Chao Chen

Chao Chen, Yu-Shen Liu, Zhizhong Han · tsinghua

Deep implicit functions have shown remarkable shape modeling ability in various 3D computer vision tasks. One drawback is that it is hard for them to represent a 3D shape as multiple parts. Current solutions learn various primitives and blend the primitives directly in the spatial space, which still struggle to approximate the 3D shape accurately. To resolve this problem, we introduce a novel implicit representation to represent a single 3D shape as a set of parts in the latent space, towards both highly accurate and plausibly interpretable shape modeling. Our insight here is that both the part learning and the part blending can be conducted much easier in the latent space than in the spatial space. We name our method Latent Partition Implicit (LPI), because of its ability of casting the global shape modeling into multiple local part modeling, which partitions the global shape unity. LPI represents a shape as Signed Distance Functions (SDFs) using surface codes. Each surface code is a latent code representing a part whose center is on the surface, which enables us to flexibly employ intrinsic attributes of shapes or additional surface properties. Eventually, LPI can reconstruct both the shape and the parts on the shape, both of which are plausible meshes. LPI is a multi-level representation, which can partition a shape into different numbers of parts after training. LPI can be learned without ground truth signed distances, point normals or any supervision for part partition. LPI outperforms the latest methods under the widely used benchmarks in terms of reconstruction accuracy and modeling interpretability. Our code, data and models are available at https://github.com/chenchao15/LPI.

Chao Chen, Yu-Shen Liu, Zhizhong Han · tsinghua

It is vital to infer signed distance functions (SDFs) from 3D point clouds. The latest methods rely on generalizing the priors learned from large scale supervision. However, the learned priors do not generalize well to various geometric variations that are unseen during training, especially for extremely sparse point clouds. To resolve this issue, we present a neural network to directly infer SDFs from single sparse point clouds without using signed distance supervision, learned priors or even normals. Our insight here is to learn surface parameterization and SDFs inference in an end-to-end manner. To make up the sparsity, we leverage parameterized surfaces as a coarse surface sampler to provide many coarse surface estimations in training iterations, according to which we mine supervision and our thin plate splines (TPS) based network infers SDFs as smooth functions in a statistical way. Our method significantly improves the generalization ability and accuracy in unseen point clouds. Our experimental results show our advantages over the state-of-the-art methods in surface reconstruction for sparse point clouds under synthetic datasets and real scans.The code is available at \url{https://github.com/chenchao15/NeuralTPS}.

Mingrui Wu, Jiaxin Gu, Yunhang Shen et al.

Most existing Human-Object Interaction~(HOI) Detection methods rely heavily on full annotations with predefined HOI categories, which is limited in diversity and costly to scale further. We aim at advancing zero-shot HOI detection to detect both seen and unseen HOIs simultaneously. The fundamental challenges are to discover potential human-object pairs and identify novel HOI categories. To overcome the above challenges, we propose a novel end-to-end zero-shot HOI Detection (EoID) framework via vision-language knowledge distillation. We first design an Interactive Score module combined with a Two-stage Bipartite Matching algorithm to achieve interaction distinguishment for human-object pairs in an action-agnostic manner. Then we transfer the distribution of action probability from the pretrained vision-language teacher as well as the seen ground truth to the HOI model to attain zero-shot HOI classification. Extensive experiments on HICO-Det dataset demonstrate that our model discovers potential interactive pairs and enables the recognition of unseen HOIs. Finally, our method outperforms the previous SOTA by 8.92% on unseen mAP and 10.18% on overall mAP under UA setting, by 6.02% on unseen mAP and 9.1% on overall mAP under UC setting. Moreover, our method is generalizable to large-scale object detection data to further scale up the action sets. The source code will be available at: https://github.com/mrwu-mac/EoID.

Yawei Li, Kai Zhang, Radu Timofte et al. · eth-zurich, tencent-ai

This paper reviews the NTIRE 2022 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The task of the challenge was to super-resolve an input image with a magnification factor of $\times$4 based on pairs of low and corresponding high resolution images. The aim was to design a network for single image super-resolution that achieved improvement of efficiency measured according to several metrics including runtime, parameters, FLOPs, activations, and memory consumption while at least maintaining the PSNR of 29.00dB on DIV2K validation set. IMDN is set as the baseline for efficiency measurement. The challenge had 3 tracks including the main track (runtime), sub-track one (model complexity), and sub-track two (overall performance). In the main track, the practical runtime performance of the submissions was evaluated. The rank of the teams were determined directly by the absolute value of the average runtime on the validation set and test set. In sub-track one, the number of parameters and FLOPs were considered. And the individual rankings of the two metrics were summed up to determine a final ranking in this track. In sub-track two, all of the five metrics mentioned in the description of the challenge including runtime, parameter count, FLOPs, activations, and memory consumption were considered. Similar to sub-track one, the rankings of five metrics were summed up to determine a final ranking. The challenge had 303 registered participants, and 43 teams made valid submissions. They gauge the state-of-the-art in efficient single image super-resolution.

Dongli Tan, Jiang-Jiang Liu, Xingyu Chen et al. · tencent-ai

Modeling sparse and dense image matching within a unified functional correspondence model has recently attracted increasing research interest. However, existing efforts mainly focus on improving matching accuracy while ignoring its efficiency, which is crucial for realworld applications. In this paper, we propose an efficient structure named Efficient Correspondence Transformer (ECO-TR) by finding correspondences in a coarse-to-fine manner, which significantly improves the efficiency of functional correspondence model. To achieve this, multiple transformer blocks are stage-wisely connected to gradually refine the predicted coordinates upon a shared multi-scale feature extraction network. Given a pair of images and for arbitrary query coordinates, all the correspondences are predicted within a single feed-forward pass. We further propose an adaptive query-clustering strategy and an uncertainty-based outlier detection module to cooperate with the proposed framework for faster and better predictions. Experiments on various sparse and dense matching tasks demonstrate the superiority of our method in both efficiency and effectiveness against existing state-of-the-arts.

Chengqian Ma, Zhiqiang Wu, Chunlei Cai et al.

In the past decades, lots of progress have been done in the video compression field including traditional video codec and learning-based video codec. However, few studies focus on using preprocessing techniques to improve the rate-distortion performance. In this paper, we propose a rate-perception optimized preprocessing (RPP) method. We first introduce an adaptive Discrete Cosine Transform loss function which can save the bitrate and keep essential high frequency components as well. Furthermore, we also combine several state-of-the-art techniques from low-level vision fields into our approach, such as the high-order degradation model, efficient lightweight network design, and Image Quality Assessment model. By jointly using these powerful techniques, our RPP approach can achieve on average, 16.27% bitrate saving with different video encoders like AVC, HEVC, and VVC under multiple quality metrics. In the deployment stage, our RPP method is very simple and efficient which is not required any changes in the setting of video encoding, streaming, and decoding. Each input frame only needs to make a single pass through RPP before sending into video encoders. In addition, in our subjective visual quality test, 87% of users think videos with RPP are better or equal to videos by only using the codec to compress, while these videos with RPP save about 12% bitrate on average. Our RPP framework has been integrated into the production environment of our video transcoding services which serve millions of users every day.

Jing He, Yiyi Zhou, Qi Zhang et al.

Pixel synthesis is a promising research paradigm for image generation, which can well exploit pixel-wise prior knowledge for generation. However, existing methods still suffer from excessive memory footprint and computation overhead. In this paper, we propose a progressive pixel synthesis network towards efficient image generation, coined as PixelFolder. Specifically, PixelFolder formulates image generation as a progressive pixel regression problem and synthesizes images via a multi-stage structure, which can greatly reduce the overhead caused by large tensor transformations. In addition, we introduce novel pixel folding operations to further improve model efficiency while maintaining pixel-wise prior knowledge for end-to-end regression. With these innovative designs, we greatly reduce the expenditure of pixel synthesis, e.g., reducing 89% computation and 53% parameters compared with the latest pixel synthesis method CIPS. To validate our approach, we conduct extensive experiments on two benchmark datasets, namely FFHQ and LSUN Church. The experimental results show that with much less expenditure, PixelFolder obtains new state-of-the-art (SOTA) performance on two benchmark datasets, i.e., 3.77 FID and 2.45 FID on FFHQ and LSUN Church, respectively.Meanwhile, PixelFolder is also more efficient than the SOTA methods like StyleGAN2, reducing about 72% computation and 31% parameters, respectively. These results greatly validate the effectiveness of the proposed PixelFolder.

Zequn Qin, Jingyu Chen, Chao Chen et al.

Bird's eye view (BEV) representation is a new perception formulation for autonomous driving, which is based on spatial fusion. Further, temporal fusion is also introduced in BEV representation and gains great success. In this work, we propose a new method that unifies both spatial and temporal fusion and merges them into a unified mathematical formulation. The unified fusion could not only provide a new perspective on BEV fusion but also brings new capabilities. With the proposed unified spatial-temporal fusion, our method could support long-range fusion, which is hard to achieve in conventional BEV methods. Moreover, the BEV fusion in our work is temporal-adaptive and the weights of temporal fusion are learnable. In contrast, conventional methods mainly use fixed and equal weights for temporal fusion. Besides, the proposed unified fusion could avoid information lost in conventional BEV fusion methods and make full use of features. Extensive experiments and ablation studies on the NuScenes dataset show the effectiveness of the proposed method and our method gains the state-of-the-art performance in the map segmentation task.

Rebecca Salles, Janio Lima, Michel Reis et al.

Time series event detection methods are evaluated mainly by standard classification metrics that focus solely on detection accuracy. However, inaccuracy in detecting an event can often result from its preceding or delayed effects reflected in neighboring detections. These detections are valuable to trigger necessary actions or help mitigate unwelcome consequences. In this context, current metrics are insufficient and inadequate for the context of event detection. There is a demand for metrics that incorporate both the concept of time and temporal tolerance for neighboring detections. This paper introduces SoftED metrics, a new set of metrics designed for soft evaluating event detection methods. They enable the evaluation of both detection accuracy and the degree to which their detections represent events. They improved event detection evaluation by associating events and their representative detections, incorporating temporal tolerance in over 36\% of experiments compared to the usual classification metrics. SoftED metrics were validated by domain specialists that indicated their contribution to detection evaluation and method selection.

Chen Li, Xiaoling Hu, Shahira Abousamra et al.

Semi-supervised crowd counting is an important yet challenging task. A popular approach is to iteratively generate pseudo-labels for unlabeled data and add them to the training set. The key is to use uncertainty to select reliable pseudo-labels. In this paper, we propose a novel method to calibrate model uncertainty for crowd counting. Our method takes a supervised uncertainty estimation strategy to train the model through a surrogate function. This ensures the uncertainty is well controlled throughout the training. We propose a matching-based patch-wise surrogate function to better approximate uncertainty for crowd counting tasks. The proposed method pays a sufficient amount of attention to details, while maintaining a proper granularity. Altogether our method is able to generate reliable uncertainty estimation, high quality pseudolabels, and achieve state-of-the-art performance in semisupervised crowd counting.

Weimin Lyu, Songzhu Zheng, Tengfei Ma et al.

Trojan attacks raise serious security concerns. In this paper, we investigate the underlying mechanism of Trojaned BERT models. We observe the attention focus drifting behavior of Trojaned models, i.e., when encountering an poisoned input, the trigger token hijacks the attention focus regardless of the context. We provide a thorough qualitative and quantitative analysis of this phenomenon, revealing insights into the Trojan mechanism. Based on the observation, we propose an attention-based Trojan detector to distinguish Trojaned models from clean ones. To the best of our knowledge, this is the first paper to analyze the Trojan mechanism and to develop a Trojan detector based on the transformer's attention.

Wei Chen, Yuxuan Liang, Yuanshao Zhu et al.

Trajectory computing is a pivotal domain encompassing trajectory data management and mining, garnering widespread attention due to its crucial role in various practical applications such as location services, urban traffic, and public safety. Traditional methods, focusing on simplistic spatio-temporal features, face challenges of complex calculations, limited scalability, and inadequate adaptability to real-world complexities. In this paper, we present a comprehensive review of the development and recent advances in deep learning for trajectory computing (DL4Traj). We first define trajectory data and provide a brief overview of widely-used deep learning models. Systematically, we explore deep learning applications in trajectory management (pre-processing, storage, analysis, and visualization) and mining (trajectory-related forecasting, trajectory-related recommendation, trajectory classification, travel time estimation, anomaly detection, and mobility generation). Notably, we encapsulate recent advancements in Large Language Models (LLMs) that hold the potential to augment trajectory computing. Additionally, we summarize application scenarios, public datasets, and toolkits. Finally, we outline current challenges in DL4Traj research and propose future directions. Relevant papers and open-source resources have been collated and are continuously updated at: \href{https://github.com/yoshall/Awesome-Trajectory-Computing}{DL4Traj Repo}.

Tianle Pu, Zijie Geng, Haoyang Liu et al.

Mixed-Integer Linear Programming (MILP) is a fundamental and powerful framework for modeling complex optimization problems across diverse domains. Recently, learning-based methods have shown great promise in accelerating MILP solvers by predicting high-quality solutions. However, most existing approaches are developed and evaluated in single-domain settings, limiting their ability to generalize to unseen problem distributions. This limitation poses a major obstacle to building scalable and general-purpose learning-based solvers. To address this challenge, we introduce RoME, a domain-Robust Mixture-of-Experts framework for predicting MILP solutions across domains. RoME dynamically routes problem instances to specialized experts based on learned task embeddings. The model is trained using a two-level distributionally robust optimization strategy: inter-domain to mitigate global shifts across domains, and intra-domain to enhance local robustness by introducing perturbations on task embeddings. We reveal that cross-domain training not only enhances the model's generalization capability to unseen domains but also improves performance within each individual domain by encouraging the model to capture more general intrinsic combinatorial patterns. Specifically, a single RoME model trained on three domains achieves an average improvement of 67.7% then evaluated on five diverse domains. We further test the pretrained model on MIPLIB in a zero-shot setting, demonstrating its ability to deliver measurable performance gains on challenging real-world instances where existing learning-based approaches often struggle to generalize.

Chao Chen, Mingzhi Zhu, Ankush Pratap Singh et al.

Humans are remarkably efficient at forming spatial understanding from just a few visual observations. When browsing real estate or navigating unfamiliar spaces, they intuitively select a small set of views that summarize the spatial layout. Inspired by this ability, we introduce scene summarization, the task of condensing long, continuous scene videos into a compact set of spatially diverse keyframes that facilitate global spatial reasoning. Unlike conventional video summarization-which focuses on user-edited, fragmented clips and often ignores spatial continuity-our goal is to mimic how humans abstract spatial layout from sparse views. We propose SceneSum, a two-stage self-supervised pipeline that first clusters video frames using visual place recognition to promote spatial diversity, then selects representative keyframes from each cluster under resource constraints. When camera trajectories are available, a lightweight supervised loss further refines clustering and selection. Experiments on real and simulated indoor datasets show that SceneSum produces more spatially informative summaries and outperforms existing video summarization baselines.

Chao Chen, Xiaowei Jin, Hui Li

The partial differential equation (PDE) plays a significantly important role in many fields of science and engineering. The conventional case of the derivation of PDE mainly relies on first principles and empirical observation. However, the development of machine learning technology allows us to mine potential control equations from the massive amounts of stored data in a fresh way. Although there has been considerable progress in the data-driven discovery of PDE, the extant literature mostly focuses on the improvements of discovery methods, without substantial breakthroughs in the discovery process itself, including the principles for the construction of candidates and how to incorporate physical priors. In this paper, through rigorous derivation of formulas, novel physically enhanced machining learning discovery methods for control equations: GSNN (Galileo Symbolic Neural Network) and LSNN (Lorentz Symbolic Neural Network) are firstly proposed based on Galileo invariance and Lorentz invariance respectively, setting forth guidelines for building the candidates of discovering equations. The adoption of mandatory embedding of physical constraints is fundamentally different from PINN in the form of the loss function, thus ensuring that the designed Neural Network strictly obeys the physical prior of invariance and enhancing the interpretability of the network. By comparing the results with PDE-NET in numerical experiments of Burgers equation and Sine-Gordon equation, it shows that the method presented in this study has better accuracy, parsimony, and interpretability.

Chao Chen, Jie Liu, Chang Zhou et al.

Lane detection is to determine the precise location and shape of lanes on the road. Despite efforts made by current methods, it remains a challenging task due to the complexity of real-world scenarios. Existing approaches, whether proposal-based or keypoint-based, suffer from depicting lanes effectively and efficiently. Proposal-based methods detect lanes by distinguishing and regressing a collection of proposals in a streamlined top-down way, yet lack sufficient flexibility in lane representation. Keypoint-based methods, on the other hand, construct lanes flexibly from local descriptors, which typically entail complicated post-processing. In this paper, we present a "Sketch-and-Refine" paradigm that utilizes the merits of both keypoint-based and proposal-based methods. The motivation is that local directions of lanes are semantically simple and clear. At the "Sketch" stage, local directions of keypoints can be easily estimated by fast convolutional layers. Then we can build a set of lane proposals accordingly with moderate accuracy. At the "Refine" stage, we further optimize these proposals via a novel Lane Segment Association Module (LSAM), which allows adaptive lane segment adjustment. Last but not least, we propose multi-level feature integration to enrich lane feature representations more efficiently. Based on the proposed "Sketch and Refine" paradigm, we propose a fast yet effective lane detector dubbed "SRLane". Experiments show that our SRLane can run at a fast speed (i.e., 278 FPS) while yielding an F1 score of 78.9\%. The source code is available at: https://github.com/passerer/SRLane.

Chaoyang Zhu, Yiyi Zhou, Yunhang Shen et al.

In this paper, we propose a simple yet universal network termed SeqTR for visual grounding tasks, e.g., phrase localization, referring expression comprehension (REC) and segmentation (RES). The canonical paradigms for visual grounding often require substantial expertise in designing network architectures and loss functions, making them hard to generalize across tasks. To simplify and unify the modeling, we cast visual grounding as a point prediction problem conditioned on image and text inputs, where either the bounding box or binary mask is represented as a sequence of discrete coordinate tokens. Under this paradigm, visual grounding tasks are unified in our SeqTR network without task-specific branches or heads, e.g., the convolutional mask decoder for RES, which greatly reduces the complexity of multi-task modeling. In addition, SeqTR also shares the same optimization objective for all tasks with a simple cross-entropy loss, further reducing the complexity of deploying hand-crafted loss functions. Experiments on five benchmark datasets demonstrate that the proposed SeqTR outperforms (or is on par with) the existing state-of-the-arts, proving that a simple yet universal approach for visual grounding is indeed feasible. Source code is available at https://github.com/sean-zhuh/SeqTR.

Chen Chao, Zhizhong Han, Yu-Shen Liu et al.

Learning to generate 3D point clouds without 3D supervision is an important but challenging problem. Current solutions leverage various differentiable renderers to project the generated 3D point clouds onto a 2D image plane, and train deep neural networks using the per-pixel difference with 2D ground truth images. However, these solutions are still struggling to fully recover fine structures of 3D shapes, such as thin tubes or planes. To resolve this issue, we propose an unsupervised approach for 3D point cloud generation with fine structures. Specifically, we cast 3D point cloud learning as a 2D projection matching problem. Rather than using entire 2D silhouette images as a regular pixel supervision, we introduce structure adaptive sampling to randomly sample 2D points within the silhouettes as an irregular point supervision, which alleviates the consistency issue of sampling from different view angles. Our method pushes the neural network to generate a 3D point cloud whose 2D projections match the irregular point supervision from different view angles. Our 2D projection matching approach enables the neural network to learn more accurate structure information than using the per-pixel difference, especially for fine and thin 3D structures. Our method can recover fine 3D structures from 2D silhouette images at different resolutions, and is robust to different sampling methods and point number in irregular point supervision. Our method outperforms others under widely used benchmarks. Our code, data and models are available at https://github.com/chenchao15/2D\_projection\_matching.

Weimin Lyu, Xiao Lin, Songzhu Zheng et al.

Textual backdoor attacks pose significant security threats. Current detection approaches, typically relying on intermediate feature representation or reconstructing potential triggers, are task-specific and less effective beyond sentence classification, struggling with tasks like question answering and named entity recognition. We introduce TABDet (Task-Agnostic Backdoor Detector), a pioneering task-agnostic method for backdoor detection. TABDet leverages final layer logits combined with an efficient pooling technique, enabling unified logit representation across three prominent NLP tasks. TABDet can jointly learn from diverse task-specific models, demonstrating superior detection efficacy over traditional task-specific methods.

Keifer Lee, Shubham Gupta, Sunglyoung Kim et al.

Despite the great success of Neural Radiance Fields (NeRF), its data-gathering process remains vague with only a general rule of thumb of sampling as densely as possible. The lack of understanding of what actually constitutes good views for NeRF makes it difficult to actively plan a sequence of views that yield the maximal reconstruction quality. We propose Surrogate Objectives for Active Radiance Fields (SOAR), which is a set of interpretable functions that evaluates the goodness of views using geometric and photometric visual cues - surface coverage, geometric complexity, textural complexity, and ray diversity. Moreover, by learning to infer the SOAR scores from a deep network, SOARNet, we are able to effectively select views in mere seconds instead of hours, without the need for prior visits to all the candidate views or training any radiance field during such planning. Our experiments show SOARNet outperforms the baselines with $\sim$80x speed-up while achieving better or comparable reconstruction qualities. We finally show that SOAR is model-agnostic, thus it generalizes across fully neural-implicit to fully explicit approaches.

Chao Chen, Hui Li, Xiaowei Jin

The discovery of partial differential equations (PDEs) from datasets has attracted increased attention. However, the discovery of governing equations from sparse data with high noise is still very challenging due to the difficulty of derivatives computation and the disturbance of noise. Moreover, the selection principles for the candidate library to meet physical laws need to be further studied. The invariance is one of the fundamental laws for governing equations. In this study, we propose an invariance constrained deep learning network (ICNet) for the discovery of PDEs. Considering that temporal and spatial translation invariance (Galilean invariance) is a fundamental property of physical laws, we filter the candidates that cannot meet the requirement of the Galilean transformations. Subsequently, we embedded the fixed and possible terms into the loss function of neural network, significantly countering the effect of sparse data with high noise. Then, by filtering out redundant terms without fixing learnable parameters during the training process, the governing equations discovered by the ICNet method can effectively approximate the real governing equations. We select the 2D Burgers equation, the equation of 2D channel flow over an obstacle, and the equation of 3D intracranial aneurysm as examples to verify the superiority of the ICNet for fluid mechanics. Furthermore, we extend similar invariance methods to the discovery of wave equation (Lorentz Invariance) and verify it through Single and Coupled Klein-Gordon equation. The results show that the ICNet method with physical constraints exhibits excellent performance in governing equations discovery from sparse and noisy data.

Meilong Xu, Di Fu, Jiaxing Zhang et al.

Vision Language Models (VLMs) are becoming increasingly integral to multimedia understanding; however, they often struggle with domain-specific video classification tasks, particularly in cases with limited data. This stems from a critical \textit{rationale gap}, where sparse domain data is insufficient to bridge the semantic distance between complex spatio-temporal content and abstract classification labels. We propose a two-stage self-improvement paradigm to bridge this gap without new annotations. First, we prompt the VLMs to generate detailed textual rationales for each video, compelling them to articulate the domain-specific logic. The VLM is then fine-tuned on these self-generated rationales, utilizing this intermediate supervision to align its representations with the nuances of the target domain. Second, conventional supervised fine-tuning (SFT) is performed on the task labels, achieving markedly higher effectiveness as a result of the model's pre-acquired domain reasoning. Extensive experiments on diverse datasets demonstrate that our method significantly outperforms direct SFT, validating self-generated rationale as an effective, annotation-efficient paradigm for adapting VLMs to domain-specific video analysis.

Lu Pang, Tao Sun, Weimin Lyu et al.

Recently, backdoor attack has become an increasing security threat to deep neural networks and drawn the attention of researchers. Backdoor attacks exploit vulnerabilities in third-party pretrained models during the training phase, enabling them to behave normally for clean samples and mispredict for samples with specific triggers. Existing backdoor attacks mainly focus on balanced datasets. However, real-world datasets often follow long-tailed distributions. In this paper, for the first time, we explore backdoor attack on such datasets. Specifically, we first analyze the influence of data imbalance on backdoor attack. Based on our analysis, we propose an effective backdoor attack named Dynamic Data Augmentation Operation (D$^2$AO). We design D$^2$AO selectors to select operations depending jointly on the class, sample type (clean vs. backdoored) and sample features. Meanwhile, we develop a trigger generator to generate sample-specific triggers. Through simultaneous optimization of the backdoored model and trigger generator, guided by dynamic data augmentation operation selectors, we achieve significant advancements. Extensive experiments demonstrate that our method can achieve the state-of-the-art attack performance while preserving the clean accuracy.

Chao Chen, Yibing Zhan, Baosheng Yu et al.

Scene Graph Generation (SGG) aims to build a structured representation of a scene using objects and pairwise relationships, which benefits downstream tasks. However, current SGG methods usually suffer from sub-optimal scene graph generation because of the long-tailed distribution of training data. To address this problem, we propose Resistance Training using Prior Bias (RTPB) for the scene graph generation. Specifically, RTPB uses a distributed-based prior bias to improve models' detecting ability on less frequent relationships during training, thus improving the model generalizability on tail categories. In addition, to further explore the contextual information of objects and relationships, we design a contextual encoding backbone network, termed as Dual Transformer (DTrans). We perform extensive experiments on a very popular benchmark, VG150, to demonstrate the effectiveness of our method for the unbiased scene graph generation. In specific, our RTPB achieves an improvement of over 10% under the mean recall when applied to current SGG methods. Furthermore, DTrans with RTPB outperforms nearly all state-of-the-art methods with a large margin.

Ran Cheng, Chao Chen, Longfei Xu et al.

Route recommendation is significant in navigation service. Two major challenges for route recommendation are route representation and user representation. Different from items that can be identified by unique IDs in traditional recommendation, routes are combinations of links (i.e., a road segment and its following action like turning left) and the number of combinations could be close to infinite. Besides, the representation of a route changes under different scenarios. These facts result in severe sparsity of routes, which increases the difficulty of route representation. Moreover, link attribute deficiencies and errors affect preciseness of route representation. Because of the sparsity of routes, the interaction data between users and routes are also sparse. This makes it not easy to acquire user representation from historical user-item interactions as traditional recommendations do. To address these issues, we propose a novel learning framework R4. In R4, we design a sparse & dense network to obtain representations of routes. The sparse unit learns link ID embeddings and aggregates them to represent a route, which captures implicit route characteristics and subsequently alleviates problems caused by link attribute deficiencies and errors. The dense unit extracts implicit local features of routes from link attributes. For user representation, we utilize a series of historical navigation to extract user preference. R4 achieves remarkable performance in both offline and online experiments.

Ning Xi, Chao Chen, Jun Zhang et al.

Mobile and IoT applications have greatly enriched our daily life by providing convenient and intelligent services. However, these smart applications have been a prime target of adversaries for stealing sensitive data. It poses a crucial threat to users' identity security, financial security, or even life security. Research communities and industries have proposed many Information Flow Control (IFC) techniques for data leakage detection and prevention, including secure modeling, type system, static analysis, dynamic analysis, \textit{etc}. According to the application's development life cycle, although most attacks are conducted during the application's execution phase, data leakage vulnerabilities have been introduced since the design phase. With a focus on lifecycle protection, this survey reviews the recent representative works adopted in different phases. We propose an information flow based defensive chain, which provides a new framework to systematically understand various IFC techniques for data leakage detection and prevention in Mobile and IoT applications. In line with the phases of the application life cycle, each reviewed work is comprehensively studied in terms of technique, performance, and limitation. Research challenges and future directions are also pointed out by consideration of the integrity of the defensive chain.

Li-Heng Chen, Christos G. Bampis, Zhi Li et al.

The layers of convolutional neural networks (CNNs) can be used to alter the resolution of their inputs, but the scaling factors are limited to integer values. However, in many image and video processing applications, the ability to resize by a fractional factor would be advantageous. One example is conversion between resolutions standardized for video compression, such as from 1080p to 720p. To solve this problem, we propose an alternative building block, formulated as a conventional convolutional layer followed by a differentiable resizer. More concretely, the convolutional layer preserves the resolution of the input, while the resizing operation is fully handled by the resizer. In this way, any CNN architecture can be adapted for non-integer resizing. As an application, we replace the resizing convolutional layer of a modern deep downsampling model by the proposed building block, and apply it to an adaptive bitrate video streaming scenario. Our experimental results show that an improvement in coding efficiency over the conventional Lanczos algorithm is attained, in terms of PSNR, SSIM, and VMAF on test videos.

Yifei Simon Shao, Chao Chen, Shreyas Kousik et al.

Reinforcement Learning (RL) algorithms have achieved remarkable performance in decision making and control tasks due to their ability to reason about long-term, cumulative reward using trial and error. However, during RL training, applying this trial-and-error approach to real-world robots operating in safety critical environment may lead to collisions. To address this challenge, this paper proposes a Reachability-based Trajectory Safeguard (RTS), which leverages reachability analysis to ensure safety during training and operation. Given a known (but uncertain) model of a robot, RTS precomputes a Forward Reachable Set of the robot tracking a continuum of parameterized trajectories. At runtime, the RL agent selects from this continuum in a receding-horizon way to control the robot; the FRS is used to identify if the agent's choice is safe or not, and to adjust unsafe choices. The efficacy of this method is illustrated on three nonlinear robot models, including a 12-D quadrotor drone, in simulation and in comparison with state-of-the-art safe motion planning methods.

Zhiyong Yu, Lei Han, Chao Chen et al.

Tracking a car or a person in a city is crucial for urban safety management. How can we complete the task with minimal number of spatiotemporal searches from massive camera records? This paper proposes a strategy named IHMs (Intermediate Searching at Heuristic Moments): each step we figure out which moment is the best to search according to a heuristic indicator, then at that moment search locations one by one in descending order of predicted appearing probabilities, until a search hits; iterate this step until we get the object's current location. Five searching strategies are compared in experiments, and IHMs is validated to be most efficient, which can save up to 1/3 total costs. This result provides an evidence that "searching at intermediate moments can save cost".

Bin Guo, Huihui Chen, Yan Liu et al.

With the development of mobile social networks, more and more crowdsourced data are generated on the Web or collected from real-world sensing. The fragment, heterogeneous, and noisy nature of online/offline crowdsourced data, however, makes it difficult to be understood. Traditional content-based analyzing methods suffer from potential issues such as computational intensiveness and poor performance. To address them, this paper presents CrowdMining. In particular, we observe that the knowledge hidden in the process of data generation, regarding individual/crowd behavior patterns (e.g., mobility patterns, community contexts such as social ties and structure) and crowd-object interaction patterns (flickering or tweeting patterns) are neglected in crowdsourced data mining. Therefore, a novel approach that leverages implicit human intelligence (implicit HI) for crowdsourced data mining and understanding is proposed. Two studies titled CrowdEvent and CrowdRoute are presented to showcase its usage, where implicit HIs are extracted either from online or offline crowdsourced data. A generic model for CrowdMining is further proposed based on a set of existing studies. Experiments based on real-world datasets demonstrate the effectiveness of CrowdMining.

Sheng Chen, Xu Wang, Chao Chen et al.

Modern neural network modules which can significantly enhance the learning power usually add too much computational complexity to the original neural networks. In this paper, we pursue very efficient neural network modules which can significantly boost the learning power of deep convolutional neural networks with negligible extra computational cost. We first present both theoretically and experimentally that elementwise square operator has a potential to enhance the learning power of neural networks. Then, we design four types of lightweight modules with elementwise square operators, named as Square-Pooling, Square-Softmin, Square-Excitation, and Square-Encoding. We add our four lightweight modules to Resnet18, Resnet50, and ShuffleNetV2 for better performance in the experiment on ImageNet 2012 dataset. The experimental results show that our modules can bring significant accuracy improvements to the base convolutional neural network models. The performance of our lightweight modules is even comparable to many complicated modules such as bilinear pooling, Squeeze-and-Excitation, and Gather-Excite. Our highly efficient modules are particularly suitable for mobile models. For example, when equipped with a single Square-Pooling module, the top-1 classification accuracy of ShuffleNetV2-0.5x on ImageNet 2012 is absolutely improved by 1.45% with no additional parameters and negligible inference time overhead.

Jiangtao Wang, Yasha Wang, Daqing Zhang et al.

With the advent of seamless connection of human, machine, and smart things, there is an emerging trend to leverage the power of crowds (e.g., citizens, mobile devices, and smart things) to monitor what is happening in a city, understand how the city is evolving, and further take actions to enable better quality of life, which is referred to as Crowd-Powered Smart City (CPSC). In this article, we provide a literature review for CPSC and identify future research opportunities. Specifically, we first define the concepts with typical CPSC applications. Then, we present the main characteristics of CPSC and further highlight the research issues. In the end, we point out existing limitations which can inform and guide future research directions.

Chao Chen, Yao-Chung Lin, Anil Kokaram et al.

HTTP video streaming is in wide use to deliver video over the Internet. With HTTP adaptive steaming, a video playback dynamically selects a video stream from a pre-encoded representation based on available bandwidth and viewport (screen) size. The viewer's video quality is therefore influenced by the encoded bitrates. We minimize the average delivered bitrate subject to a quality lower bound on a per-chunk basis by modeling the probability that a player selects a particular encoding. Through simulation and real-world experiments, the proposed method saves 9.6% of bandwidth while average delivered video quality comparing with state of the art while keeping average delivered video quality.

Bin Guo, Chao Chen, Daqing Zhang et al.

With the development of mobile sensing and mobile social networking techniques, Mobile Crowd Sensing and Computing (MCSC), which leverages heterogeneous crowdsourced data for large-scale sensing, has become a leading paradigm. Built on top of the participatory sensing vision, MCSC has two characterizing features: (1) it leverages heterogeneous crowdsourced data from two data sources: participatory sensing and participatory social media; and (2) it presents the fusion of human and machine intelligence (HMI) in both the sensing and computing process. This paper characterizes the unique features and challenges of MCSC. We further present early efforts on MCSC to demonstrate the benefits of aggregating heterogeneous crowdsourced data.

Chao Chen, Robert W. Heath, Alan C. Bovik et al.

We propose two scheduling algorithms that seek to optimize the quality of scalably coded videos that have been stored at a video server before transmission.} The first scheduling algorithm is derived from a Markov Decision Process (MDP) formulation developed here. We model the dynamics of the channel as a Markov chain and reduce the problem of dynamic video scheduling to a tractable Markov decision problem over a finite state space. Based on the MDP formulation, a near-optimal scheduling policy is computed that minimize the mean square error. Using insights taken from the development of the optimal MDP-based scheduling policy, the second proposed scheduling algorithm is an online scheduling method that only requires easily measurable knowledge of the channel dynamics, and is thus viable in practice. Simulation results show that the performance of both scheduling algorithms is close to a performance upper bound also derived in this paper.