Search

The key of the text-to-video retrieval (TVR) task lies in learning the unique similarity between each pair of text (consisting of words) and video (consisting of audio and image frames) representations. However, some problems exist in the representation alignment of video and text, such as a text, and further each word, are of different importance for video frames. Besides, audio usually carries additional or critical information for TVR in the case that frames carry little valid information. Therefore, in TVR task, multi-granularity representation of text, including whole sentence and every word, and the modal of audio are salutary which are underutilized in most existing works. To address this, we propose a novel multi-granularity feature interaction module called MGFI, consisting of text-frame and word-frame, for video-text representations alignment. Moreover, we introduce a cross-modal feature interaction module of audio and text called CMFI to solve the problem of insufficient expression of frames in the video. Experiments on benchmark datasets such as MSR-VTT, MSVD, DiDeMo show that the proposed method outperforms the existing state-of-the-art methods.

We study source-free unsupervised domain adaptation (SFUDA) for semantic segmentation, which aims to adapt a source-trained model to the target domain without accessing the source data. Many works have been proposed to address this challenging problem, among which uncertainty-based self-training is a predominant approach. However, without comprehensive denoising mechanisms, they still largely fall into biased estimates when dealing with different domains and confirmation bias. In this paper, we observe that pseudo-label noise is mainly contained in unstable samples in which the predictions of most pixels undergo significant variations during self-training. Inspired by this, we propose a novel mechanism to denoise unstable samples with stable ones. Specifically, we introduce the Stable Neighbor Denoising (SND) approach, which effectively discovers highly correlated stable and unstable samples by nearest neighbor retrieval and guides the reliable optimization of unstable samples by bi-level learning. Moreover, we compensate for the stable set by object-level object paste, which can further eliminate the bias caused by less learned classes. Our SND enjoys two advantages. First, SND does not require a specific segmentor structure, endowing its universality. Second, SND simultaneously addresses the issues of class, domain, and confirmation biases during adaptation, ensuring its effectiveness. Extensive experiments show that SND consistently outperforms state-of-the-art methods in various SFUDA semantic segmentation settings. In addition, SND can be easily integrated with other approaches, obtaining further improvements., Comment: 2024 Conference on Computer Vision and Pattern Recognition

Audio-driven talking face generation has garnered significant interest within the domain of digital human research. Existing methods are encumbered by intricate model architectures that are intricately dependent on each other, complicating the process of re-editing image or video inputs. In this work, we present ControlTalk, a talking face generation method to control face expression deformation based on driven audio, which can construct the head pose and facial expression including lip motion for both single image or sequential video inputs in a unified manner. By utilizing a pre-trained video synthesis renderer and proposing the lightweight adaptation, ControlTalk achieves precise and naturalistic lip synchronization while enabling quantitative control over mouth opening shape. Our experiments show that our method is superior to state-of-the-art performance on widely used benchmarks, including HDTF and MEAD. The parameterized adaptation demonstrates remarkable generalization capabilities, effectively handling expression deformation across same-ID and cross-ID scenarios, and extending its utility to out-of-domain portraits, regardless of languages.

Some photonic crystals (PCs) with Dirac-like conical dispersions exhibit the property of double zero refractive index (that is, both epsilon and mu near zero (EMNZ)), wherein the electromagnetic waves have an infinite effective wavelength and do not experience any spatial phase change. The Dirac-like cones that support EMNZ are previously thought to present only at the center of the Brillouin zone ($\Gamma$ point) with a zero wavevector (we refer to as type-I EMNZ), which is constrained by the proportional relationship between phase refractive index and wavevector ($n=kc/\omega$). Here, we demonstrate the existence of an anomalous type-II EMNZ in PCs, which is associated with the Dirac-like point at off-$\Gamma$ points. By introducing a wave modulation approach, we theoretically elucidate its physical mechanism, and resolve the paradox of type-II EMNZ with non-zero wavevectors. We then fabricate a type-II EMNZ PC operating at the X point, and experimentally demonstrate that both its effective permittivity and permeability are zero at the Dirac-like point. Type-II EMNZ PCs exhibit a range of intriguing phenomena, including angle-selective transmission, wavefront flattening, a 180$^{\circ}$ phase shift upon transmission, and waveguiding with natural zero radiation loss. The extraordinary properties of type-II EMNZ PCs may open new avenues for the development of angle-selective optical filters, directional light sources, phase-controlled optical switches, ultracompact photonic circuits, nanolasers, and on-chip nonlinear enhancement., Comment: 38 pages, 13 figures

Remote Sensing Image-Text Retrieval (RSITR) is pivotal for knowledge services and data mining in the remote sensing (RS) domain. Considering the multi-scale representations in image content and text vocabulary can enable the models to learn richer representations and enhance retrieval. Current multi-scale RSITR approaches typically align multi-scale fused image features with text features, but overlook aligning image-text pairs at distinct scales separately. This oversight restricts their ability to learn joint representations suitable for effective retrieval. We introduce a novel Multi-Scale Alignment (MSA) method to overcome this limitation. Our method comprises three key innovations: (1) Multi-scale Cross-Modal Alignment Transformer (MSCMAT), which computes cross-attention between single-scale image features and localized text features, integrating global textual context to derive a matching score matrix within a mini-batch, (2) a multi-scale cross-modal semantic alignment loss that enforces semantic alignment across scales, and (3) a cross-scale multi-modal semantic consistency loss that uses the matching matrix from the largest scale to guide alignment at smaller scales. We evaluated our method across multiple datasets, demonstrating its efficacy with various visual backbones and establishing its superiority over existing state-of-the-art methods. The GitHub URL for our project is: https://github.com/yr666666/MSA, Comment: 16 pages, 9 figures

Millimeter wave radar is gaining traction recently as a promising modality for enabling pervasive and privacy-preserving gesture recognition. However, the lack of rich and fine-grained radar datasets hinders progress in developing generalized deep learning models for gesture recognition across various user postures (e.g., standing, sitting), positions, and scenes. To remedy this, we resort to designing a software pipeline that exploits wealthy 2D videos to generate realistic radar data, but it needs to address the challenge of simulating diversified and fine-grained reflection properties of user gestures. To this end, we design G3R with three key components: (i) a gesture reflection point generator expands the arm's skeleton points to form human reflection points; (ii) a signal simulation model simulates the multipath reflection and attenuation of radar signals to output the human intensity map; (iii) an encoder-decoder model combines a sampling module and a fitting module to address the differences in number and distribution of points between generated and real-world radar data for generating realistic radar data. We implement and evaluate G3R using 2D videos from public data sources and self-collected real-world radar data, demonstrating its superiority over other state-of-the-art approaches for gesture recognition., Comment: 18 pages, 29 figures

The discovery of photonic higher-order topological insulators (HOTIs) has significantly expanded our understanding of band topology and provided unprecedented lower-dimensional topological boundary states for robust photonic devices. However, due to the vectorial and leaky nature of electromagnetic waves, it is challenging to discover three-dimensional (3D) topological photonic systems and photonic HOTIs have so far still been limited to two dimensions (2D). Here, we report on the first experimental realization of a 3D Wannier-type photonic HOTI in a tight-binding-like metal-cage photonic crystal, whose band structure matches well with that of a 3D tight-binding model due to the confined Mie resonances. By microwave near-field measurements, we directly observe coexisting topological surface, hinge, and corner states in a single 3D photonic HOTI, as predicted by the tight-binding model and simulation results. Moreover, we demonstrate that all-order topological boundary states are self-guided even in the light cone continuum and can be exposed to air without ancillary cladding, making them well-suited for practical applications. Our work thus opens routes to the multi-dimensional robust manipulation of electromagnetic waves at the outer surfaces of 3D cladding-free photonic bandgap materials and may find novel applications in 3D topological integrated photonics devices., Comment: 23 pages,4 figures

Single atoms trapped in tightly focused optical dipole traps provide an excellent experimental platform for quantum computing, precision measurement, and fundamental physics research. In this work, we propose and demonstrate a novel approach to enhancing the loading of single atoms by introducing a weak ancillary dipole beam. The loading rate of single atoms in a dipole trap can be significantly improved by only a few tens of microwatts of counter-propagating beam. It was also demonstrated that multiple atoms could be loaded with the assistance of a counter-propagating beam. By reducing the power requirements for trapping single atoms and enabling the trapping of multiple atoms, our method facilitates the extension of single-atom arrays and the investigation of collective light-atom interactions., Comment: 10 pages, 7 figures

Topological polaritons characterized by light-matter interactions have become a pivotal platform in exploring new topological phases of matter. Recent theoretical advances unveiled a novel mechanism for tuning topological phases of polaritons by modifying the surrounding photonic environment (light-matter interactions) without altering the lattice structure. Here, by embedding a dimerized chain of microwave helical resonators (electric dipole emitters) in a metallic cavity waveguide, we report the pioneering observation of tunable topological phases of polaritons by varying the cavity width which governs the surrounding photonic environment and the strength of light-matter interactions. Moreover, we experimentally identified a new type of topological phase transition which includes three non-coincident critical points in the parameter space: the closure of the polaritonic bandgap, the transition of the Zak phase, and the hybridization of the topological edge states with the bulk states. These results reveal some remarkable and uncharted properties of topological matter when strongly coupled to light and provide an innovative design principle for tunable topological photonic devices., Comment: 6 pages, 4 figures

Presently, self-training stands as a prevailing approach in cross-domain semantic segmentation, enhancing model efficacy by training with pixels assigned with reliable pseudo-labels. However, we find two critical limitations in this paradigm. (1) The majority of reliable pixels exhibit a speckle-shaped pattern and are primarily located in the central semantic region. This presents challenges for the model in accurately learning semantics. (2) Category noise in speckle pixels is difficult to locate and correct, leading to error accumulation in self-training. To address these limitations, we propose a novel approach called Semantic Connectivity-driven pseudo-labeling (SeCo). This approach formulates pseudo-labels at the connectivity level and thus can facilitate learning structured and low-noise semantics. Specifically, SeCo comprises two key components: Pixel Semantic Aggregation (PSA) and Semantic Connectivity Correction (SCC). Initially, PSA divides semantics into 'stuff' and 'things' categories and aggregates speckled pseudo-labels into semantic connectivity through efficient interaction with the Segment Anything Model (SAM). This enables us not only to obtain accurate boundaries but also simplifies noise localization. Subsequently, SCC introduces a simple connectivity classification task, which enables locating and correcting connectivity noise with the guidance of loss distribution. Extensive experiments demonstrate that SeCo can be flexibly applied to various cross-domain semantic segmentation tasks, including traditional unsupervised, source-free, and black-box domain adaptation, significantly improving the performance of existing state-of-the-art methods. The code is available at https://github.com/DZhaoXd/SeCo.

There is a significant relevance of federated learning (FL) in the realm of Artificial Intelligence of Things (AIoT). However, most existing FL works do not use datasets collected from authentic IoT devices and thus do not capture unique modalities and inherent challenges of IoT data. To fill this critical gap, in this work, we introduce FedAIoT, an FL benchmark for AIoT. FedAIoT includes eight datasets collected from a wide range of IoT devices. These datasets cover unique IoT modalities and target representative applications of AIoT. FedAIoT also includes a unified end-to-end FL framework for AIoT that simplifies benchmarking the performance of the datasets. Our benchmark results shed light on the opportunities and challenges of FL for AIoT. We hope FedAIoT could serve as an invaluable resource to foster advancements in the important field of FL for AIoT. The repository of FedAIoT is maintained at https://github.com/AIoT-MLSys-Lab/FedAIoT., Comment: Camera-ready version of the Journal of Data-centric Machine Learning Research (DMLR)

A unified solution to adaptive approximation-based control for nonlinear systems with accurate and inaccurate state measurement is synthesized in this study. Starting from the standard adaptive approximation-based controller with accurate state measurement, its corresponding physical interpretation, stability conclusion, and learning ability are rigorously addressed when facing additive measurement inaccuracy, and explicit answers are obtained in the framework of both controller matching and system matching. Finally, it proves that, with a certain condition, the standard adaptive approximation-based controller works as a unified solution for the cases with accurate and inaccurate measurement, and the solution can be extended to the nonlinear system control problems with extra unknown dynamics or faults in actuator and/or process dynamics. A single-link robot arm example is used for the simulation demonstration of the unified solution.

The replay attack detection problem is studied from a new perspective based on parity space method in this paper. The proposed detection methods have the ability to distinguish system fault and replay attack, handle both input and output data replay, maintain certain control performance, and can be implemented conveniently and efficiently. First, the replay attack effect on the residual is derived and analyzed. The residual change induced by replay attack is characterized explicitly and the detection performance analysis based on two different test statistics are given. Second, based on the replay attack effect characterization, targeted passive and active design for detection performance enhancement are proposed. Regarding the passive design, four optimization schemes regarding different cost functions are proposed with optimal parity matrix solutions, and the unified solution to the passive optimization schemes is obtained; the active design is enabled by a marginally stable filter so as to enlarge the replay attack effect on the residual for detection. Simulations and comparison studies are given to show the effectiveness of the proposed methods.