Search

Accurately reconstructing the global ocean deoxygenation over a century is crucial for assessing and protecting marine ecosystem. Existing expert-dominated numerical simulations fail to catch up with the dynamic variation caused by global warming and human activities. Besides, due to the high-cost data collection, the historical observations are severely sparse, leading to big challenge for precise reconstruction. In this work, we propose OxyGenerator, the first deep learning based model, to reconstruct the global ocean deoxygenation from 1920 to 2023. Specifically, to address the heterogeneity across large temporal and spatial scales, we propose zoning-varying graph message-passing to capture the complex oceanographic correlations between missing values and sparse observations. Additionally, to further calibrate the uncertainty, we incorporate inductive bias from dissolved oxygen (DO) variations and chemical effects. Compared with in-situ DO observations, OxyGenerator significantly outperforms CMIP6 numerical simulations, reducing MAPE by 38.77%, demonstrating a promising potential to understand the "breathless ocean" in data-driven manner., Comment: Accepted to ICML 2024

Automatic evaluation metrics for generated texts play an important role in the NLG field, especially with the rapid growth of LLMs. However, existing metrics are often limited to specific scenarios, making it challenging to meet the evaluation requirements of expanding LLM applications. Therefore, there is a demand for new, flexible, and effective metrics. In this study, we introduce RepEval, the first metric leveraging the projection of LLM representations for evaluation. RepEval requires minimal sample pairs for training, and through simple prompt modifications, it can easily transition to various tasks. Results on ten datasets from three tasks demonstrate the high effectiveness of our method, which exhibits stronger correlations with human judgments compared to previous metrics, even outperforming GPT-4. Our work underscores the richness of information regarding text quality embedded within LLM representations, offering insights for the development of new metrics.

This paper aims to present objective methods for constructing new fuzzy sets from known fuzzy or classical sets, defined over the elements of a finite universe's superstructure. The paper proposes rules for assigning membership functions to these new fuzzy sets, leading to two important findings. Firstly, the property concerning the cardinality of a power set in classical theory has been extended to the fuzzy setting, whereby the scalar cardinality of a fuzzy set $\tilde B$ defined on the power set of a finite universe of a fuzzy set $\tilde A$ satisfies $\text{card}(\tilde B)=2^{\text{card}(\tilde A)}$. Secondly, the novel algorithms allow for an arbitrary membership value to be objectively achieved and represented by a specific binary sequence.

In the realm of robotic grasping, achieving accurate and reliable interactions with the environment is a pivotal challenge. Traditional methods of grasp planning methods utilizing partial point clouds derived from depth image often suffer from reduced scene understanding due to occlusion, ultimately impeding their grasping accuracy. Furthermore, scene reconstruction methods have primarily relied upon static techniques, which are susceptible to environment change during manipulation process limits their efficacy in real-time grasping tasks. To address these limitations, this paper introduces a novel two-stage pipeline for dynamic scene reconstruction. In the first stage, our approach takes scene scanning as input to register each target object with mesh reconstruction and novel object pose tracking. In the second stage, pose tracking is still performed to provide object poses in real-time, enabling our approach to transform the reconstructed object point clouds back into the scene. Unlike conventional methodologies, which rely on static scene snapshots, our method continuously captures the evolving scene geometry, resulting in a comprehensive and up-to-date point cloud representation. By circumventing the constraints posed by occlusion, our method enhances the overall grasp planning process and empowers state-of-the-art 6-DoF robotic grasping algorithms to exhibit markedly improved accuracy., Comment: ICRA 2024

The majority of automatic metrics for evaluating NLG systems are reference-based. However, the challenge of collecting human annotation results in a lack of reliable references in numerous application scenarios. Despite recent advancements in reference-free metrics, it has not been well understood when and where they can be used as an alternative to reference-based metrics. In this study, by employing diverse analytical approaches, we comprehensively assess the performance of both metrics across a wide range of NLG tasks, encompassing eight datasets and eight evaluation models. Based on solid experiments, the results show that reference-free metrics exhibit a higher correlation with human judgment and greater sensitivity to deficiencies in language quality. However, their effectiveness varies across tasks and is influenced by the quality of candidate texts. Therefore, it's important to assess the performance of reference-free metrics before applying them to a new task, especially when inputs are in uncommon form or when the answer space is highly variable. Our study can provide insight into the appropriate application of automatic metrics and the impact of metric choice on evaluation performance.

Structured light, particularly in the terahertz frequency range, holds considerable potential for a diverse range of applications. However, the generation and control of structured terahertz radiation pose major challenges. In this work, we demonstrate a novel programmable spintronic emitter that can flexibly generate a variety of structured terahertz waves. This is achieved through the precise and high-resolution programming of the magnetization pattern on the emitter surface, utilizing laser-assisted local field cooling of an exchange-biased ferromagnetic heterostructure. Moreover, we outline a generic design strategy for realizing specific complex structured terahertz fields in the far field. Our device successfully demonstrates the generation of terahertz waves with diverse structured polarization states, including spatially separated circular polarizations, azimuthal or radial polarization states, and a full Poincare beam. This innovation opens a new avenue for designing and generating structured terahertz radiations, with potential applications in terahertz microscopy, communication, quantum information, and light-matter interactions.

Precision motion systems are at the core of various manufacturing equipment. The rapidly increasing demand for higher productivity necessitates higher control bandwidth in the motion systems to effectively reject disturbances while maintaining excellent positioning accuracy. However, most existing optimal control methods do not explicitly optimize for control bandwidth, and the classic loop-shaping method suffers from conservative designs and fails to address cross-couplings, which motivates the development of new control solutions for bandwidth optimization. This paper proposes a novel bandwidth optimal control formulation based on nonsmooth optimization for precision motion systems. Our proposed method explicitly optimizes the system's MIMO control bandwidth while constraining the H-infinity norm of the closed-loop sensitivity function for robustness. A nonsmooth optimization solver, GRANSO, is used to solve the proposed program, and an augmented quadratic programming (QP)--based descent direction search is proposed to facilitate convergence. Simulation evaluations show that the bandwidth optimal control method can achieve a 23% higher control bandwidth than conventional loop-shaping design, and the QP-based descent direction search can reduce iteration number by 60%, which illustrates the effectiveness and efficiency of the proposed approach.

This paper tackles the challenges of self-supervised monocular depth estimation in indoor scenes caused by large rotation between frames and low texture. We ease the learning process by obtaining coarse camera poses from monocular sequences through multi-view geometry to deal with the former. However, we found that limited by the scale ambiguity across different scenes in the training dataset, a na\"ive introduction of geometric coarse poses cannot play a positive role in performance improvement, which is counter-intuitive. To address this problem, we propose to refine those poses during training through rotation and translation/scale optimization. To soften the effect of the low texture, we combine the global reasoning of vision transformers with an overfitting-aware, iterative self-distillation mechanism, providing more accurate depth guidance coming from the network itself. Experiments on NYUv2, ScanNet, 7scenes, and KITTI datasets support the effectiveness of each component in our framework, which sets a new state-of-the-art for indoor self-supervised monocular depth estimation, as well as outstanding generalization ability. Code and models are available at https://github.com/zxcqlf/GasMono, Comment: ICCV 2023. Code: https://github.com/zxcqlf/GasMono

Micro/Nano-positioning stages are of great importance in a wide range of manufacturing machines and instruments. In recent years, the drastically growing demand for higher throughput and reduced power consumption in various IC manufacturing equipment calls for the development of next-generation precision positioning systems with unprecedented acceleration capability while maintaining exceptional positioning accuracy and high control bandwidth. Reducing the stage's weight is an effective approach to achieving this goal. However, the reduction of stages' weight tends to decrease its structural resonance frequency, which limits the closed-loop control bandwidth and can even cause stability issues. Aiming to overcome the aforementioned challenge and thus create new lightweight precision stages with substantially improved acceleration capability without sacrificing stage control performance, this research presents a novel sequential structure and control design framework for lightweight stages with low-frequency flexible modes of the stage being actively controlled. Additional actuators and sensors are placed to actively control the flexible structural dynamics of the lightweight stage to attain high control bandwidth. A case study is simulated to evaluate the effectiveness of the proposed approach, where a stage weight reduction of 24% is demonstrated compared to a baseline case, which demonstrates the potential of the proposed design framework. Experimental evaluation of the designed stage's motion performance will be performed on a magnetically levitated linear motor platform for performance demonstration., Comment: arXiv admin note: substantial text overlap with arXiv:2301.04208; text overlap with arXiv:2309.11735

Understanding the evolution of cooperation in structured populations represented by networks is a problem of long research interest, and a most fundamental and widespread property of social networks related to cooperation phenomena is that the node's degree (i.e., number of edges connected to the node) is heterogeneously distributed. Previous results indicate that static heterogeneous (i.e., degree-heterogeneous) networks promote cooperation in stationarity compared to static regular (i.e., degree-homogeneous) networks if equilibrium dynamics starting from many cooperators and defectors is employed. However, the above conclusion reverses if we employ non-equilibrium stochastic processes to measure the fixation probability for cooperation, i.e., the probability that a single cooperator successfully invades a population. Here we resolve this conundrum by analyzing the fixation of cooperation on temporal (i.e., time-varying) networks. We theoretically prove and numerically confirm that on both synthetic and empirical networks, contrary to the case of static networks, temporal heterogeneous networks can promote cooperation more than temporal regular networks in terms of the fixation probability of cooperation. Given that the same conclusion is known for the equilibrium fraction of cooperators on temporal networks, the present results provide a unified understanding of the effect of temporal degree heterogeneity on promoting cooperation across two main analytical frameworks, i.e., equilibrium and non-equilibrium ones., Comment: 7 pages, 4 figures

Precision motion stages play a critical role in various manufacturing and inspection equipment, for example, the wafer/reticle scanning in photolithography scanners and positioning stages in wafer inspection systems. To meet the growing demand for higher throughput in chip manufacturing and inspection, it is critical to create new precision motion stages with higher acceleration capability with high control bandwidth, which calls for the development of lightweight precision stages. However, in today's precision motion systems, only the rigid body motion of the system are under control, and the flexible dynamic systems are in open loop. For these systems, the motion control bandwidth is limited by the first structural resonance frequency of the stage, which enforces a fundamental trade-off between the stage's bandwidth and acceleration capability. Aiming to overcome this trade-off, we have introduced a sequential structure and control design framework for lightweight stages with the low-frequency flexible modes of the stage are under active control. To facilitate the controller design, we further propose to minimize the resonance frequency of the stage's mode being controlled and to maximize the resonance frequency of the uncontrolled mode. The system's control bandwidth is placed in between the resonance frequencies. This paper presents the design, optimization, building, and experimental evaluations for a lightweight magnetically levitated planar stage, which we call FleXstage, with first flexible mode actively controlled via over-actuation. Simulations show the proposed design is highly promising in enabling stages with lightweight without sacrificing control bandwidth. We have some preliminary results now and are still working on the experimental evaluations for the closed-loop system, and will present the results in the oral presentation.

Category-level object pose estimation involves estimating the 6D pose and the 3D metric size of objects from predetermined categories. While recent approaches take categorical shape prior information as reference to improve pose estimation accuracy, the single-stage network design and training manner lead to sub-optimal performance since there are two distinct tasks in the pipeline. In this paper, the advantage of two-stage pipeline over single-stage design is discussed. To this end, we propose a two-stage deformation-and registration pipeline called DR-Pose, which consists of completion-aided deformation stage and scaled registration stage. The first stage uses a point cloud completion method to generate unseen parts of target object, guiding subsequent deformation on the shape prior. In the second stage, a novel registration network is designed to extract pose-sensitive features and predict the representation of object partial point cloud in canonical space based on the deformation results from the first stage. DR-Pose produces superior results to the state-of-the-art shape prior-based methods on both CAMERA25 and REAL275 benchmarks. Codes are available at https://github.com/Zray26/DR-Pose.git., Comment: Camera-ready version accepted to IROS 2023