Zhe Zhang

Zhe Zhang, Gaochang Wu, Jing Zhang et al.

Unsupervised Domain Adaptive Semantic Segmentation (UDA-SS) aims to transfer the supervision from a labeled source domain to an unlabeled target domain. The majority of existing UDA-SS works typically consider images whilst recent attempts have extended further to tackle videos by modeling the temporal dimension. Although the two lines of research share the major challenges -- overcoming the underlying domain distribution shift, their studies are largely independent, resulting in fragmented insights, a lack of holistic understanding, and missed opportunities for cross-pollination of ideas. This fragmentation prevents the unification of methods, leading to redundant efforts and suboptimal knowledge transfer across image and video domains. Under this observation, we advocate unifying the study of UDA-SS across video and image scenarios, enabling a more comprehensive understanding, synergistic advancements, and efficient knowledge sharing. To that end, we explore the unified UDA-SS from a general data augmentation perspective, serving as a unifying conceptual framework, enabling improved generalization, and potential for cross-pollination of ideas, ultimately contributing to the overall progress and practical impact of this field of research. Specifically, we propose a Quad-directional Mixup (QuadMix) method, characterized by tackling distinct point attributes and feature inconsistencies through four-directional paths for intra- and inter-domain mixing in a feature space. To deal with temporal shifts with videos, we incorporate optical flow-guided feature aggregation across spatial and temporal dimensions for fine-grained domain alignment. Extensive experiments show that our method outperforms the state-of-the-art works by large margins on four challenging UDA-SS benchmarks. Our source code and models will be released at https://github.com/ZHE-SAPI/UDASS.

Ryan Humble, Zhe Zhang, Finn O'Shea et al.

Anomaly detection is an important task for complex systems (e.g., industrial facilities, manufacturing, large-scale science experiments), where failures in a sub-system can lead to low yield, faulty products, or even damage to components. While complex systems often have a wealth of data, labeled anomalies are typically rare (or even nonexistent) and expensive to acquire. Unsupervised approaches are therefore common and typically search for anomalies either by distance or density of examples in the input feature space (or some associated low-dimensional representation). This paper presents a novel approach called CoAD, which is specifically designed for multi-modal tasks and identifies anomalies based on \textit{coincident} behavior across two different slices of the feature space. We define an \textit{unsupervised} metric, $\hat{F}_β$, out of analogy to the supervised classification $F_β$ statistic. CoAD uses $\hat{F}_β$ to train an anomaly detection algorithm on \textit{unlabeled data}, based on the expectation that anomalous behavior in one feature slice is coincident with anomalous behavior in the other. The method is illustrated using a synthetic outlier data set and a MNIST-based image data set, and is compared to prior state-of-the-art on two real-world tasks: a metal milling data set and a data set from a particle accelerator.

Yaozhong Gan, Zhe Zhang, Xiaoyang Tan

Advantage learning (AL) aims to improve the robustness of value-based reinforcement learning against estimation errors with action-gap-based regularization. Unfortunately, the method tends to be unstable in the case of function approximation. In this paper, we propose a simple variant of AL, named smoothing advantage learning (SAL), to alleviate this problem. The key to our method is to replace the original Bellman Optimal operator in AL with a smooth one so as to obtain more reliable estimation of the temporal difference target. We give a detailed account of the resulting action gap and the performance bound for approximate SAL. Further theoretical analysis reveals that the proposed value smoothing technique not only helps to stabilize the training procedure of AL by controlling the trade-off between convergence rate and the upper bound of the approximation errors, but is beneficial to increase the action gap between the optimal and sub-optimal action value as well.

Zhe Zhang, Yaozhong Gan, Xiaoyang Tan

Advantage Learning (AL) seeks to increase the action gap between the optimal action and its competitors, so as to improve the robustness to estimation errors. However, the method becomes problematic when the optimal action induced by the approximated value function does not agree with the true optimal action. In this paper, we present a novel method, named clipped Advantage Learning (clipped AL), to address this issue. The method is inspired by our observation that increasing the action gap blindly for all given samples while not taking their necessities into account could accumulate more errors in the performance loss bound, leading to a slow value convergence, and to avoid that, we should adjust the advantage value adaptively. We show that our simple clipped AL operator not only enjoys fast convergence guarantee but also retains proper action gaps, hence achieving a good balance between the large action gap and the fast convergence. The feasibility and effectiveness of the proposed method are verified empirically on several RL benchmarks with promising performance.

Ming Zhao, Pingping Liu, Tongshun Zhang et al.

Low-light image enhancement presents two primary challenges: 1) Significant variations in low-light images across different conditions, and 2) Enhancement levels influenced by subjective preferences and user intent. To address these issues, we propose ReF-LLE, a novel personalized low-light image enhancement method that operates in the Fourier frequency domain and incorporates deep reinforcement learning. ReF-LLE is the first to integrate deep reinforcement learning into this domain. During training, a zero-reference image evaluation strategy is introduced to score enhanced images, providing reward signals that guide the model to handle varying degrees of low-light conditions effectively. In the inference phase, ReF-LLE employs a personalized adaptive iterative strategy, guided by the zero-frequency component in the Fourier domain, which represents the overall illumination level. This strategy enables the model to adaptively adjust low-light images to align with the illumination distribution of a user-provided reference image, ensuring personalized enhancement results. Extensive experiments on benchmark datasets demonstrate that ReF-LLE outperforms state-of-the-art methods, achieving superior perceptual quality and adaptability in personalized low-light image enhancement.

Zhe Zhang, Shiyao Ma, Zhaohui Yang et al.

Air access networks have been recognized as a significant driver of various Internet of Things (IoT) services and applications. In particular, the aerial computing network infrastructure centered on the Internet of Drones has set off a new revolution in automatic image recognition. This emerging technology relies on sharing ground truth labeled data between Unmanned Aerial Vehicle (UAV) swarms to train a high-quality automatic image recognition model. However, such an approach will bring data privacy and data availability challenges. To address these issues, we first present a Semi-supervised Federated Learning (SSFL) framework for privacy-preserving UAV image recognition. Specifically, we propose model parameters mixing strategy to improve the naive combination of FL and semi-supervised learning methods under two realistic scenarios (labels-at-client and labels-at-server), which is referred to as Federated Mixing (FedMix). Furthermore, there are significant differences in the number, features, and distribution of local data collected by UAVs using different camera modules in different environments, i.e., statistical heterogeneity. To alleviate the statistical heterogeneity problem, we propose an aggregation rule based on the frequency of the client's participation in training, namely the FedFreq aggregation rule, which can adjust the weight of the corresponding local model according to its frequency. Numerical results demonstrate that the performance of our proposed method is significantly better than those of the current baseline and is robust to different non-IID levels of client data.

Zhe Zhang, Shiyao Ma, Jiangtian Nie et al.

Federated Learning (FL) allows edge devices (or clients) to keep data locally while simultaneously training a shared high-quality global model. However, current research is generally based on an assumption that the training data of local clients have ground-truth. Furthermore, FL faces the challenge of statistical heterogeneity, i.e., the distribution of the client's local training data is non-independent identically distributed (non-IID). In this paper, we present a robust semi-supervised FL system design, where the system aims to solve the problem of data availability and non-IID in FL. In particular, this paper focuses on studying the labels-at-server scenario where there is only a limited amount of labeled data on the server and only unlabeled data on the clients. In our system design, we propose a novel method to tackle the problems, which we refer to as Federated Mixing (FedMix). FedMix improves the naive combination of FL and semi-supervised learning methods and designs parameter decomposition strategies for disjointed learning of labeled, unlabeled data, and global models. To alleviate the non-IID problem, we propose a novel aggregation rule based on the frequency of the client's participation in training, namely the FedFreq aggregation algorithm, which can adjust the weight of the corresponding local model according to this frequency. Extensive evaluations conducted on CIFAR-10 dataset show that the performance of our proposed method is significantly better than those of the current baseline. It is worth noting that our system is robust to different non-IID levels of client data.

Anirudh Raju, Milind Rao, Gautam Tiwari et al.

Spoken language understanding (SLU) systems extract both text transcripts and semantics associated with intents and slots from input speech utterances. SLU systems usually consist of (1) an automatic speech recognition (ASR) module, (2) an interface module that exposes relevant outputs from ASR, and (3) a natural language understanding (NLU) module. Interfaces in SLU systems carry information on text transcriptions or richer information like neural embeddings from ASR to NLU. In this paper, we study how interfaces affect joint-training for spoken language understanding. Most notably, we obtain the state-of-the-art results on the publicly available 50-hr SLURP dataset. We first leverage large-size pretrained ASR and NLU models that are connected by a text interface, and then jointly train both models via a sequence loss function. For scenarios where pretrained models are not utilized, the best results are obtained through a joint sequence loss training using richer neural interfaces. Finally, we show the overall diminishing impact of leveraging pretrained models with increased training data size.

Jingyuan Chen, Guanchen Ding, Yuchen Yang et al.

Traffic anomaly detection has played a crucial role in Intelligent Transportation System (ITS). The main challenges of this task lie in the highly diversified anomaly scenes and variational lighting conditions. Although much work has managed to identify the anomaly in homogenous weather and scene, few resolved to cope with complex ones. In this paper, we proposed a dual-modality modularized methodology for the robust detection of abnormal vehicles. We introduced an integrated anomaly detection framework comprising the following modules: background modeling, vehicle tracking with detection, mask construction, Region of Interest (ROI) backtracking, and dual-modality tracing. Concretely, we employed background modeling to filter the motion information and left the static information for later vehicle detection. For the vehicle detection and tracking module, we adopted YOLOv5 and multi-scale tracking to localize the anomalies. Besides, we utilized the frame difference and tracking results to identify the road and obtain the mask. In addition, we introduced multiple similarity estimation metrics to refine the anomaly period via backtracking. Finally, we proposed a dual-modality bilateral tracing module to refine the time further. The experiments conducted on the Track 4 testset of the NVIDIA 2021 AI City Challenge yielded a result of 0.9302 F1-Score and 3.4039 root mean square error (RMSE), indicating the effectiveness of our framework.

Rui Xiang, Feng Zheng, Huapeng Su et al.

In this paper, we propose an end-to-end deep learning network named 3dDepthNet, which produces an accurate dense depth image from a single pair of sparse LiDAR depth and color image for robotics and autonomous driving tasks. Based on the dimensional nature of depth images, our network offers a novel 3D-to-2D coarse-to-fine dual densification design that is both accurate and lightweight. Depth densification is first performed in 3D space via point cloud completion, followed by a specially designed encoder-decoder structure that utilizes the projected dense depth from 3D completion and the original RGB-D images to perform 2D image completion. Experiments on the KITTI dataset show our network achieves state-of-art accuracy while being more efficient. Ablation and generalization tests prove that each module in our network has positive influences on the final results, and furthermore, our network is resilient to even sparser depth.

Feng Zheng, Grace Tsai, Zhe Zhang et al.

In this paper, we present the Trifo Visual Inertial Odometry (Trifo-VIO), a tightly-coupled filtering-based stereo VIO system using both points and lines. Line features help improve system robustness in challenging scenarios when point features cannot be reliably detected or tracked, e.g. low-texture environment or lighting change. In addition, we propose a novel lightweight filtering-based loop closing technique to reduce accumulated drift without global bundle adjustment or pose graph optimization. We formulate loop closure as EKF updates to optimally relocate the current sliding window maintained by the filter to past keyframes. We also present the Trifo Ironsides dataset, a new visual-inertial dataset, featuring high-quality synchronized stereo camera and IMU data from the Ironsides sensor [3] with various motion types and textures and millimeter-accuracy groundtruth. To validate the performance of the proposed system, we conduct extensive comparison with state-of-the-art approaches (OKVIS, VINS-MONO and S-MSCKF) using both the public EuRoC dataset and the Trifo Ironsides dataset.

Zhe Zhang, Shaoshan Liu, Grace Tsai et al.

In this paper, we present the PerceptIn Robotics Vision System (PIRVS) system, a visual-inertial computing hardware with embedded simultaneous localization and mapping (SLAM) algorithm. The PIRVS hardware is equipped with a multi-core processor, a global-shutter stereo camera, and an IMU with precise hardware synchronization. The PIRVS software features a novel and flexible sensor fusion approach to not only tightly integrate visual measurements with inertial measurements and also to loosely couple with additional sensor modalities. It runs in real-time on both PC and the PIRVS hardware. We perform a thorough evaluation of the proposed system using multiple public visual-inertial datasets. Experimental results demonstrate that our system reaches comparable accuracy of state-of-the-art visual-inertial algorithms on PC, while being more efficient on the PIRVS hardware.