arXiv subjects

Chenxi Huang

Publications and source records attributed to Chenxi Huang.

35 records · Page 2Linked to original sources

arXiv2024

S$^2$-FPN: Scale-ware Strip Attention Guided Feature Pyramid Network for Real-time Semantic Segmentation

Modern high-performance semantic segmentation methods employ a heavy backbone and dilated convolution to extract the relevant feature. Although extracting features with both contextual and semantic information is critical for the segmentation tasks, it brings a memory footprint and high computation cost for real-time applications. This paper presents a new model to achieve a trade-off between accuracy/speed for real-time road scene semantic segmentation. Specifically, we proposed a lightweight model named Scale-aware Strip Attention Guided Feature Pyramid Network (S$^2$-FPN). Our network consists of three main modules: Attention Pyramid Fusion (APF) module, Scale-aware Strip Attention Module (SSAM), and Global Feature Upsample (GFU) module. APF adopts an attention mechanisms to learn discriminative multi-scale features and help close the semantic gap between different levels. APF uses the scale-aware attention to encode global context with vertical stripping operation and models the long-range dependencies, which helps relate pixels with similar semantic label. In addition, APF employs channel-wise reweighting block (CRB) to emphasize the channel features. Finally, the decoder of S$^2$-FPN then adopts GFU, which is used to fuse features from APF and the encoder. Extensive experiments have been conducted on two challenging semantic segmentation benchmarks, which demonstrate that our approach achieves better accuracy/speed trade-off with different model settings. The proposed models have achieved a results of 76.2\%mIoU/87.3FPS, 77.4\%mIoU/67FPS, and 77.8\%mIoU/30.5FPS on Cityscapes dataset, and 69.6\%mIoU,71.0\% mIoU, and 74.2\% mIoU on Camvid dataset. The code for this work will be made available at \url{https://github.com/mohamedac29/S2-FPN

cs.CV↗

arXiv2024

Interaction-driven breakdown of Aharonov--Bohm caging in flat-band Rydberg lattices

Flat bands play a central role in hosting emergent states of matter in many condensed matter systems, from the nascent insulating states of twisted bilayer graphene to the fractionalized excitations found in frustrated magnets and quantum Hall materials. Here, we report on the experimental realization of highly tunable flat-band models populated by strongly interacting Rydberg atoms. Using the approach of synthetic dimensions, we engineer a flat-band rhombic lattice with twisted boundaries, and through nonequilibrium dynamics we explore the control of Aharonov--Bohm (AB) caging via a tunable $U(1)$ gauge field. Through microscopic measurements of Rydberg pairs, we explore the interaction-driven breakdown of AB caging in the limit of strong dipolar interactions that mix the lattice bands. In the limit of weak interactions, where caging remains intact, we observe an effective magnetism that arises due to the interaction-driven mixing of degenerate flat-band states. These observations of strongly correlated flat-band dynamics open the door to explorations of new emergent phenomena in synthetic quantum materials.

cond-mat.quant-gas↗

arXiv2024

NeRF-Det++: Incorporating Semantic Cues and Perspective-aware Depth Supervision for Indoor Multi-View 3D Detection

NeRF-Det has achieved impressive performance in indoor multi-view 3D detection by innovatively utilizing NeRF to enhance representation learning. Despite its notable performance, we uncover three decisive shortcomings in its current design, including semantic ambiguity, inappropriate sampling, and insufficient utilization of depth supervision. To combat the aforementioned problems, we present three corresponding solutions: 1) Semantic Enhancement. We project the freely available 3D segmentation annotations onto the 2D plane and leverage the corresponding 2D semantic maps as the supervision signal, significantly enhancing the semantic awareness of multi-view detectors. 2) Perspective-aware Sampling. Instead of employing the uniform sampling strategy, we put forward the perspective-aware sampling policy that samples densely near the camera while sparsely in the distance, more effectively collecting the valuable geometric clues. 3)Ordinal Residual Depth Supervision. As opposed to directly regressing the depth values that are difficult to optimize, we divide the depth range of each scene into a fixed number of ordinal bins and reformulate the depth prediction as the combination of the classification of depth bins as well as the regression of the residual depth values, thereby benefiting the depth learning process. The resulting algorithm, NeRF-Det++, has exhibited appealing performance in the ScanNetV2 and ARKITScenes datasets. Notably, in ScanNetV2, NeRF-Det++ outperforms the competitive NeRF-Det by +1.9% in mAP@0.25 and +3.5% in mAP@0.50$. The code will be publicly at https://github.com/mrsempress/NeRF-Detplusplus.

cs.CV↗

arXiv2024

OBMO: One Bounding Box Multiple Objects for Monocular 3D Object Detection

Compared to typical multi-sensor systems, monocular 3D object detection has attracted much attention due to its simple configuration. However, there is still a significant gap between LiDAR-based and monocular-based methods. In this paper, we find that the ill-posed nature of monocular imagery can lead to depth ambiguity. Specifically, objects with different depths can appear with the same bounding boxes and similar visual features in the 2D image. Unfortunately, the network cannot accurately distinguish different depths from such non-discriminative visual features, resulting in unstable depth training. To facilitate depth learning, we propose a simple yet effective plug-and-play module, \underline{O}ne \underline{B}ounding Box \underline{M}ultiple \underline{O}bjects (OBMO). Concretely, we add a set of suitable pseudo labels by shifting the 3D bounding box along the viewing frustum. To constrain the pseudo-3D labels to be reasonable, we carefully design two label scoring strategies to represent their quality. In contrast to the original hard depth labels, such soft pseudo labels with quality scores allow the network to learn a reasonable depth range, boosting training stability and thus improving final performance. Extensive experiments on KITTI and Waymo benchmarks show that our method significantly improves state-of-the-art monocular 3D detectors by a significant margin (The improvements under the moderate setting on KITTI validation set are $\mathbf{1.82\sim 10.91\%}$ \textbf{mAP in BEV} and $\mathbf{1.18\sim 9.36\%}$ \textbf{mAP in 3D}). Codes have been released at \url{https://github.com/mrsempress/OBMO}.

cs.CV↗

arXiv2023

Technical Report on Subspace Pyramid Fusion Network for Semantic Segmentation

The following is a technical report to test the validity of the proposed Subspace Pyramid Fusion Module (SPFM) to capture multi-scale feature representations, which is more useful for semantic segmentation. In this investigation, we have proposed the Efficient Shuffle Attention Module(ESAM) to reconstruct the skip-connections paths by fusing multi-level global context features. Experimental results on two well-known semantic segmentation datasets, including Camvid and Cityscapes, show the effectiveness of our proposed method.

cs.CV↗

arXiv2023

Towards Head Computed Tomography Image Reconstruction Standardization with Deep Learning Assisted Automatic Detection

Three-dimensional (3D) reconstruction of head Computed Tomography (CT) images elucidates the intricate spatial relationships of tissue structures, thereby assisting in accurate diagnosis. Nonetheless, securing an optimal head CT scan without deviation is challenging in clinical settings, owing to poor positioning by technicians, patient's physical constraints, or CT scanner tilt angle restrictions. Manual formatting and reconstruction not only introduce subjectivity but also strain time and labor resources. To address these issues, we propose an efficient automatic head CT images 3D reconstruction method, improving accuracy and repeatability, as well as diminishing manual intervention. Our approach employs a deep learning-based object detection algorithm, identifying and evaluating orbitomeatal line landmarks to automatically reformat the images prior to reconstruction. Given the dearth of existing evaluations of object detection algorithms in the context of head CT images, we compared ten methods from both theoretical and experimental perspectives. By exploring their precision, efficiency, and robustness, we singled out the lightweight YOLOv8 as the aptest algorithm for our task, with an mAP of 92.77% and impressive robustness against class imbalance. Our qualitative evaluation of standardized reconstruction results demonstrates the clinical practicability and validity of our method.

cs.CV↗

arXiv2023

Shuffled Differentially Private Federated Learning for Time Series Data Analytics

Trustworthy federated learning aims to achieve optimal performance while ensuring clients' privacy. Existing privacy-preserving federated learning approaches are mostly tailored for image data, lacking applications for time series data, which have many important applications, like machine health monitoring, human activity recognition, etc. Furthermore, protective noising on a time series data analytics model can significantly interfere with temporal-dependent learning, leading to a greater decline in accuracy. To address these issues, we develop a privacy-preserving federated learning algorithm for time series data. Specifically, we employ local differential privacy to extend the privacy protection trust boundary to the clients. We also incorporate shuffle techniques to achieve a privacy amplification, mitigating the accuracy decline caused by leveraging local differential privacy. Extensive experiments were conducted on five time series datasets. The evaluation results reveal that our algorithm experienced minimal accuracy loss compared to non-private federated learning in both small and large client scenarios. Under the same level of privacy protection, our algorithm demonstrated improved accuracy compared to the centralized differentially private federated learning in both scenarios.

cs.LG↗

arXiv2023

Strongly interacting Rydberg atoms in synthetic dimensions with a magnetic flux

Synthetic dimensions, wherein dynamics occurs in a set of internal states, have found great success in recent years in exploring topological effects in cold atoms and photonics. However, the phenomena thus far explored have largely been restricted to the non-interacting or weakly interacting regimes. Here, we extend the synthetic dimensions playbook to strongly interacting systems of Rydberg atoms prepared in optical tweezer arrays. We use precise control over driving microwave fields to introduce a tunable $U(1)$ flux in a four-site lattice of coupled Rydberg levels. We find highly coherent dynamics, in good agreement with theory. Single atoms show oscillatory dynamics controllable by the gauge field. Small arrays of interacting atoms exhibit behavior suggestive of the emergence of ergodic and arrested dynamics in the regimes of intermediate and strong interactions, respectively. These demonstrations pave the way for future explorations of strongly interacting dynamics and many-body phases in Rydberg synthetic lattices.

cond-mat.quant-gas↗

arXiv2023

Neural Collapse Inspired Federated Learning with Non-iid Data

One of the challenges in federated learning is the non-independent and identically distributed (non-iid) characteristics between heterogeneous devices, which cause significant differences in local updates and affect the performance of the central server. Although many studies have been proposed to address this challenge, they only focus on local training and aggregation processes to smooth the changes and fail to achieve high performance with deep learning models. Inspired by the phenomenon of neural collapse, we force each client to be optimized toward an optimal global structure for classification. Specifically, we initialize it as a random simplex Equiangular Tight Frame (ETF) and fix it as the unit optimization target of all clients during the local updating. After guaranteeing all clients are learning to converge to the global optimum, we propose to add a global memory vector for each category to remedy the parameter fluctuation caused by the bias of the intra-class condition distribution among clients. Our experimental results show that our method can improve the performance with faster convergence speed on different-size datasets.

cs.LG↗

arXiv2023

Self-supervised and Weakly Supervised Contrastive Learning for Frame-wise Action Representations

Previous work on action representation learning focused on global representations for short video clips. In contrast, many practical applications, such as video alignment, strongly demand learning the intensive representation of long videos. In this paper, we introduce a new framework of contrastive action representation learning (CARL) to learn frame-wise action representation in a self-supervised or weakly-supervised manner, especially for long videos. Specifically, we introduce a simple but effective video encoder that considers both spatial and temporal context by combining convolution and transformer. Inspired by the recent massive progress in self-supervised learning, we propose a new sequence contrast loss (SCL) applied to two related views obtained by expanding a series of spatio-temporal data in two versions. One is the self-supervised version that optimizes embedding space by minimizing KL-divergence between sequence similarity of two augmented views and prior Gaussian distribution of timestamp distance. The other is the weakly-supervised version that builds more sample pairs among videos using video-level labels by dynamic time wrapping (DTW). Experiments on FineGym, PennAction, and Pouring datasets show that our method outperforms previous state-of-the-art by a large margin for downstream fine-grained action classification and even faster inference. Surprisingly, although without training on paired videos like in previous works, our self-supervised version also shows outstanding performance in video alignment and fine-grained frame retrieval tasks.

cs.CV↗

arXiv2022

Sparse Fuse Dense: Towards High Quality 3D Detection with Depth Completion

Current LiDAR-only 3D detection methods inevitably suffer from the sparsity of point clouds. Many multi-modal methods are proposed to alleviate this issue, while different representations of images and point clouds make it difficult to fuse them, resulting in suboptimal performance. In this paper, we present a novel multi-modal framework SFD (Sparse Fuse Dense), which utilizes pseudo point clouds generated from depth completion to tackle the issues mentioned above. Different from prior works, we propose a new RoI fusion strategy 3D-GAF (3D Grid-wise Attentive Fusion) to make fuller use of information from different types of point clouds. Specifically, 3D-GAF fuses 3D RoI features from the couple of point clouds in a grid-wise attentive way, which is more fine-grained and more precise. In addition, we propose a SynAugment (Synchronized Augmentation) to enable our multi-modal framework to utilize all data augmentation approaches tailored to LiDAR-only methods. Lastly, we customize an effective and efficient feature extractor CPConv (Color Point Convolution) for pseudo point clouds. It can explore 2D image features and 3D geometric features of pseudo point clouds simultaneously. Our method holds the highest entry on the KITTI car 3D object detection leaderboard, demonstrating the effectiveness of our SFD. Codes are available at https://github.com/LittlePey/SFD.

cs.CV↗

arXiv2022

Simulating Quantum Mechanics with a $θ$-term and an 't Hooft Anomaly on a Synthetic Dimension

A topological $θ$-term in gauge theories, including quantum chromodynamics in 3+1 dimensions, gives rise to a sign problem that makes classical Monte Carlo simulations impractical. Quantum simulations are not subject to such sign problems and are a promising approach to studying these theories in the future. In the near term, it is interesting to study simpler models that retain some of the physical phenomena of interest and their implementation on quantum hardware. For example, dimensionally-reducing gauge theories on small spatial tori produces quantum mechanical models which, despite being relatively simple to solve, retain interesting vacuum and symmetry structures from the parent gauge theories. Here we consider quantum mechanical particle-on-a-circle models, related by dimensional reduction to the 1+1d Schwinger model, that possess a $θ$-term and realize an 't Hooft anomaly or global inconsistency at $θ= π$. These models also exhibit the related phenomena of spontaneous symmetry breaking and instanton-anti-instanton interference in real time. We propose an experimental scheme for the real-time simulation of a particle on a circle with a $θ$-term and a $\mathbb{Z}_n$ potential using a synthetic dimension encoded in a Rydberg atom. Simulating the Rydberg atom with realistic experimental parameters, we demonstrate that the essential physics can be well-captured by the experiment, with expected behavior in the tunneling rate as a function of $θ$. Similar phenomena and observables can also arise in more complex quantum mechanical models connected to higher-dimensional nonabelian gauge theories by dimensional reduction.

quant-ph↗

arXiv2021

Gray molasses cooling of $^{39}$K atoms in optical tweezers

Robust cooling and nondestructive imaging are prerequisites for many emerging applications of neutral atoms trapped in optical tweezers, such as their use in quantum information science and analog quantum simulation. The tasks of cooling and imaging can be challenged, however, by the presence of large trap-induced shifts of their respective optical transitions. Here, we explore a system of $^{39}$K atoms trapped in a near-detuned ($780$ nm) optical tweezer, which leads to relatively minor differential (ground vs. excited state) Stark shifts. We demonstrate that simple and robust loading, cooling, and imaging can be achieved through a combined addressing of the D$_\textrm{1}$ and D$_\textrm{2}$ transitions. While imaging on the D$_\textrm{2}$ transition, we can simultaneously apply $Λ$-enhanced gray molasses (GM) on the D$_\textrm{1}$ transition, preserving low backgrounds for single-atom imaging through spectral filtering. Using D$_\textrm{1}$ cooling during and after trap loading, we demonstrate enhanced loading efficiencies as well as cooling to low temperatures. These results suggest a simple and robust path for loading and cooling large arrays of potassium atoms in optical tweezers through the use of resource-efficient near-detuned optical tweezers and GM cooling.

physics.atom-ph↗

arXiv2020

Stochastic Greedy Algorithms For Multiple Measurement Vectors

Sparse representation of a single measurement vector (SMV) has been explored in a variety of compressive sensing applications. Recently, SMV models have been extended to solve multiple measurement vectors (MMV) problems, where the underlying signal is assumed to have joint sparse structures. To circumvent the NP-hardness of the $\ell_0$ minimization problem, many deterministic MMV algorithms solve the convex relaxed models with limited efficiency. In this paper, we develop stochastic greedy algorithms for solving the joint sparse MMV reconstruction problem. In particular, we propose the MMV Stochastic Iterative Hard Thresholding (MStoIHT) and MMV Stochastic Gradient Matching Pursuit (MStoGradMP) algorithms, and we also utilize the mini-batching technique to further improve their performance. Convergence analysis indicates that the proposed algorithms are able to converge faster than their SMV counterparts, i.e., concatenated StoIHT and StoGradMP, under certain conditions. Numerical experiments have illustrated the superior effectiveness of the proposed algorithms over their SMV counterparts.

math.OC↗

arXiv2019

Learning Rate Dropout

The performance of a deep neural network is highly dependent on its training, and finding better local optimal solutions is the goal of many optimization algorithms. However, existing optimization algorithms show a preference for descent paths that converge slowly and do not seek to avoid bad local optima. In this work, we propose Learning Rate Dropout (LRD), a simple gradient descent technique for training related to coordinate descent. LRD empirically aids the optimizer to actively explore in the parameter space by randomly setting some learning rates to zero; at each iteration, only parameters whose learning rate is not 0 are updated. As the learning rate of different parameters is dropped, the optimizer will sample a new loss descent path for the current update. The uncertainty of the descent path helps the model avoid saddle points and bad local minima. Experiments show that LRD is surprisingly effective in accelerating training while preventing overfitting.

cs.CV↗

arXiv2018

Compressed Anomaly Detection with Multiple Mixed Observations

We consider a collection of independent random variables that are identically distributed, except for a small subset which follows a different, anomalous distribution. We study the problem of detecting which random variables in the collection are governed by the anomalous distribution. Recent work proposes to solve this problem by conducting hypothesis tests based on mixed observations (e.g. linear combinations) of the random variables. Recognizing the connection between taking mixed observations and compressed sensing, we view the problem as recovering the "support" (index set) of the anomalous random variables from multiple measurement vectors (MMVs). Many algorithms have been developed for recovering jointly sparse signals and their support from MMVs. We establish the theoretical and empirical effectiveness of these algorithms at detecting anomalies. We also extend the LASSO algorithm to an MMV version for our purpose. Further, we perform experiments on synthetic data, consisting of samples from the random variables, to explore the trade-off between the number of mixed observations per sample and the number of samples required to detect anomalies.

cs.IT↗

arXiv2018

Sparse Randomized Kaczmarz for Support Recovery of Jointly Sparse Corrupted Multiple Measurement Vectors

While single measurement vector (SMV) models have been widely studied in signal processing, there is a surging interest in addressing the multiple measurement vectors (MMV) problem. In the MMV setting, more than one measurement vector is available and the multiple signals to be recovered share some commonalities such as a common support. Applications in which MMV is a naturally occurring phenomenon include online streaming, medical imaging, and video recovery. This work presents a stochastic iterative algorithm for the support recovery of jointly sparse corrupted MMV. We present a variant of the Sparse Randomized Kaczmarz algorithm for corrupted MMV and compare our proposed method with an existing Kaczmarz type algorithm for MMV problems. We also showcase the usefulness of our approach in the online (streaming) setting and provide empirical evidence that suggests the robustness of the proposed method to the distribution of the corruption and the number of corruptions occurring.

eess.SP↗