Search

In this paper, we investigate ultra-wideband (UWB) localization and tracking in cluttered environments. Instead of mitigating the multipath, we exploit the specular reflections to enhance the localizability and improve the positioning accuracy. With the assistance of the multipath, it is also possible to achieve localization purposes using fewer anchors or when the line-of-sight propagations are blocked. Rather than using single-value distance, angle, or Doppler estimates for the localization, we model the likelihoods of both the line-of-sight and specular multipath components, namely soft multipath information, and propose the multipath-assisted probabilistic UWB tracking algorithm. Experimental results in a cluttered industrial scenario show that the proposed algorithm achieves 46.4 cm and 33.1 cm 90th percentile errors in the cases of 3 and 4 anchors, respectively, which outperforms conventional methods with more than 61.8% improvement given fewer anchors and strong multipath effect.

Molecule-and-text cross-modal representation learning has emerged as a promising direction for enhancing the quality of molecular representation, thereby improving performance in various scientific fields, including drug discovery and materials science. Existing studies adopt a global alignment approach to learn the knowledge from different modalities. These global alignment approaches fail to capture fine-grained information, such as molecular fragments and their corresponding textual description, which is crucial for downstream tasks. Furthermore, it is incapable to model such information using a similar global alignment strategy due to data scarcity of paired local part annotated data from existing datasets. In this paper, we propose Atomas, a multi-modal molecular representation learning framework to jointly learn representations from SMILES string and text. We design a Hierarchical Adaptive Alignment model to concurrently learn the fine-grained fragment correspondence between two modalities and align these representations of fragments in three levels. Additionally, Atomas's end-to-end training framework incorporates the tasks of understanding and generating molecule, thereby supporting a wider range of downstream tasks. In the retrieval task, Atomas exhibits robust generalization ability and outperforms the baseline by 30.8% of recall@1 on average. In the generation task, Atomas achieves state-of-the-art results in both molecule captioning task and molecule generation task. Moreover, the visualization of the Hierarchical Adaptive Alignment model further confirms the chemical significance of our approach. Our codes can be found at https://anonymous.4open.science/r/Atomas-03C3.

The core challenge of de novo protein design lies in creating proteins with specific functions or properties, guided by certain conditions. Current models explore to generate protein using structural and evolutionary guidance, which only provide indirect conditions concerning functions and properties. However, textual annotations of proteins, especially the annotations for protein domains, which directly describe the protein's high-level functionalities, properties, and their correlation with target amino acid sequences, remain unexplored in the context of protein design tasks. In this paper, we propose Protein-Annotation Alignment Generation (PAAG), a multi-modality protein design framework that integrates the textual annotations extracted from protein database for controllable generation in sequence space. Specifically, within a multi-level alignment module, PAAG can explicitly generate proteins containing specific domains conditioned on the corresponding domain annotations, and can even design novel proteins with flexible combinations of different kinds of annotations. Our experimental results underscore the superiority of the aligned protein representations from PAAG over 7 prediction tasks. Furthermore, PAAG demonstrates a nearly sixfold increase in generation success rate (24.7% vs 4.7% in zinc finger, and 54.3% vs 8.7% in the immunoglobulin domain) in comparison to the existing model.

Vision-language foundation models have exhibited remarkable success across a multitude of downstream tasks due to their scalability on extensive image-text paired data. However, these models also display significant limitations when applied to downstream tasks, such as fine-grained image classification, as a result of ``decision shortcuts'' that hinder their generalization capabilities. In this work, we find that the CLIP model possesses a rich set of features, encompassing both \textit{desired invariant causal features} and \textit{undesired decision shortcuts}. Moreover, the underperformance of CLIP on downstream tasks originates from its inability to effectively utilize pre-trained features in accordance with specific task requirements. To address this challenge, we propose a simple yet effective method, Spurious Feature Eraser (SEraser), to alleviate the decision shortcuts by erasing the spurious features. Specifically, we introduce a test-time prompt tuning paradigm that optimizes a learnable prompt, thereby compelling the model to exploit invariant features while disregarding decision shortcuts during the inference phase. The proposed method effectively alleviates excessive dependence on potentially misleading spurious information. We conduct comparative analysis of the proposed method against various approaches which validates the significant superiority.

We present General Time Transformer (GTT), an encoder-only style foundation model for zero-shot multivariate time series forecasting. GTT is pretrained on a large dataset of 200M high-quality time series samples spanning diverse domains. In our proposed framework, the task of multivariate time series forecasting is formulated as a channel-wise next curve shape prediction problem, where each time series sample is represented as a sequence of non-overlapping curve shapes with a unified numerical magnitude. GTT is trained to predict the next curve shape based on a window of past curve shapes in a channel-wise manner. Experimental results demonstrate that GTT exhibits superior zero-shot multivariate forecasting capabilities on unseen time series datasets, even surpassing state-of-the-art supervised baselines. Additionally, we investigate the impact of varying GTT model parameters and training dataset scales, observing that the scaling law also holds in the context of zero-shot multivariate time series forecasting.

In this paper, we construct the rotating multistate Proca stars (RMPSs) composed of two coexisting Proca fields including the ground state and first excited state. We consider the solution families of RMPSs in synchronized frequency and non-synchronized frequency scenarios, respectively. We discuss the characteristics of the matter fields changing with the synchronized frequency or non-synchronized frequency, and then we explore the relationship between the ADM mass $M$ and the frequency. We also analyze the characteristics of ADM mass $M$ as a function of the angular momentum $J$ simultaneously. Furthermore, we calculate the binding energy of RMPSs and analyze the stability of the solutions. Finally, we discuss the ergosphere of RMPSs., Comment: 18 pages, 9 figures

In this paper, we study solutions of a static spherically symmetric system, which is composed of the coupling with the Bardeen action and two Dirac fields. For the case where only the Bardeen action is present, the magnetic charge $q$ can be infinite, then when the magnetic charge is greater than a certain value $q_b$, there exists a black hole solution, which is called the Bardeen black hole (BBH). However, if the Dirac field is introduced, we find that the magnetic charge can only be smaller than the critical value $q_b$, in which there is no black hole solution. Moreover, in the region $q

We develop a symmetric traversable wormhole model, integrating Einstein's gravitational coupling phantom field and a nonlinear electromagnetic field. This work indicates the emergence of negative ADM mass within a specific parameter range, coinciding with distinct alterations in the wormhole's spacetime properties. Despite violating the Null Energy Condition (NEC) and other energy conditions, the solution exhibits unique characteristics in certain energy-momentum tensor components, potentially accounting for the manifestation of negative mass., Comment: 5 pages, 4 figures

Let $\Omega$ be a bounded John domain in $\mathbb R^n$ with $n\ge 2$, and let $\mathcal{H}_{\infty }^{\delta}$ denote the Hausdorff content of dimension $\delta\in (0,n]$. In this article, the authors prove the Poincar\'e and the Poincar\'e--Sobolev inequalities on Choquet--Lorentz integrals with respect to $\mathcal{H}_{\infty }^{\delta}$ for all continuously differentiable functions on $\Omega$. These results not only extend the recent Poincar\'e and Poincar\'e--Sobolev inequalities, obtained by Harjulehto and Hurri-Syrj\"{a}nen in [J. Funct. Anal. 284 (2023), Paper No. 109862], to the Choquet--Lorentz integrals, but also provide some endpoint estimates (weak type) of their results in the critical case. One of the main novelties exists in that, to achieve the goals, the authors develop some new tools associated with Choquet--Lorentz integrals on $\mathcal{H}_{\infty }^{\delta}$, such as the fractional Hardy--Littlewood maximal inequality and the Hedberg-type pointwise estimate on the Riesz potential. As an application, the sharp boundedness of the Riesz potential on Choquet--Lorentz integrals is obtained. Moreover, even for classical Lorentz integrals, these Poincar\'e and Poincar\'e--Sobolev inequalities are also new., Comment: 26 pages

We investigate the tidal deformability of spherically symmetric axion stars on the stable branches, including the Newtonian and relativistic branches. The results suggest that on the stable branch, the electric Love numbers of axion star are positive, while the magnetic Love numbers are negative. On the Newtonian stable branch, the electric tidal Love numbers are much larger than the magnetic ones, while on the relativistic stable branch, they are slightly larger. Furthermore, the relativistic stable branch has much smaller tidal Love numbers than the Newtonian stable branch, indicating weaker deformability of axion stars on the relativistic stable branch. This could be attributed to the fact that on the relativistic branch, axion stars are more compact, resulting hardly distorted by tidal forces., Comment: 22 pages, 5 figures

A multitude of prevalent pre-trained models mark a major milestone in the development of artificial intelligence, while fine-tuning has been a common practice that enables pretrained models to figure prominently in a wide array of target datasets. Our empirical results reveal that off-the-shelf finetuning techniques are far from adequate to mitigate negative transfer caused by two types of underperforming features in a pre-trained model, including rare features and spuriously correlated features. Rooted in structural causal models of predictions after fine-tuning, we propose a Concept-wise fine-tuning (Concept-Tuning) approach which refines feature representations in the level of patches with each patch encoding a concept. Concept-Tuning minimizes the negative impacts of rare features and spuriously correlated features by (1) maximizing the mutual information between examples in the same category with regard to a slice of rare features (a patch) and (2) applying front-door adjustment via attention neural networks in channels and feature slices (patches). The proposed Concept-Tuning consistently and significantly (by up to 4.76%) improves prior state-of-the-art fine-tuning methods on eleven datasets, diverse pre-training strategies (supervised and self-supervised ones), various network architectures, and sample sizes in a target dataset.

In this paper, we construct a Dirac star model composed of $|\kappa|$ pairs of spinor fields. The azimuthal harmonic indeces $m$ of these spinor fields are half-integers, and they satisfiy $-(|\kappa|-\frac{1}{2})\leq m \leq |\kappa|-\frac{1}{2}$. When $\kappa=1$, it corresponds to the conventional Dirac star model, formed by two spinor fields with $m=\frac{1}{2}$ and $m=-\frac{1}{2}$. When $|\kappa|>1$, among these $|\kappa|$ pairs of spinor fields, there exist spinor fields with azimuthal harmonic indices $m>\frac{1}{2}$, and all spinor fields still conform to the same Dirac field equation. Different families of solutions are distinguished by the value of $\kappa$, so we named these solutions $\kappa$-Dirac stars. We obtain solutions for $\kappa=\pm1,\pm2,\pm3,\pm4,\pm5,\pm6$ by using numerical methods. Additionally, we compute their ADM mass $M$, Noether charge $Q$, and binding energy $E$, and illustrate how these quantities change with the spinor field's frequency $\omega$ for different $\kappa$. We observe significant differences between solutions for $|\kappa|>1$ and the $|\kappa|=1$ case. Furthermore, we provide the energy density distribution of the Dirac stars, wherein for $|\kappa|>1$ scenarios, the Dirac stars exhibit a spherical shell-like structure. Moreover, we employ three-dimensional diagrams to intuitively depict how $\kappa$ influences the combination of spinor fields to form a spherically symmetric configuration., Comment: 19 pages, 5 figures. The case of negative value $\kappa$ added, and some discussion expanded

In this paper, we investigate the properties of the first excited state Dirac stars (DSs) with higher azimuthal harmonic index (specifically, the azimuthal harmonic indexes $m_D$ = $3/2$, $5/2$, $7/2$), as well as the relationship between the ADM mass and angular momentum of Dirac stars with respect to frequency. Moreover, We find that the ergospheres of DSs appear at lower spinor field frequencies, and both the ergospheres and the distribution of the spinor field functions are asymmetric about the equatorial plane. Furthermore, we introduce the ground state scalar field and examine its impact on this system, which is known as the multi-state Dirac-boson stars (DBSs) model. We show various types of solution families for DBSs under both synchronized frequency $\omega$ and nonsynchronized frequencies and find that similar to DSs, the spinor field and the ergospheres of DBSs are also asymmetric about the equatorial plane, but the ergospheres appear at higher spinor field frequencies., Comment: 22 pages, 8 figures

In this paper, we study the spherically symmetric Dirac star model in the presence of a phantom field, obtaining a traversable wormhole solution in non-trivial topological spacetime. This solution exhibits asymmetry in both the field configuration and the metric and possesses a finite ADM mass $M$ and Noether charge $Q$. Furthermore, we find that due to the presence of a wormhole at the center, this solution exhibits many differences from the Dirac star under trivial spacetime. Notably, when the wormhole throat size is small, our numerical calculations indicate the emergence of an extremely approximate black hole solution on one side of the wormhole spacetime, a phenomenon unexplored. At this time, the Kretschmann scalar near the throat tends to infinity, indicating the wormhole becomes untraversable., Comment: 32 pages, 23 figures

Atomic nonlinear interferometry has wide applications in quantum metrology and quantum information science. Here we propose a nonlinear time-reversal interferometry scheme with high robustness and metrological gain based on the spin squeezing generated by arbitrary quadratic collective-spin interaction, which could be described by the Lipkin-Meshkov-Glick (LMG) model. We optimize the squeezing process, encoding process, and anti-squeezing process, finding that the two particular cases of the LMG model, one-axis twisting and two-axis twisting outperform in robustness and precision, respectively. Moreover, we propose a Floquet driving method to realize equivalent time reverse in the atomic system, which leads to high performance in precision, robustness, and operability. Our study sets a benchmark in achieving high precision and robustness in atomic nonlinear interferometry.

Graph anomaly detection (GAD) is a vital task since even a few anomalies can pose huge threats to benign users. Recent semi-supervised GAD methods, which can effectively leverage the available labels as prior knowledge, have achieved superior performances than unsupervised methods. In practice, people usually need to identify anomalies on new (sub)graphs to secure their business, but they may lack labels to train an effective detection model. One natural idea is to directly adopt a trained GAD model to the new (sub)graph for testing. However, we find that existing semi-supervised GAD methods suffer from poor generalization issue, i.e., well-trained models could not perform well on an unseen area (i.e., not accessible in training) of the same graph. It may cause great troubles. In this paper, we base on the phenomenon and propose a general and novel research problem of generalized graph anomaly detection that aims to effectively identify anomalies on both the training-domain graph and unseen testing graph to eliminate potential dangers. Nevertheless, it is a challenging task since only limited labels are available, and the normal background may differ between training and testing data. Accordingly, we propose a data augmentation method named \textit{AugAN} (\uline{Aug}mentation for \uline{A}nomaly and \uline{N}ormal distributions) to enrich training data and boost the generalizability of GAD models. Experiments verify the effectiveness of our method in improving model generalizability., Comment: Accepted to IEEE Transactions on Knowledge and Data Engineering (TKDE). arXiv admin note: substantial text overlap with arXiv:2209.10168

Spin squeezing is vitally important in quantum metrology and quantum information science. The noise reduction resulting from spin squeezing can surpass the standard quantum limit and even reach the Heisenberg Limit (HL) in some special circumstances. However, systems that can reach the HL are very limited. Here we study the spin squeezing in atomic systems with a generic form of quadratic collective-spin interaction, which can be described by the Lipkin-Meshkov-Glick(LMG) model. We find that the squeezing properties are determined by the initial states and the anisotropic parameters. Moreover, we propose a pulse rotation scheme to transform the model into two-axis twisting model with Heisenberg-limited spin squeezing. Our study paves the way for reaching HL in a broad variety of systems.

In this work, we provide a complete characterization of the boundedness of two classes of multiparameter Forelli-Rudin type operators from one mixed-norm Lebesgue space $L^{\vec p}$ to another space $L^{\vec q}$, when $1\leq \vec{p}\leq \vec q<\infty$, equipped with possibly different weights. Using these characterizations, we establish the necessary and sufficient conditions for both $L^{\vec p}-L^{\vec q}$ boundedness of the weighted multiparameter Berezin transform and $L^{\vec p}-A^{\vec q}$ boundedness of the weighted multiparameter Bergman projection, where $A^{\vec q}$ denotes the mixed-norm Bergman space. Our approach presents several novelties. Firstly, we conduct refined integral estimates of holomorphic functions on the unit ball in $\mathbb{C}^n$. Secondly, we adapt the classical Schur's test to different weighted mixed-norm Lebesgue spaces. These improvements are crucial in our proofs and allow us to establish the desired characterization and sharp conditions., Comment: 43 pages, to appear in Math. Zeit

Spin squeezing plays a crucial role in quantum metrology and quantum information science. Its generation is the prerequisite for further applications but still faces an enormous challenge since the existing physical systems rarely contain the required squeezing interactions. Here we propose a universal scheme to generate spin squeezing in coupled spin models with collective spin-spin interactions, which commonly exist in various systems. Our scheme can transform the coupled spin interactions into squeezing interactions, and reach the extreme squeezing with Heisenberg-limited measurement precision scaling as $1/N$ for $N$ particles. Only constant and continuous driving fields are required, which is accessible to a series of current realistic experiments. This work greatly enriches the variety of systems that can generate the Heisenberg-limited spin squeezing, with broad applications in quantum precision measurement.

Generation of mechanical squeezing has attracted a lot of interest for its nonclassical properties, applications in quantum information, and high-sensitivity measurement. Here we propose a detuning-switched method that can rapidly generate strong and stationary mechanical squeezing. The pulsed driving can dynamically transpose the optomechanical coupling into a linear optical force and maintain an effective mechanical frequency, which can introduce strong mechanical squeezing in a short time. Moreover, we show the obtained strong mechanical squeezing can be frozen by increasing the pulse intervals, leading to stationary mechanical squeezing with a fixed squeezing angle. Thus, our proposal provides fascinating insights and applications of modulated optomechanical systems.

In this work, the efficient quantum error-correction protocol against the general independent noise is constructed with the three-qubit codes. The rules of concatenation are summarized according to the error-correcting capability of the codes. The codes not only play the role of correcting errors, but the role of polarizing the effective channel. For any independent noise, the most suitable error-correction protocol can be constructed based on the rules of concatenation. The most significant aspect of using the concatenated three-qubit codes is to realize quantum error-correction with the non-ideal gate operations, because the error-correction schemes of the three-qubit codes are simple and modular. For example, for the amplitude damping noise with the initial channel fidelity 0.9, the effective channel fidelity can reach 0.94731 (or 0.961634 with the ideal quantum gate operations) when using the 4 levels concatenated quantum error-correction with the three-qubit codes. The protocol with the three-qubit codes needs 81 qubits and 366 quantum gate operations when the accuracy rate is 0.9995. Meanwhile, when using the 3 levels concatenated quantum error-correction with the five-qubit code, the effective channel fidelity can only reach 0.922798 (or 0.975488 with the ideal quantum gate operations). The protocol with the five-qubit code needs 125 qubits and 1147 quantum gate operations when the accuracy rate is 0.9995. The physical resources costed is multiple of the physical resources costed by realizing the three-qubit quantum error-correction, the protocol has a higher fault tolerance threshold, and no increase in complexity. So, we believe it will be helpful for realizing quantum error-correction in the physical system., Comment: 5 pages, 1 figure, comments are welcome! arXiv admin note: text overlap with arXiv:2110.05130

Graph anomaly detection (GAD) is a vital task since even a few anomalies can pose huge threats to benign users. Recent semi-supervised GAD methods, which can effectively leverage the available labels as prior knowledge, have achieved superior performances than unsupervised methods. In practice, people usually need to identify anomalies on new (sub)graphs to secure their business, but they may lack labels to train an effective detection model. One natural idea is to directly adopt a trained GAD model to the new (sub)graph for testing. However, we find that existing semi-supervised GAD methods suffer from poor generalization issue, i.e., well-trained models could not perform well on an unseen area (i.e., not accessible in training) of the same graph. It may cause great troubles. In this paper, we base on the phenomenon and propose a general and novel research problem of generalized graph anomaly detection that aims to effectively identify anomalies on both the training-domain graph and unseen testing graph to eliminate potential dangers. Nevertheless, it is a challenging task since only limited labels are available, and the normal background may differ between training and testing data. Accordingly, we propose a data augmentation method named \textit{AugAN} (\uline{Aug}mentation for \uline{A}nomaly and \uline{N}ormal distributions) to enrich training data and boost the generalizability of GAD models. Experiments verify the effectiveness of our method in improving model generalizability., Comment: The updated version is accepted by TKDE 2023. Please refer to arXiv:2306.10534v1

Self-supervised pre-training is gaining increasingly more popularity in AI-aided drug discovery, leading to more and more pre-trained models with the promise that they can extract better feature representations for molecules. Yet, the quality of learned representations have not been fully explored. In this work, inspired by the two phenomena of Activity Cliffs (ACs) and Scaffold Hopping (SH) in traditional Quantitative Structure-Activity Relationship (QSAR) analysis, we propose a method named Representation-Property Relationship Analysis (RePRA) to evaluate the quality of the representations extracted by the pre-trained model and visualize the relationship between the representations and properties. The concepts of ACs and SH are generalized from the structure-activity context to the representation-property context, and the underlying principles of RePRA are analyzed theoretically. Two scores are designed to measure the generalized ACs and SH detected by RePRA, and therefore the quality of representations can be evaluated. In experiments, representations of molecules from 10 target tasks generated by 7 pre-trained models are analyzed. The results indicate that the state-of-the-art pre-trained models can overcome some shortcomings of canonical Extended-Connectivity FingerPrints (ECFP), while the correlation between the basis of the representation space and specific molecular substructures are not explicit. Thus, some representations could be even worse than the canonical fingerprints. Our method enables researchers to evaluate the quality of molecular representations generated by their proposed self-supervised pre-trained models. And our findings can guide the community to develop better pre-training techniques to regularize the occurrence of ACs and SH.

The success of meta-learning on existing benchmarks is predicated on the assumption that the distribution of meta-training tasks covers meta-testing tasks. Frequent violation of the assumption in applications with either insufficient tasks or a very narrow meta-training task distribution leads to memorization or learner overfitting. Recent solutions have pursued augmentation of meta-training tasks, while it is still an open question to generate both correct and sufficiently imaginary tasks. In this paper, we seek an approach that up-samples meta-training tasks from the task representation via a task up-sampling network. Besides, the resulting approach named Adversarial Task Up-sampling (ATU) suffices to generate tasks that can maximally contribute to the latest meta-learner by maximizing an adversarial loss. On few-shot sine regression and image classification datasets, we empirically validate the marked improvement of ATU over state-of-the-art task augmentation strategies in the meta-testing performance and also the quality of up-sampled tasks.

This book is devoted to exploring properties of generalized Herz spaces and establishing a complete real-variable theory of Hardy spaces associated with local and global generalized Herz spaces via a totally fresh perspective which means that the authors view these generalized Herz spaces as special cases of ball quasi-Banach function spaces. To be precise, in this book, the authors first study some basic properties of generalized Herz spaces and obtain boundedness and compactness characterizations of commutators on them. Then the authors introduce the associated Herz-Hardy spaces, local Herz-Hardy spaces, and weak Herz-Hardy spaces, and develop a complete real-variable theory of these Herz-Hardy spaces, including their various maximal function, atomic, finite atomic, molecular as well as various Littlewood-Paley function characterizations. As applications, the authors establish the boundedness of some important operators arising from harmonic analysis on these Herz-Hardy spaces. Finally, the inhomogeneous Herz-Hardy spaces and their complete real-variable theory are also investigated. Due to the deficiency of the associate space of the global Herz space, the known real-variable characterizations about Hardy-type spaces associated with ball quasi-Banach function spaces are not applicable to Hardy spaces associated with global generalized Herz spaces which need an improved generalization of the existing one, done by the authors also in this book and having more additional anticipating applications. The authors should also point out that, with the fresh perspective and the improved conclusions on the real-variable theory of Hardy spaces associated with ball quasi-Banach function spaces, the exponents in all the obtained results of this book are sharp. Moreover, all of these results in this book are new and have never been published before., Comment: Submitted

Recently, the pretrain-finetuning paradigm has attracted tons of attention in graph learning community due to its power of alleviating the lack of labels problem in many real-world applications. Current studies use existing techniques, such as weight constraint, representation constraint, which are derived from images or text data, to transfer the invariant knowledge from the pre-train stage to fine-tuning stage. However, these methods failed to preserve invariances from graph structure and Graph Neural Network (GNN) style models. In this paper, we present a novel optimal transport-based fine-tuning framework called GTOT-Tuning, namely, Graph Topology induced Optimal Transport fine-Tuning, for GNN style backbones. GTOT-Tuning is required to utilize the property of graph data to enhance the preservation of representation produced by fine-tuned networks. Toward this goal, we formulate graph local knowledge transfer as an Optimal Transport (OT) problem with a structural prior and construct the GTOT regularizer to constrain the fine-tuned model behaviors. By using the adjacency relationship amongst nodes, the GTOT regularizer achieves node-level optimal transport procedures and reduces redundant transport procedures, resulting in efficient knowledge transfer from the pre-trained models. We evaluate GTOT-Tuning on eight downstream tasks with various GNN backbones and demonstrate that it achieves state-of-the-art fine-tuning performance for GNNs., Comment: 24 pages, 13 figures