Specifically, our framework first constructs a multi-scale heterogeneity-aware hypergraph and assigns each node with its biological entity type. It then mines diverse interactions between nodes on the graph structure to obtain a global representation. Experimental results demonstrate that our method outperforms state-of-the-art approaches on three benchmark datasets. Code is publicly available at \href{https://github.com/Hanminghao/H2GT}{https://github.com/Hanminghao/H2GT}., Comment: 9 pages, 6 figures, accepted by ICME2024

Multimodal sentiment analysis (MSA) aims to understand human sentiment through multimodal data. Most MSA efforts are based on the assumption of modality completeness. However, in real-world applications, some practical factors cause uncertain modality missingness, which drastically degrades the model's performance. To this end, we propose a Correlation-decoupled Knowledge Distillation (CorrKD) framework for the MSA task under uncertain missing modalities. Specifically, we present a sample-level contrastive distillation mechanism that transfers comprehensive knowledge containing cross-sample correlations to reconstruct missing semantics. Moreover, a category-guided prototype distillation mechanism is introduced to capture cross-category correlations using category prototypes to align feature distributions and generate favorable joint representations. Eventually, we design a response-disentangled consistency distillation strategy to optimize the sentiment decision boundaries of the student network through response disentanglement and mutual information maximization. Comprehensive experiments on three datasets indicate that our framework can achieve favorable improvements compared with several baselines., Comment: Accepted by CVPR 2024

Adversarial patch attacks present a significant threat to real-world object detectors due to their practical feasibility. Existing defense methods, which rely on attack data or prior knowledge, struggle to effectively address a wide range of adversarial patches. In this paper, we show two inherent characteristics of adversarial patches, semantic independence and spatial heterogeneity, independent of their appearance, shape, size, quantity, and location. Semantic independence indicates that adversarial patches operate autonomously within their semantic context, while spatial heterogeneity manifests as distinct image quality of the patch area that differs from original clean image due to the independent generation process. Based on these observations, we propose PAD, a novel adversarial patch localization and removal method that does not require prior knowledge or additional training. PAD offers patch-agnostic defense against various adversarial patches, compatible with any pre-trained object detectors. Our comprehensive digital and physical experiments involving diverse patch types, such as localized noise, printable, and naturalistic patches, exhibit notable improvements over state-of-the-art works. Our code is available at https://github.com/Lihua-Jing/PAD., Comment: Accepted by CVPR 2024

In the realm of Medical Visual Language Models (Med-VLMs), the quest for universal efficient fine-tuning mechanisms remains paramount, especially given researchers in interdisciplinary fields are often extremely short of training resources, yet largely unexplored. Given the unique challenges in the medical domain, such as limited data scope and significant domain-specific requirements, evaluating and adapting Parameter-Efficient Fine-Tuning (PEFT) methods specifically for Med-VLMs is essential. Most of the current PEFT methods on Med-VLMs have yet to be comprehensively investigated but mainly focus on adding some components to the model's structure or input. However, fine-tuning intrinsic model components often yields better generality and consistency, and its impact on the ultimate performance of Med-VLMs has been widely overlooked and remains understudied. In this paper, we endeavour to explore an alternative to traditional PEFT methods, especially the impact of fine-tuning LayerNorm layers, FFNs and Attention layers on the Med-VLMs. Our comprehensive studies span both small-scale and large-scale Med-VLMs, evaluating their performance under various fine-tuning paradigms across tasks such as Medical Visual Question Answering and Medical Imaging Report Generation. The findings reveal unique insights into the effects of intrinsic parameter fine-tuning methods on fine-tuning Med-VLMs to downstream tasks and expose fine-tuning solely the LayerNorm layers not only surpasses the efficiency of traditional PEFT methods but also retains the model's accuracy and generalization capabilities across a spectrum of medical downstream tasks. The experiments show LayerNorm fine-tuning's superior adaptability and scalability, particularly in the context of large-scale Med-VLMs.

Recently, backdoor attacks have posed a serious security threat to the training process of deep neural networks (DNNs). The attacked model behaves normally on benign samples but outputs a specific result when the trigger is present. However, compared with the rocketing progress of backdoor attacks, existing defenses are difficult to deal with these threats effectively or require benign samples to work, which may be unavailable in real scenarios. In this paper, we find that the poisoned samples and benign samples can be distinguished with prediction entropy. This inspires us to propose a novel dual-network training framework: The Victim and The Beneficiary (V&B), which exploits a poisoned model to train a clean model without extra benign samples. Firstly, we sacrifice the Victim network to be a powerful poisoned sample detector by training on suspicious samples. Secondly, we train the Beneficiary network on the credible samples selected by the Victim to inhibit backdoor injection. Thirdly, a semi-supervised suppression strategy is adopted for erasing potential backdoors and improving model performance. Furthermore, to better inhibit missed poisoned samples, we propose a strong data augmentation method, AttentionMix, which works well with our proposed V&B framework. Extensive experiments on two widely used datasets against 6 state-of-the-art attacks demonstrate that our framework is effective in preventing backdoor injection and robust to various attacks while maintaining the performance on benign samples. Our code is available at https://github.com/Zixuan-Zhu/VaB., Comment: 13 pages, 6 figures, published to ICCV

Visual Odometry (VO) is vital for the navigation of autonomous systems, providing accurate position and orientation estimates at reasonable costs. While traditional VO methods excel in some conditions, they struggle with challenges like variable lighting and motion blur. Deep learning-based VO, though more adaptable, can face generalization problems in new environments. Addressing these drawbacks, this paper presents a novel hybrid visual odometry (VO) framework that leverages pose-only supervision, offering a balanced solution between robustness and the need for extensive labeling. We propose two cost-effective and innovative designs: a self-supervised homographic pre-training for enhancing optical flow learning from pose-only labels and a random patch-based salient point detection strategy for more accurate optical flow patch extraction. These designs eliminate the need for dense optical flow labels for training and significantly improve the generalization capability of the system in diverse and challenging environments. Our pose-only supervised method achieves competitive performance on standard datasets and greater robustness and generalization ability in extreme and unseen scenarios, even compared to dense optical flow-supervised state-of-the-art methods., Comment: Accepted by IEEE Robotics and Automation Letters

Point-based interactive editing serves as an essential tool to complement the controllability of existing generative models. A concurrent work, DragDiffusion, updates the diffusion latent map in response to user inputs, causing global latent map alterations. This results in imprecise preservation of the original content and unsuccessful editing due to gradient vanishing. In contrast, we present DragNoise, offering robust and accelerated editing without retracing the latent map. The core rationale of DragNoise lies in utilizing the predicted noise output of each U-Net as a semantic editor. This approach is grounded in two critical observations: firstly, the bottleneck features of U-Net inherently possess semantically rich features ideal for interactive editing; secondly, high-level semantics, established early in the denoising process, show minimal variation in subsequent stages. Leveraging these insights, DragNoise edits diffusion semantics in a single denoising step and efficiently propagates these changes, ensuring stability and efficiency in diffusion editing. Comparative experiments reveal that DragNoise achieves superior control and semantic retention, reducing the optimization time by over 50% compared to DragDiffusion. Our codes are available at https://github.com/haofengl/DragNoise., Comment: Accepted by CVPR 2024

Data-Free Knowledge Distillation (DFKD) is a promising task to train high-performance small models to enhance actual deployment without relying on the original training data. Existing methods commonly avoid relying on private data by utilizing synthetic or sampled data. However, a long-overlooked issue is that the severe distribution shifts between their substitution and original data, which manifests as huge differences in the quality of images and class proportions. The harmful shifts are essentially the confounder that significantly causes performance bottlenecks. To tackle the issue, this paper proposes a novel perspective with causal inference to disentangle the student models from the impact of such shifts. By designing a customized causal graph, we first reveal the causalities among the variables in the DFKD task. Subsequently, we propose a Knowledge Distillation Causal Intervention (KDCI) framework based on the backdoor adjustment to de-confound the confounder. KDCI can be flexibly combined with most existing state-of-the-art baselines. Experiments in combination with six representative DFKD methods demonstrate the effectiveness of our KDCI, which can obviously help existing methods under almost all settings, \textit{e.g.}, improving the baseline by up to 15.54\% accuracy on the CIFAR-100 dataset., Comment: Accepted by CVPR24

Human pose estimation (HPE) from Radio Frequency vision (RF-vision) performs human sensing using RF signals that penetrate obstacles without revealing privacy (e.g., facial information). Recently, mmWave radar has emerged as a promising RF-vision sensor, providing radar point clouds by processing RF signals. However, the mmWave radar has a limited resolution with severe noise, leading to inaccurate and inconsistent human pose estimation. This work proposes mmDiff, a novel diffusion-based pose estimator tailored for noisy radar data. Our approach aims to provide reliable guidance as conditions to diffusion models. Two key challenges are addressed by mmDiff: (1) miss-detection of parts of human bodies, which is addressed by a module that isolates feature extraction from different body parts, and (2) signal inconsistency due to environmental interference, which is tackled by incorporating prior knowledge of body structure and motion. Several modules are designed to achieve these goals, whose features work as the conditions for the subsequent diffusion model, eliminating the miss-detection and instability of HPE based on RF-vision. Extensive experiments demonstrate that mmDiff outperforms existing methods significantly, achieving state-of-the-art performances on public datasets.

Inspired by the behavior of birds, we present AirCrab, a hybrid aerial ground manipulator (HAGM) with a single active wheel and a 3-degree of freedom (3-DoF) manipulator. AirCrab leverages a single point of contact with the ground to reduce position drift and improve manipulation accuracy. The single active wheel enables locomotion on narrow surfaces without adding significant weight to the robot. To realize accurate attitude maintenance using propellers on the ground, we design a control allocation method for AirCrab that prioritizes attitude control and dynamically adjusts the thrust input to reduce energy consumption. Experiments verify the effectiveness of the proposed control method and the gain in manipulation accuracy with ground contact. A series of operations to complete the letters 'NTU' demonstrates the capability of the robot to perform challenging hybrid aerial-ground manipulation missions.

The precise measurement of the gravity of the earth plays a pivotal role in various fundamental research and application fields. Although a few gravimeters have been reported to achieve this goal, miniaturization of high-precision gravimetry remains a challenge. In this work, we have proposed and demonstrated a miniaturized gravimetry operating at room temperature based on a diamagnetic levitated micro-oscillator with a proof mass of only 215 mg. Compared with the latest reported miniaturized gravimeters based on Micro-Electro-Mechanical Systems, the performance of our gravimetry has substantial improvements in that an acceleration sensitivity of 15 $\mu Gal/\sqrt{Hz}$ and a drift as low as 61 $\mu Gal$ per day have been reached. Based on this diamagnetic levitation gravimetry, we observed the earth tides, and the correlation coefficient between the experimental data and theoretical data reached 0.97. Some moderate foreseeable improvements can develop this diamagnetic levitation gravimetry into chip size device, making it suitable for mobile platforms such as drones. Our advancement in gravimetry is expected to facilitate a multitude of applications, including underground density surveying and the forecasting of natural hazards., Comment: 5 pages, 3 figures

Compact wearable mapping system (WMS) has gained significant attention due to their convenience in various applications. Specifically, it provides an efficient way to collect prior maps for 3D structure inspection and robot-based "last-mile delivery" in complex environments. However, vibrations in human motion and the uneven distribution of point cloud features in complex environments often lead to rapid drift, which is a prevalent issue when applying existing LiDAR Inertial Odometry (LIO) methods on low-cost WMS. To address these limitations, we propose a novel LIO for WMSs based on Hybrid Continuous Time Optimization (HCTO) considering the optimality of Lidar correspondences. First, HCTO recognizes patterns in human motion (high-frequency part, low-frequency part, and constant velocity part) by analyzing raw IMU measurements. Second, HCTO constructs hybrid IMU factors according to different motion states, which enables robust and accurate estimation against vibration-induced noise in the IMU measurements. Third, the best point correspondences are selected using optimal design to achieve real-time performance and better odometry accuracy. We conduct experiments on head-mounted WMS datasets to evaluate the performance of our system, demonstrating significant advantages over state-of-the-art methods. Video recordings of experiments can be found on the project page of HCTO: \href{https://github.com/kafeiyin00/HCTO}{https://github.com/kafeiyin00/HCTO}.

It is commonly understood that the strong magnetic field produced in heavy ion collisions is short-lived. The electric conductivity of the quark-gluon plasma is unable to significantly extend the life time of magnetic field. We propose an alternative scenario to achieve this: with finite baryon density and spin polarization by the initial magnetic field, the quark-gluon plasma behaves as a paramagnet, which may continue to polarize quark after fading of initial magnetic field. We confirm this picture by calculations in both quantum electrodynamics and quantum chromodynamics. In the former case, we find a splitting in the damping rates of probe fermion with opposite spin component along the magnetic field. In the latter case, we find a similar splitting in damping rate of probe quark in quark-gluon plasma in both high density and low density limits. The splitting provides a way of polarizing strange quarks by the quark-gluon plasma paramagnet consisting of light quarks, which effectively extends the lifetime of magnetic field in heavy ion collisions.

Perception plays a crucial role in various robot applications. However, existing well-annotated datasets are biased towards autonomous driving scenarios, while unlabelled SLAM datasets are quickly over-fitted, and often lack environment and domain variations. To expand the frontier of these fields, we introduce a comprehensive dataset named MCD (Multi-Campus Dataset), featuring a wide range of sensing modalities, high-accuracy ground truth, and diverse challenging environments across three Eurasian university campuses. MCD comprises both CCS (Classical Cylindrical Spinning) and NRE (Non-Repetitive Epicyclic) lidars, high-quality IMUs (Inertial Measurement Units), cameras, and UWB (Ultra-WideBand) sensors. Furthermore, in a pioneering effort, we introduce semantic annotations of 29 classes over 59k sparse NRE lidar scans across three domains, thus providing a novel challenge to existing semantic segmentation research upon this largely unexplored lidar modality. Finally, we propose, for the first time to the best of our knowledge, continuous-time ground truth based on optimization-based registration of lidar-inertial data on large survey-grade prior maps, which are also publicly released, each several times the size of existing ones. We conduct a rigorous evaluation of numerous state-of-the-art algorithms on MCD, report their performance, and highlight the challenges awaiting solutions from the research community., Comment: Accepted by The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2024

Multi-modal test-time adaptation (MM-TTA) is proposed to adapt models to an unlabeled target domain by leveraging the complementary multi-modal inputs in an online manner. Previous MM-TTA methods for 3D segmentation rely on predictions of cross-modal information in each input frame, while they ignore the fact that predictions of geometric neighborhoods within consecutive frames are highly correlated, leading to unstable predictions across time. To fulfill this gap, we propose ReLiable Spatial-temporal Voxels (Latte), an MM-TTA method that leverages reliable cross-modal spatial-temporal correspondences for multi-modal 3D segmentation. Motivated by the fact that reliable predictions should be consistent with their spatial-temporal correspondences, Latte aggregates consecutive frames in a slide window manner and constructs Spatial-Temopral (ST) voxels to capture temporally local prediction consistency for each modality. After filtering out ST voxels with high ST entropy, Latte conducts cross-modal learning for each point and pixel by attending to those with reliable and consistent predictions among both spatial and temporal neighborhoods. Experimental results show that Latte achieves state-of-the-art performance on three different MM-TTA benchmarks compared to previous MM-TTA or TTA methods. Visit our project site https://sites.google.com/view/eccv24-latte.

While Large Language Models (LLMs) excel in world knowledge understanding, adapting them to specific subfields requires precise adjustments. Due to the model's vast scale, traditional global fine-tuning methods for large models can be computationally expensive and impact generalization. To address this challenge, a range of innovative Parameters-Efficient Fine-Tuning (PEFT) methods have emerged and achieved remarkable success in both LLMs and Large Vision-Language Models (LVLMs). In the medical domain, fine-tuning a medical Vision-Language Pretrained (VLP) model is essential for adapting it to specific tasks. Can the fine-tuning methods for large models be transferred to the medical field to enhance transfer learning efficiency? In this paper, we delve into the fine-tuning methods of LLMs and conduct extensive experiments to investigate the impact of fine-tuning methods for large models on the existing multimodal model in the medical domain from the training data level and the model structure level. We show the different impacts of fine-tuning methods for large models on medical VLMs and develop the most efficient ways to fine-tune medical VLP models. We hope this research can guide medical domain researchers in optimizing VLMs' training costs, fostering the broader application of VLMs in healthcare fields. The code and dataset have been released at https://github.com/TIMMY-CHAN/MILE., Comment: Accepted by MICCAI 2024

Accurate and consistent construction of point clouds from LiDAR scanning data is fundamental for 3D modeling applications. Current solutions, such as multiview point cloud registration and LiDAR bundle adjustment, predominantly depend on the local plane assumption, which may be inadequate in complex environments lacking of planar geometries or substantial initial pose errors. To mitigate this problem, this paper presents a LiDAR bundle adjustment with progressive spatial smoothing, which is suitable for complex environments and exhibits improved convergence capabilities. The proposed method consists of a spatial smoothing module and a pose adjustment module, which combines the benefits of local consistency and global accuracy. With the spatial smoothing module, we can obtain robust and rich surface constraints employing smoothing kernels across various scales. Then the pose adjustment module corrects all poses utilizing the novel surface constraints. Ultimately, the proposed method simultaneously achieves fine poses and parametric surfaces that can be directly employed for high-quality point cloud reconstruction. The effectiveness and robustness of our proposed approach have been validated on both simulation and real-world datasets. The experimental results demonstrate that the proposed method outperforms the existing methods and achieves better accuracy in complex environments with low planar structures.