arXiv subjects

Yutao Tang

Publications and source records attributed to Yutao Tang.

At least 19 recordsLinked to original sources

arXiv2025

Scenes as Tokens: Multi-Scale Normal Distributions Transform Tokenizer for General 3D Vision-Language Understanding

Recent advances in 3D vision-language models (VLMs) highlight a strong potential for 3D scene understanding and reasoning. However, effectively tokenizing 3D scenes into holistic scene tokens, and leveraging these tokens across diverse 3D understanding tasks, remain highly challenging. We present NDTokenizer3D, a generalist 3D VLM that performs a wide range of 3D scene understanding tasks while naturally supporting human interactions, thereby bridging language-level reasoning with 3D spatial understanding. The core of our approach is a novel three-stage scene tokenization pipeline built upon a Multi-Scale Normal Distributions Transform (NDT) representation, paired with a Multi-Scale NDT Decoder (MSDec). Specifically, NDTokenizer3D first constructs a multi-scale NDT representation from raw high-resolution point clouds, preserving both global context and fine-grained geometric details. Next, the MSDec progressively fuses cross-scale NDT features, producing holistic scene tokens consumable by LLM endpoints. Beyond tokenization, MSDec is repurposed as a general interface for human-interactive prompting (points, boxes, masks) and segmentation-mask decoding, unifying diverse 3D scene understanding tasks within a single architecture. With this compact and unified design, NDTokenizer3D offers a fine-grained, general-purpose 3D VLM, achieving remarkable improvements in 3D Referring Segmentation, 3D Visual Question Answering, and 3D Dense Captioning.

cs.CV↗

arXiv2025

Dynamic Nash Equilibrium Seeking for a Class of Nonlinear Uncertain Multi-agent Systems

We consider seeking a Nash equilibrium (NE) of a monotone game, played by dynamic agents which are modeled as a class of lower-triangular nonlinear uncertain dynamics with external disturbances. We establish a general framework that converts the problem into a distributed robust stabilization problem of an appropriately augmented system. To be specific, we construct a virtual single-integrator multi-agent system, as a reference signal generator, to compute an NE in a fully distributed manner. By introducing internal models to tackle the disturbances, as well as embedding the virtual system, we derive an augmented system. Following that, we show that the outputs of all agents reach an NE of the game if the augmented system can be stabilized by a control law. Finally, resorting to a backstepping procedure, we design a distributed state-feedback controller to stabilize the augmented system semi-globally.

math.OC↗

arXiv2025

MS-GS: Multi-Appearance Sparse-View 3D Gaussian Splatting in the Wild

In-the-wild photo collections often contain limited volumes of imagery and exhibit multiple appearances, e.g., taken at different times of day or seasons, posing significant challenges to scene reconstruction and novel view synthesis. Although recent adaptations of Neural Radiance Field (NeRF) and 3D Gaussian Splatting (3DGS) have improved in these areas, they tend to oversmooth and are prone to overfitting. In this paper, we present MS-GS, a novel framework designed with Multi-appearance capabilities in Sparse-view scenarios using 3DGS. To address the lack of support due to sparse initializations, our approach is built on the geometric priors elicited from monocular depth estimations. The key lies in extracting and utilizing local semantic regions with a Structure-from-Motion (SfM) points anchored algorithm for reliable alignment and geometry cues. Then, to introduce multi-view constraints, we propose a series of geometry-guided supervision steps at virtual views in pixel and feature levels to encourage 3D consistency and reduce overfitting. We also introduce a dataset and an in-the-wild experiment setting to set up more realistic benchmarks. We demonstrate that MS-GS achieves photorealistic renderings under various challenging sparse-view and multi-appearance conditions, and outperforms existing approaches significantly across different datasets.

cs.CV↗

arXiv2024

SPARS3R: Semantic Prior Alignment and Regularization for Sparse 3D Reconstruction

Recent efforts in Gaussian-Splat-based Novel View Synthesis can achieve photorealistic rendering; however, such capability is limited in sparse-view scenarios due to sparse initialization and over-fitting floaters. Recent progress in depth estimation and alignment can provide dense point cloud with few views; however, the resulting pose accuracy is suboptimal. In this work, we present SPARS3R, which combines the advantages of accurate pose estimation from Structure-from-Motion and dense point cloud from depth estimation. To this end, SPARS3R first performs a Global Fusion Alignment process that maps a prior dense point cloud to a sparse point cloud from Structure-from-Motion based on triangulated correspondences. RANSAC is applied during this process to distinguish inliers and outliers. SPARS3R then performs a second, Semantic Outlier Alignment step, which extracts semantically coherent regions around the outliers and performs local alignment in these regions. Along with several improvements in the evaluation process, we demonstrate that SPARS3R can achieve photorealistic rendering with sparse images and significantly outperforms existing approaches.

cs.CV↗

arXiv2024

Reversible optical isolators and quasi-circulators using a magneto-optical Fabry-Pérot cavity

Nonreciprocal optical devices are essential for laser protection, modern optical communication and quantum information processing by enforcing one-way light propagation. The conventional Faraday magneto-optical nonreciprocal devices rely on a strong magnetic field, which is provided by a permanent magnet. As a result, the isolation direction of such devices is fixed and severely restricts their applications in quantum networks.In this work, we experimentally demonstrate the simultaneous one-way transmission and unidirectional reflection by using a magneto-optical Fabry-Pérot cavity and a magnetic field strength of $50~\milli\tesla$. An optical isolator and a three-port quasi-circulator are realized based on this nonreciprocal cavity system. The isolator achieves an isolation ratio of up to $22~\deci\bel$ and an averaged insertion loss down to $0.97~\deci\bel$. The quasi-circulator is realized with a fidelity exceeding $99\%$ and an overall survival probability of $89.9\%$, corresponding to an insertion loss of $\sim 0.46~\deci\bel$. The magnetic field is provided by an electromagnetic coil, thereby allowing for reversing the light circulating path. The reversible quasi-circulator paves the way for building reconfigurable quantum networks.

physics.optics↗

arXiv2024

BAGS: Blur Agnostic Gaussian Splatting through Multi-Scale Kernel Modeling

Recent efforts in using 3D Gaussians for scene reconstruction and novel view synthesis can achieve impressive results on curated benchmarks; however, images captured in real life are often blurry. In this work, we analyze the robustness of Gaussian-Splatting-based methods against various image blur, such as motion blur, defocus blur, downscaling blur, \etc. Under these degradations, Gaussian-Splatting-based methods tend to overfit and produce worse results than Neural-Radiance-Field-based methods. To address this issue, we propose Blur Agnostic Gaussian Splatting (BAGS). BAGS introduces additional 2D modeling capacities such that a 3D-consistent and high quality scene can be reconstructed despite image-wise blur. Specifically, we model blur by estimating per-pixel convolution kernels from a Blur Proposal Network (BPN). BPN is designed to consider spatial, color, and depth variations of the scene to maximize modeling capacity. Additionally, BPN also proposes a quality-assessing mask, which indicates regions where blur occur. Finally, we introduce a coarse-to-fine kernel optimization scheme; this optimization scheme is fast and avoids sub-optimal solutions due to a sparse point cloud initialization, which often occurs when we apply Structure-from-Motion on blurry images. We demonstrate that BAGS achieves photorealistic renderings under various challenging blur conditions and imaging geometry, while significantly improving upon existing approaches.

cs.CV↗

arXiv2023

Semantic-aware Video Representation for Few-shot Action Recognition

Recent work on action recognition leverages 3D features and textual information to achieve state-of-the-art performance. However, most of the current few-shot action recognition methods still rely on 2D frame-level representations, often require additional components to model temporal relations, and employ complex distance functions to achieve accurate alignment of these representations. In addition, existing methods struggle to effectively integrate textual semantics, some resorting to concatenation or addition of textual and visual features, and some using text merely as an additional supervision without truly achieving feature fusion and information transfer from different modalities. In this work, we propose a simple yet effective Semantic-Aware Few-Shot Action Recognition (SAFSAR) model to address these issues. We show that directly leveraging a 3D feature extractor combined with an effective feature-fusion scheme, and a simple cosine similarity for classification can yield better performance without the need of extra components for temporal modeling or complex distance functions. We introduce an innovative scheme to encode the textual semantics into the video representation which adaptively fuses features from text and video, and encourages the visual encoder to extract more semantically consistent features. In this scheme, SAFSAR achieves alignment and fusion in a compact way. Experiments on five challenging few-shot action recognition benchmarks under various settings demonstrate that the proposed SAFSAR model significantly improves the state-of-the-art performance.

cs.CV↗

arXiv2023

Observer-based Leader-following Consensus for Positive Multi-agent Systems Over Time-varying Graphs

This paper addresses the leader-following consensus problem for discrete-time positive multi-agent systems over time-varying graphs. We assume that the followers may have mutually different positive dynamics which can also be different from the leader. Compared with most existing positive consensus works for homogeneous multi-agent systems, the formulated problem is more general and challenging due to the interplay between the positivity requirement and high-order heterogeneous dynamics. To solve the problem, we present an extended version of existing observer-based design for positive multi-agent systems. By virtue of the common quadratic Lyapunov function technique, we show the followers will maintain their state variables in the positive orthant and finally achieve an output consensus specified by the leader. A numerical example is used to verify the efficacy of our algorithms.

eess.SY↗

arXiv2023

Robust Positive Consensus for Heterogeneous Multi-agent Systems

This paper investigates a robust positive consensus problem for a class of heterogeneous high-order multi-agent systems subject to external inputs. Compared with existing multi-agent consensus results, the most distinct feature of the formulated problem is that the state variables of all heterogeneous agents are confined in the positive orthant. To solve this problem, we present a two-step design procedure. By constructing an auxiliary multi-agent system as positive local reference generators, we incorporate the reference generator into some applicable decentralized robust tracking controller for each agent. The proposed distributed algorithm is proven to ensure a robust consensus fulfilling certain prescribed pattern for the multi-agent system under switching topology in the sense of finite-gain stability with respect to the external inputs. A simulation example is finally given to illustrate the effectiveness of our design.

eess.SY↗

arXiv2023

On Measurement Disturbances in Distributed Least Squares Solvers for Linear Equations

This paper aims at distributed algorithms for solving a system of linear algebraic equations. Different from most existing formulations for this problem, we assume that the local data at each node is not accurately measured but subject to some disturbances. To be specific, the local measurement consists of two parts: a nominal value and a multiple sinusoidal disturbance. By introducing an identifier-enhanced observer to estimate the disturbance, we present a novel distributed least squares solver for the linear equations using noisy measurements. The proposed solver is proven to be able to recover the least squares solution to the linear equations associated with the nominal values irrespective of any multi-sinusoidal disturbance even with unknown frequencies. We also show the robustness of the distributed solvers under standard conditions against unstructured perturbations. The effectiveness of our design is verified by a numerical example.

math.OC↗

arXiv2022

Event-triggered Design for Optimal Output Consensus of High-order Multi-agent Systems

This paper studies the optimal output consensus problem for a group of heterogeneous linear multi-agent systems. Different from existing results, we aim at effective controllers for these high-order agents under both event-triggered control and event-triggered communication settings. We conduct an embedded design for the problem and constructively propose a multi-rate event-triggered controller with a set of applicable parameters. The proposed event-triggered rules are shown to be free of Zeno behaviors and can achieve the optimal output consensus goal for these high-order agents. A simulation example is given to verify the efficacy of our designs.

eess.SY↗

arXiv2022

Resonant chiral effects in nonlinear dielectric metasurfaces

We study the resonant enhancement of linear and nonlinear chiroptical effects in asymmetric silicon metasurfaces supporting multipolar Mie resonances and quasi-bound states in the continuum (quasi-BICs). We demonstrate theoretically and observe in experiment the pronounced linear circular dichroism at the quasi-BIC resonances. We further find that both local field enhancement and third-harmonic signal are large for Mie resonances and some quasi-BIC modes. We explain the selectivity of the nonlinear enhancement by employing the concept of critical coupling being more favorable for the modes with moderately large radiative quality factors ($Q$ factors). We demonstrate experimentally strong nonlinear chiroptical response associated with high efficiency of the third-harmonic generation and large nonlinear circular dichroism varying from $+0.918\pm0.049$ to $-0.771\pm0.004$ for the samples with different asymmetries. We believe our results suggest a general strategy for engineering nonlinear chiroptical response in dielectric resonant metasurfaces.

physics.optics↗

arXiv2022

Facial Tic Detection in Untrimmed Videos of Tourette Syndrome Patients

Tourette Syndrome (TS) is a behavior disorder that onsets in childhood and is characterized by the expression of involuntary movements and sounds commonly referred to as tics. Behavioral therapy is the first-line treatment for patients with TS, and it helps patients raise awareness about tic occurrence as well as develop tic inhibition strategies. However, the limited availability of therapists and the difficulties for in-home follow up work limits its effectiveness. An automatic tic detection system that is easy to deploy could alleviate the difficulties of home-therapy by providing feedback to the patients while exercising tic awareness. In this work, we propose a novel architecture (T-Net) for automatic tic detection and classification from untrimmed videos. T-Net combines temporal detection and segmentation and operates on features that are interpretable to a clinician. We compare T-Net to several state-of-the-art systems working on deep features extracted from the raw videos and T-Net achieves comparable performance in terms of average precision while relying on interpretable features needed in clinical practice.

cs.CV↗

arXiv2022

Distributed Optimization with Inexact Oracle

In this paper, we study the distributed optimization problem using approximate first-order information. We suppose the agent can repeatedly call an inexact first-order oracle of each individual objective function and exchange information with its time-varying neighbors. We revisit the distributed subgradient method in this circumstance and show its suboptimality under square summable but not summable step sizes. We also present several conditions on the inexactness of the local oracles to ensure an exact convergence of the iterative sequences towards the global optimal solution. A numerical example is given to verify the efficiency of our algorithm.

math.OC↗

arXiv2022

Neural Network-based Constrained Optimal Coordination for Heterogeneous Uncertain Nonlinear Multi-agent Systems

In this paper, we investigate a constrained optimal coordination problem for a class of heterogeneous nonlinear multi-agent systems described by high-order dynamics subject to both unknown nonlinearities and external disturbances. Each agent has a private objective function and a steady-state constraint about its output. We develop a composite distributed controller for each agent by a combination of internal model and neural network. All agent outputs are proven to reach the constrained minimal point of the aggregate objective function with bounded residual errors irrespective of the unknown nonlinearities and external disturbances. Two examples are finally given to demonstrate the effectiveness of the algorithm.

math.OC↗

arXiv2022

Nash Equilibrium Seeking Over Directed Graphs

In this paper, we aim to develop distributed continuous-time algorithms over directed graphs to seek the Nash equilibrium in a noncooperative game. Motivated by the recent consensus-based designs, we present a distributed algorithm with a proportional gain for weight-balanced directed graphs. By further embedding a distributed estimator of the left eigenvector associated with zero eigenvalue of the graph Laplacian, we extend it to the case with arbitrary strongly connected directed graphs having possible unbalanced weights. In both cases, the Nash equilibrium is proven to be exactly reached with an exponential convergence rate. An example is given to illustrate the validity of the theoretical results.

math.OC↗

arXiv2021

Optimal Consensus for Uncertain High-order Multi-agent Systems by Output Feedback

The distributed optimal output consensus problem for high-order multi-agent systems has been studied recently. In this paper, we further focus on the same problem for high-order multi-agent systems subject to parametric uncertainties and aim at distributed robust controllers by measurement output feedback. We first develop a dynamic compensator to estimate the expected optimal consensus point and convert the problem into several decentralized robust tracking problems. Then, by combining the integral control technique and dirty derivative observer technique, we constructively propose a distributed output feedback integral controller to solve this problem under a mild graph connectivity condition.

eess.SY↗

arXiv2021

Primal-dual $\varepsilon$-Subgradient Method for Distributed Optimization

This paper studies the distributed optimization problem when the objective functions might be nondifferentiable and subject to heterogeneous set constraints. Unlike existing subgradient methods, we focus on the case when the exact subgradients of the local objective functions can not be accessed by the agents. To solve this problem, we propose a projected primal-dual dynamics using only the objective function's approximate subgradients. We first prove that the formulated optimization problem can generally be solved with an error depending upon the accuracy of the available subgradients. Then, we show the exact solvability of this distributed optimization problem when the accumulated approximation error of inexact subgradients is not too large. After that, we also give a novel componentwise normalized variant to improve the transient behavior of the convergent sequence. The effectiveness of our algorithms is verified by a numerical example.

math.OC↗