arXiv subjects

Peng Wang

Publications and source records attributed to Peng Wang.

At least 649 records · Page 36Linked to original sources

arXiv2022

Attacking Digital Twins of Robotic Systems to Compromise Security and Safety

Security and safety are of paramount importance to human-robot interaction, either for autonomous robots or human-robot collaborative manufacturing. The intertwined relationship between security and safety has imposed new challenges on the emerging digital twin systems of various types of robots. To be specific, the attack of either the cyber-physical system or the digital-twin system could cause severe consequences to the other. Particularly, the attack of a digital-twin system that is synchronized with a cyber-physical system could cause lateral damage to humans and other surrounding facilities. This paper demonstrates that for Robot Operating System (ROS) driven systems, attacks such as the person-in-the-middle attack of the digital-twin system could eventually lead to a collapse of the cyber-physical system, whether it is an industrial robot or an autonomous mobile robot, causing unexpected consequences. We also discuss potential solutions to alleviate such attacks.

cs.RO↗

arXiv2022

A Human-friendly Verbal Communication Platform for Multi-Robot Systems: Design and Principles

While multi-robot systems have been broadly researched and deployed, their success is built chiefly upon the dependency on network infrastructures, whether wired or wireless. Aiming at the first steps toward de-coupling the application of multi-robot systems from the reliance on network infrastructures, this paper proposes a human-friendly verbal communication platform for multi-robot systems, following the deliberately designed principles of being adaptable, transparent, and secure. The platform is network independent and is subsequently capable of functioning in network infrastructure lacking environments from underwater to planet explorations. A series of experiments were conducted to demonstrate the platform's capability in multi-robot systems communication and task coordination, showing its potential in infrastructure-free applications. To benefit the community, we have made the codes open source at https://github.com/jynxmagic/MSc_AI_project

cs.RO↗

arXiv2022

Levenshtein OCR

A novel scene text recognizer based on Vision-Language Transformer (VLT) is presented. Inspired by Levenshtein Transformer in the area of NLP, the proposed method (named Levenshtein OCR, and LevOCR for short) explores an alternative way for automatically transcribing textual content from cropped natural images. Specifically, we cast the problem of scene text recognition as an iterative sequence refinement process. The initial prediction sequence produced by a pure vision model is encoded and fed into a cross-modal transformer to interact and fuse with the visual features, to progressively approximate the ground truth. The refinement process is accomplished via two basic character-level operations: deletion and insertion, which are learned with imitation learning and allow for parallel decoding, dynamic length change and good interpretability. The quantitative experiments clearly demonstrate that LevOCR achieves state-of-the-art performances on standard benchmarks and the qualitative analyses verify the effectiveness and advantage of the proposed LevOCR algorithm. Code is available at https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/OCR/LevOCR.

cs.CV↗

arXiv2022

Two-dimensional Thouless pumping of light in photonic moiré lattices

Continuous and quantized transports are profoundly different. The latter is determined by the global rather than local properties of a system, it exhibits unique topological features, and its ubiquitous nature causes its occurrence in many areas of science. Here we report the first observation of fully-two-dimensional Thouless pumping of light by bulk modes in a purpose-designed tilted moiré lattices imprinted in a photorefractive crystal. Pumping in such unconfined system occurs due to the longitudinal adiabatic and periodic modulation of the refractive index. The topological nature of this phenomenon manifests itself in the magnitude and direction of shift of the beam center-of-mass averaged over one pumping cycle. \rev{Our experimental results are supported by systematic numerical simulations in the frames of the continuous Schrödinger equation governing propagation of probe light beam in optically-induced photorefractive moiré lattice.} Our system affords a powerful platform for the exploration of topological pumping in tunable commensurate and incommensurate geometries.

physics.optics↗

arXiv2022

Event-triggered Finite-time Control Using Inverse-optimal Implicit Lyapunov Function

This work deals with the event-triggered finite-time control for high-order systems based on an implicit Lyapunov function (ILF). With the construction of an inverse optimal problem, a novel expression of ILF is obtained. By designing the event-triggering mechanism elaborately, it is guaranteed that the trivial solution of the closed-loop system is globally finite-time stable and there exists no Zeno phenomenon. Extensions to the scenario with a multi-agent system are studied where a finite-time tracking control drives all the agents to reach a consensus. The obtained theoretical results are supported by numerical simulations.

math.OC↗

arXiv2022

One for All: One-stage Referring Expression Comprehension with Dynamic Reasoning

Referring Expression Comprehension (REC) is one of the most important tasks in visual reasoning that requires a model to detect the target object referred by a natural language expression. Among the proposed pipelines, the one-stage Referring Expression Comprehension (OSREC) has become the dominant trend since it merges the region proposal and selection stages. Many state-of-the-art OSREC models adopt a multi-hop reasoning strategy because a sequence of objects is frequently mentioned in a single expression which needs multi-hop reasoning to analyze the semantic relation. However, one unsolved issue of these models is that the number of reasoning steps needs to be pre-defined and fixed before inference, ignoring the varying complexity of expressions. In this paper, we propose a Dynamic Multi-step Reasoning Network, which allows the reasoning steps to be dynamically adjusted based on the reasoning state and expression complexity. Specifically, we adopt a Transformer module to memorize & process the reasoning state and a Reinforcement Learning strategy to dynamically infer the reasoning steps. The work achieves the state-of-the-art performance or significant improvements on several REC datasets, ranging from RefCOCO (+, g) with short expressions, to Ref-Reasoning, a dataset with long and complex compositional expressions.

cs.CV↗

arXiv2022

A Linearly Convergent Algorithm for Rotationally Invariant $\ell_1$-Norm Principal Component Analysis

To do dimensionality reduction on the datasets with outliers, the $\ell_1$-norm principal component analysis (L1-PCA) as a typical robust alternative of the conventional PCA has enjoyed great popularity over the past years. In this work, we consider a rotationally invariant L1-PCA, which is hardly studied in the literature. To tackle it, we propose a proximal alternating linearized minimization method with a nonlinear extrapolation for solving its two-block reformulation. Moreover, we show that the proposed method converges at least linearly to a limiting critical point of the reformulated problem. Such a point is proved to be a critical point of the original problem under a condition imposed on the step size. Finally, we conduct numerical experiments on both synthetic and real datasets to support our theoretical developments and demonstrate the efficacy of our approach.

math.OC↗

arXiv2022

Observational Appearance of a Freely-falling Star in an Asymmetric Thin-shell Wormhole

It has been recently reported that, at late times, the total luminosity of a star freely falling in black holes decays exponentially with time, and one or two series of flashes with decreasing intensity are seen by a specific observer, depending on the number of photon spheres. In this paper, we examine observational appearances of an infalling star in a reflection-asymmetric wormhole, which has two photon spheres, one on each side of the wormhole. We find that the late-time total luminosity measured by distant observers gradually decays with time or remains roughly constant due to the absence of the event horizon. Moreover, a specific observer would detect a couple of light flashes in a bright background at late times. These observations would offer a new tool to distinguish wormholes from black holes, even those with multiple photon spheres.

gr-qc↗

arXiv2022

Multi-Granularity Prediction for Scene Text Recognition

Scene text recognition (STR) has been an active research topic in computer vision for years. To tackle this challenging problem, numerous innovative methods have been successively proposed and incorporating linguistic knowledge into STR models has recently become a prominent trend. In this work, we first draw inspiration from the recent progress in Vision Transformer (ViT) to construct a conceptually simple yet powerful vision STR model, which is built upon ViT and outperforms previous state-of-the-art models for scene text recognition, including both pure vision models and language-augmented methods. To integrate linguistic knowledge, we further propose a Multi-Granularity Prediction strategy to inject information from the language modality into the model in an implicit way, i.e. , subword representations (BPE and WordPiece) widely-used in NLP are introduced into the output space, in addition to the conventional character level representation, while no independent language model (LM) is adopted. The resultant algorithm (termed MGP-STR) is able to push the performance envelop of STR to an even higher level. Specifically, it achieves an average recognition accuracy of 93.35% on standard benchmarks. Code is available at https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/OCR/MGP-STR.

cs.CV↗

arXiv2022

NeuRIS: Neural Reconstruction of Indoor Scenes Using Normal Priors

Reconstructing 3D indoor scenes from 2D images is an important task in many computer vision and graphics applications. A main challenge in this task is that large texture-less areas in typical indoor scenes make existing methods struggle to produce satisfactory reconstruction results. We propose a new method, named NeuRIS, for high quality reconstruction of indoor scenes. The key idea of NeuRIS is to integrate estimated normal of indoor scenes as a prior in a neural rendering framework for reconstructing large texture-less shapes and, importantly, to do this in an adaptive manner to also enable the reconstruction of irregular shapes with fine details. Specifically, we evaluate the faithfulness of the normal priors on-the-fly by checking the multi-view consistency of reconstruction during the optimization process. Only the normal priors accepted as faithful will be utilized for 3D reconstruction, which typically happens in the regions of smooth shapes possibly with weak texture. However, for those regions with small objects or thin structures, for which the normal priors are usually unreliable, we will only rely on visual features of the input images, since such regions typically contain relatively rich visual features (e.g., shade changes and boundary contours). Extensive experiments show that NeuRIS significantly outperforms the state-of-the-art methods in terms of reconstruction quality.

cs.CV↗

arXiv2022

Möbius Homogeneous Hypersurfaces in $\mathbb{S}^{n+1}$

Let $\mathbb{M}(\mathbb{S}^{n+1})$ denote the Möbius transformation group of the $(n+1)$-dimensional sphere $\mathbb{S}^{n+1}$. A hypersurface $x:M^n\to \mathbb{S}^{n+1}$ is called a Möbius homogeneous hypersurface if there exists a subgroup $G$ of $\mathbb{M}(\mathbb{S}^{n+1})$ such that the orbit $G\cdot p=x(M^n), p\in x(M^n)$. In this paper, the Möbius homogeneous hypersurfaces are classified completely up to a Möbius transformation of $\mathbb{S}^{n+1}$.

math.DG↗

arXiv2022

Adaptive Graph Convolutional Networks for Weakly Supervised Anomaly Detection in Videos

For weakly supervised anomaly detection, most existing work is limited to the problem of inadequate video representation due to the inability of modeling long-term contextual information. To solve this, we propose a novel weakly supervised adaptive graph convolutional network (WAGCN) to model the complex contextual relationship among video segments. By which, we fully consider the influence of other video segments on the current one when generating the anomaly probability score for each segment. Firstly, we combine the temporal consistency as well as feature similarity of video segments to construct a global graph, which makes full use of the association information among spatial-temporal features of anomalous events in videos. Secondly, we propose a graph learning layer in order to break the limitation of setting topology manually, which can extract graph adjacency matrix based on data adaptively and effectively. Extensive experiments on two public datasets (i.e., UCF-Crime dataset and ShanghaiTech dataset) demonstrate the effectiveness of our approach which achieves state-of-the-art performance.

cs.CV↗

arXiv2022

Two-dimensional nonlinear Thouless pumping of matter waves

We consider theoretically the nonlinear quantized Thouless pumping of a Bose-Einstein condensate loaded in a two-dimensional dynamical optical lattices. We encountered three different scenarios of the pumping: quasi-linear one occurring for gradually dispersing wave packets, transport carried by a single two-dimensional soliton, and multi-soliton regime when initial wave packet splits into several solitons. The scenario to be realized depends on the number of atoms in the initial wave packet and on the strength of the two-body interactions. The magnitude and direction of the displacement of a wavepacket are determined by Chern numbers of the populated energy bands and by the inter-band transitions induced by two-body interactions. As a case example we explore a separable potential created by optical lattices whose constitutive sublattices undergo relative motion in the orthogonal directions. For such potentials, obeying parity-time symmetry, fractional Chern numbers, computed over half period of the evolution, acquire relevance. We focus mainly on solitonic scenarios, showing that one-soliton pumping occurs at relatively small as well as at sufficiently large amplitudes of the initial wavepacket, while at intermediate amplitudes the transport is multi-solitonic. We also describe peculiarities of the pumping characterized by two different commensurate periods of the modulations of the lattices in the orthogonal directions.

nlin.PS↗

arXiv2022

Phase transitions and thermodynamic geometry of a Kerr-Newman black hole in a cavity

Being placed in a cavity is an effective way of reaching thermodynamic equilibrium for black holes. We investigate a Kerr-Newman black hole in a cavity as well as compare it with two reduced cases, i.e., a RN black hole in a cavity and a Kerr black hole in a cavity. We derive the quasi-local energy from the Hamiltonian, and construct the first law of thermodynamics accordingly. In a canonical ensemble, these black holes could undergo a van der Waals-like phase transition, which is very similar to that in AdS space. We further investigate the black holes' thermodynamic geometry, which is a powerful tool to diagnose microscopic interactions of a thermodynamic system. Our results show that in a cavity, although phase structures of these black holes are similar, their thermodynamic geometry show strong dissimilarities, implying that the microstructure of a black hole is sensitive to its states.

gr-qc↗

arXiv2022

Anisotropic satellite accretion onto the Local Group with HESTIA

How the cosmic web feeds halos, and fuels galaxy formation is an open question with wide implications. This study explores the mass assembly in the Local Group within the context of the local cosmography by employing simulations whose initial conditions have been constrained to reproduce the local environment. The goal of this study is to inspect whether the direction of accretion of satellites on to the Milky Way and Andromeda galaxies, is related to the cosmic web. The analysis considers the three high-resolution simulations available in the HESTIA simulation suite, as well as the derived velocity shear and tidal tensors. We notice two eras in the Local Group accretion history, delimited by an epoch around $z \approx 0.7$. We also find that satellites can travel up to $\sim 4$ Mpc, relative to their parent halo before crossing its viral radius $R_{200}$. Finally, we observe a strong alignment of the infall direction with the axis of slowest collapse $\vec{e_3}$ of both tidal and shear tensors, implying satellites of the Local Group originated from one particular region of the cosmic web and were channeled towards us via the process of accretion.This alignment is dominated by the satellites that enter during the early infall era, i.e $z>0.7$.

astro-ph.GA↗

arXiv2022

Instance Image Retrieval by Learning Purely From Within the Dataset

Quality feature representation is key to instance image retrieval. To attain it, existing methods usually resort to a deep model pre-trained on benchmark datasets or even fine-tune the model with a task-dependent labelled auxiliary dataset. Although achieving promising results, this approach is restricted by two issues: 1) the domain gap between benchmark datasets and the dataset of a given retrieval task; 2) the required auxiliary dataset cannot be readily obtained. In light of this situation, this work looks into a different approach which has not been well investigated for instance image retrieval previously: {can we learn feature representation \textit{specific to} a given retrieval task in order to achieve excellent retrieval?} Our finding is encouraging. By adding an object proposal generator to generate image regions for self-supervised learning, the investigated approach can successfully learn feature representation specific to a given dataset for retrieval. This representation can be made even more effective by boosting it with image similarity information mined from the dataset. As experimentally validated, such a simple ``self-supervised learning + self-boosting'' approach can well compete with the relevant state-of-the-art retrieval methods. Ablation study is conducted to show the appealing properties of this approach and its limitation on generalisation across datasets.

cs.CV↗

arXiv2022

Identical Photons from Multiple Tin-Vacancy Centers in Diamond

We report the narrow inhomogeneous distribution of the zero-phonon line from tin-vacancy (SnV) centers in diamond and the overlap of spectra from multiple separated SnV centers. Photoluminescence excitation spectroscopy measurements at a cryogenic temperature showed that SnV centers exhibit stable fluorescence and linewidths close to the Fourier transform-limited linewidth. The inhomogeneous distribution was as low as ~4 GHz, which enabled the observation of Sn isotope-dependent resonant frequencies. Owing to the narrow inhomogeneous distribution, we observed multiple SnV centers showing identical photons with almost the same wavelength and linewidth. Identical SnV centers were also observed even in different diamond samples, confirming the reliable fabrication of the high-quality SnV centers.

quant-ph↗

arXiv2022

Prompt Tuning for Generative Multimodal Pretrained Models

Prompt tuning has become a new paradigm for model tuning and it has demonstrated success in natural language pretraining and even vision pretraining. In this work, we explore the transfer of prompt tuning to multimodal pretraining, with a focus on generative multimodal pretrained models, instead of contrastive ones. Specifically, we implement prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks. Experimental results demonstrate that the light-weight prompt tuning can achieve comparable performance with finetuning and surpass other light-weight tuning methods. Besides, in comparison with finetuned models, the prompt-tuned models demonstrate improved robustness against adversarial attacks. We further figure out that experimental factors, including the prompt length, prompt depth, and reparameteratization, have great impacts on the model performance, and thus we empirically provide a recommendation for the setups of prompt tuning. Despite the observed advantages, we still find some limitations in prompt tuning, and we correspondingly point out the directions for future studies. Codes are available at \url{https://github.com/OFA-Sys/OFA}

cs.CL↗