Bin Li

Weijie Su, Xizhou Zhu, Chenxin Tao et al.

To effectively exploit the potential of large-scale models, various pre-training strategies supported by massive data from different sources are proposed, including supervised pre-training, weakly-supervised pre-training, and self-supervised pre-training. It has been proved that combining multiple pre-training strategies and data from various modalities/sources can greatly boost the training of large-scale models. However, current works adopt a multi-stage pre-training system, where the complex pipeline may increase the uncertainty and instability of the pre-training. It is thus desirable that these strategies can be integrated in a single-stage manner. In this paper, we first propose a general multi-modal mutual information formula as a unified optimization target and demonstrate that all existing approaches are special cases of our framework. Under this unified perspective, we propose an all-in-one single-stage pre-training approach, named Maximizing Multi-modal Mutual Information Pre-training (M3I Pre-training). Our approach achieves better performance than previous pre-training methods on various vision benchmarks, including ImageNet classification, COCO object detection, LVIS long-tailed object detection, and ADE20k semantic segmentation. Notably, we successfully pre-train a billion-level parameter image backbone and achieve state-of-the-art performance on various benchmarks. Code shall be released at https://github.com/OpenGVLab/M3I-Pretraining.

Shangchao Su, Mingzhao Yang, Bin Li et al.

Federated learning (FL) enables multiple clients to collaboratively train a global model without disclosing their data. Previous researches often require training the complete model parameters. However, the emergence of powerful pre-trained models makes it possible to achieve higher performance with fewer learnable parameters in FL. In this paper, we propose a federated adaptive prompt tuning algorithm, FedAPT, for multi-domain collaborative image classification with powerful foundation models, like CLIP. Compared with direct federated prompt tuning, our core idea is to adaptively unlock specific domain knowledge for each test sample in order to provide them with personalized prompts. To implement this idea, we design an adaptive prompt tuning module, which consists of a meta prompt, an adaptive network, and some keys. The server randomly generates a set of keys and assigns a unique key to each client. Then all clients cooperatively train the global adaptive network and meta prompt with the local datasets and the frozen keys. Ultimately, the global aggregation model can assign a personalized prompt to CLIP based on the domain features of each test sample. We perform extensive experiments on two multi-domain image classification datasets across two different settings -- supervised and unsupervised. The results show that FedAPT can achieve better performance with less than 10\% of the number of parameters of the fully trained model, and the global model can perform well in diverse client domains simultaneously. The source code is available at \url{https://github.com/leondada/FedAPT}.

Jie Wang, Zhihao Shi, Xize Liang et al.

Subgraph-wise sampling -- a promising class of mini-batch training techniques for graph neural networks (GNNs -- is critical for real-world applications. During the message passing (MP) in GNNs, subgraph-wise sampling methods discard messages outside the mini-batches in backward passes to avoid the well-known neighbor explosion problem, i.e., the exponentially increasing dependencies of nodes with the number of MP iterations. However, discarding messages may sacrifice the gradient estimation accuracy, posing significant challenges to their convergence analysis and convergence speeds. To address this challenge, we propose a novel subgraph-wise sampling method with a convergence guarantee, namely Local Message Compensation (LMC). To the best of our knowledge, LMC is the first subgraph-wise sampling method with provable convergence. The key idea is to retrieve the discarded messages in backward passes based on a message passing formulation of backward passes. By efficient and effective compensations for the discarded messages in both forward and backward passes, LMC computes accurate mini-batch gradients and thus accelerates convergence. Moreover, LMC is applicable to various MP-based GNN architectures, including convolutional GNNs (finite message passing iterations with different layers) and recurrent GNNs (infinite message passing iterations with a shared layer). Experiments on large-scale benchmarks demonstrate that LMC is significantly faster than state-of-the-art subgraph-wise sampling methods.

Tonglin Chen, Bin Li, Zhimeng Shen et al.

The appearance of the same object may vary in different scene images due to perspectives and occlusions between objects. Humans can easily identify the same object, even if occlusions exist, by completing the occluded parts based on its canonical image in the memory. Achieving this ability is still a challenge for machine learning, especially under the unsupervised learning setting. Inspired by such an ability of humans, this paper proposes a compositional scene modeling method to infer global representations of canonical images of objects without any supervision. The representation of each object is divided into an intrinsic part, which characterizes globally invariant information (i.e. canonical representation of an object), and an extrinsic part, which characterizes scene-dependent information (e.g., position and size). To infer the intrinsic representation of each object, we employ a patch-matching strategy to align the representation of a potentially occluded object with the canonical representations of objects, and sample the most probable canonical representation based on the category of object determined by amortized variational inference. Extensive experiments are conducted on four object-centric learning benchmarks, and experimental results demonstrate that the proposed method not only outperforms state-of-the-arts in terms of segmentation and reconstruction, but also achieves good global object identification performance.

Weisi Guo, Schyler Sun, Bin Li et al.

Deep Learning (DL) is penetrating into a diverse range of mass mobility, smart living, and industrial applications, rapidly transforming the way we live and work. DL is at the heart of many AI implementations. A key set of challenges is to produce AI modules that are: (1) "circular" - can solve new tasks without forgetting how to solve previous ones, (2) "secure" - have immunity to adversarial data attacks, and (3) "tiny" - implementable in low power low cost embedded hardware. Clearly it is difficult to achieve all three aspects on a single horizontal layer of platforms, as the techniques require transformed deep representations that incur different computation and communication requirements. Here we set out the vision to achieve transformed DL representations across a 5G and Beyond networked architecture. We first detail the cross-sectoral motivations for each challenge area, before demonstrating recent advances in DL research that can achieve circular, secure, and tiny AI (CST-AI). Recognising the conflicting demand of each transformed deep representation, we federate their deep learning transformations and functionalities across the network to achieve connected run-time capabilities.

Jinyang Yuan, Tonglin Chen, Bin Li et al.

Visual scenes are composed of visual concepts and have the property of combinatorial explosion. An important reason for humans to efficiently learn from diverse visual scenes is the ability of compositional perception, and it is desirable for artificial intelligence to have similar abilities. Compositional scene representation learning is a task that enables such abilities. In recent years, various methods have been proposed to apply deep neural networks, which have been proven to be advantageous in representation learning, to learn compositional scene representations via reconstruction, advancing this research direction into the deep learning era. Learning via reconstruction is advantageous because it may utilize massive unlabeled data and avoid costly and laborious data annotation. In this survey, we first outline the current progress on reconstruction-based compositional scene representation learning with deep neural networks, including development history and categorizations of existing methods from the perspectives of the modeling of visual scenes and the inference of scene representations; then provide benchmarks, including an open source toolbox to reproduce the benchmark experiments, of representative methods that consider the most extensively studied problem setting and form the foundation for other methods; and finally discuss the limitations of existing methods and future directions of this research topic.

Xizhou Zhu, Weijie Su, Lewei Lu et al.

DETR has been recently proposed to eliminate the need for many hand-designed components in object detection while demonstrating good performance. However, it suffers from slow convergence and limited feature spatial resolution, due to the limitation of Transformer attention modules in processing image feature maps. To mitigate these issues, we proposed Deformable DETR, whose attention modules only attend to a small set of key sampling points around a reference. Deformable DETR can achieve better performance than DETR (especially on small objects) with 10 times less training epochs. Extensive experiments on the COCO benchmark demonstrate the effectiveness of our approach. Code is released at https://github.com/fundamentalvision/Deformable-DETR.

Hong Wang, Zhongkai Hao, Jie Wang et al.

Learning neural operators for solving partial differential equations (PDEs) has attracted great attention due to its high inference efficiency. However, training such operators requires generating a substantial amount of labeled data, i.e., PDE problems together with their solutions. The data generation process is exceptionally time-consuming, as it involves solving numerous systems of linear equations to obtain numerical solutions to the PDEs. Many existing methods solve these systems independently without considering their inherent similarities, resulting in extremely redundant computations. To tackle this problem, we propose a novel method, namely Sorting Krylov Recycling (SKR), to boost the efficiency of solving these systems, thus significantly accelerating data generation for neural operators training. To the best of our knowledge, SKR is the first attempt to address the time-consuming nature of data generation for learning neural operators. The working horse of SKR is Krylov subspace recycling, a powerful technique for solving a series of interrelated systems by leveraging their inherent similarities. Specifically, SKR employs a sorting algorithm to arrange these systems in a sequence, where adjacent systems exhibit high similarities. Then it equips a solver with Krylov subspace recycling to solve the systems sequentially instead of independently, thus effectively enhancing the solving efficiency. Both theoretical analysis and extensive experiments demonstrate that SKR can significantly accelerate neural operator data generation, achieving a remarkable speedup of up to 13.9 times.

Hongyu Liu, Haoyang Liu, Yufei Kuang et al.

Combinatorial optimization (CO) is one of the most fundamental mathematical models in real-world applications. Traditional CO solvers, such as Branch-and-Bound (B&B) solvers, heavily rely on expert-designed heuristics, which are reliable but require substantial manual tuning. Recent studies have leveraged deep learning (DL) models as an alternative to capture rich feature patterns for improved performance on GPU machines. Nonetheless, the drawbacks of high training and inference costs, as well as limited interpretability, severely hinder the adoption of DL methods in real-world applications. To address these challenges, we propose a novel deep symbolic optimization learning framework that combines their advantages. Specifically, we focus on the node selection module within B&B solvers -- namely, deep symbolic optimization for node selection (Dso4NS). With data-driven approaches, Dso4NS guides the search for mathematical expressions within the high-dimensional discrete symbolic space and then incorporates the highest-performing mathematical expressions into a solver. The data-driven model captures the rich feature information in the input data and generates symbolic expressions, while the expressions deployed in solvers enable fast inference with high interpretability. Experiments demonstrate the effectiveness of Dso4NS in learning high-quality expressions, outperforming existing approaches on a CPU machine. Encouragingly, the learned CPU-based policies consistently achieve performance comparable to state-of-the-art GPU-based approaches.

Xihua Sheng, Jiahao Li, Bin Li et al.

We address end-to-end learned video compression with a special focus on better learning and utilizing temporal contexts. For temporal context mining, we propose to store not only the previously reconstructed frames, but also the propagated features into the generalized decoded picture buffer. From the stored propagated features, we propose to learn multi-scale temporal contexts, and re-fill the learned temporal contexts into the modules of our compression scheme, including the contextual encoder-decoder, the frame generator, and the temporal context encoder. Our scheme discards the parallelization-unfriendly auto-regressive entropy model to pursue a more practical decoding time. We compare our scheme with x264 and x265 (representing industrial software for H.264 and H.265, respectively) as well as the official reference software for H.264, H.265, and H.266 (JM, HM, and VTM, respectively). When intra period is 32 and oriented to PSNR, our scheme outperforms H.265--HM by 14.4% bit rate saving; when oriented to MS-SSIM, our scheme outperforms H.266--VTM by 21.1% bit rate saving.

Zhiming Liu, Xuefei Zhang, Chongyang Liu et al.

Object detection is a basic and important task in the field of aerial image processing and has gained much attention in computer vision. However, previous aerial image object detection approaches have insufficient use of scene semantic information between different regions of large-scale aerial images. In addition, complex background and scale changes make it difficult to improve detection accuracy. To address these issues, we propose a relationship representation network for object detection in aerial images (RelationRS): 1) Firstly, multi-scale features are fused and enhanced by a dual relationship module (DRM) with conditional convolution. The dual relationship module learns the potential relationship between features of different scales and learns the relationship between different scenes from different patches in a same iteration. In addition, the dual relationship module dynamically generates parameters to guide the fusion of multi-scale features. 2) Secondly, The bridging visual representations module (BVR) is introduced into the field of aerial images to improve the object detection effect in images with complex backgrounds. Experiments with a publicly available object detection dataset for aerial images demonstrate that the proposed RelationRS achieves a state-of-the-art detection performance.

Zhuangkun Wei, Liang Wang, Schyler Chengyao Sun et al.

The proliferation of low-cost Internet of Things (IoT) devices has led to a race between wireless security and channel attacks. Traditional cryptography requires high-computational power and is not suitable for low-power IoT scenarios. Whist, recently developed physical layer security (PLS) can exploit common wireless channel state information (CSI), its sensitivity to channel estimation makes them vulnerable from attacks. In this work, we exploit an alternative common physics shared between IoT transceivers: the monitored channel-irrelevant physical networked dynamics (e.g., water/oil/gas/electrical signal-flows). Leveraging this, we propose for the first time, graph layer security (GLS), by exploiting the dependency in physical dynamics among network nodes for information encryption and decryption. A graph Fourier transform (GFT) operator is used to characterize such dependency into a graph-bandlimted subspace, which allows the generations of channel-irrelevant cipher keys by maximizing the secrecy rate. We evaluate our GLS against designed active and passive attackers, using IEEE 39-Bus system. Results demonstrate that, GLS is not reliant on wireless CSI, and can combat attackers that have partial networked dynamic knowledge (realistic access to full dynamic and critical nodes remains challenging). We believe this novel GLS has widespread applicability in secure health monitoring and for Digital Twins in adversarial radio environments.

Bin Li, Hu Cao, Zhongnan Qu et al.

Robotic grasping plays an important role in the field of robotics. The current state-of-the-art robotic grasping detection systems are usually built on the conventional vision, such as RGB-D camera. Compared to traditional frame-based computer vision, neuromorphic vision is a small and young community of research. Currently, there are limited event-based datasets due to the troublesome annotation of the asynchronous event stream. Annotating large scale vision dataset often takes lots of computation resources, especially the troublesome data for video-level annotation. In this work, we consider the problem of detecting robotic grasps in a moving camera view of a scene containing objects. To obtain more agile robotic perception, a neuromorphic vision sensor (DAVIS) attaching to the robot gripper is introduced to explore the potential usage in grasping detection. We construct a robotic grasping dataset named Event-Stream Dataset with 91 objects. A spatio-temporal mixed particle filter (SMP Filter) is proposed to track the led-based grasp rectangles which enables video-level annotation of a single grasp rectangle per object. As leds blink at high frequency, the Event-Stream dataset is annotated in a high frequency of 1 kHz. Based on the Event-Stream dataset, we develop a deep neural network for grasping detection which consider the angle learning problem as classification instead of regression. The method performs high detection accuracy on our Event-Stream dataset with 93% precision at object-wise level. This work provides a large-scale and well-annotated dataset, and promotes the neuromorphic vision applications in agile robot.

Zhuangkun Wei, Weisi Guo, Bin Li et al.

In emerging Internet-of-Nano-Thing (IoNT), information will be embedded and conveyed in the form of molecules through complex and diffusive medias. One main challenge lies in the long-tail nature of the channel response causing inter-symbol-interference (ISI), which deteriorates the detection performance. If the channel is unknown, we cannot easily achieve traditional coherent channel estimation and cancellation, and the impact of ISI will be more severe. In this paper, we develop a novel high-dimensional non-coherent scheme for blind detection of molecular signals. We achieve this in a higher-dimensional metric space by combining different non-coherent metrics that exploit the transient features of the signals. By deducing the theoretical bit error rate (BER) for any constructed high-dimensional non-coherent metric, we prove that, higher dimensionality always achieves a lower BER in the same sample space. Then, we design a generalised blind detection algorithm that utilizes the Parzen approximation and its probabilistic neural network (Parzen-PNN) to detect information bits. Taking advantages of its fast convergence and parallel implementation, our proposed scheme can meet the needs of detection accuracy and real-time computing. Numerical simulations demonstrate that our proposed scheme can gain 10dB BER compared with other state of the art methods.

Zhuangkun Wei, Bin Li, Weisi Guo et al.

Extracting and detecting spike activities from the fluorescence observations is an important step in understanding how neuron systems work. The main challenge lies in that the combination of the ambient noise with dynamic baseline fluctuation, often contaminates the observations, thereby deteriorating the reliability of spike detection. This may be even worse in the face of the nonlinear biological process, the coupling interactions between spikes and baseline, and the unknown critical parameters of an underlying physiological model, in which erroneous estimations of parameters will affect the detection of spikes causing further error propagation. In this paper, we propose a random finite set (RFS) based Bayesian approach. The dynamic behaviors of spike sequence, fluctuated baseline and unknown parameters are formulated as one RFS. This RFS state is capable of distinguishing the hidden active/silent states induced by spike and non-spike activities respectively, thereby \emph{negating the interaction role} played by spikes and other factors. Then, premised on the RFS states, a Bayesian inference scheme is designed to simultaneously estimate the model parameters, baseline, and crucial spike activities. Our results demonstrate that the proposed scheme can gain an extra $12\%$ detection accuracy in comparison with the state-of-the-art MLSpike method.

Chaoqiong Fan, Bin Li, Jia Hou et al.

In this paper, we consider a mesh-structured unmanned aerial vehicle (UAV) networks exploiting partially overlapping channels (POCs). For general data-collection tasks in UAV networks, we aim to optimize the network throughput with constraints on transmission power and quality of service (QoS). As far as the highly mobile and constantly changing UAV networks are concerned, unfortunately, most existing methods rely on definite information which is vulnerable to the dynamic environment, rendering system performance to be less effective. In order to combat dynamic topology and varying interference of UAV networks, a robust and distributed learning scheme is proposed. Rather than the perfect channel state information (CSI), we introduce uncertainties to characterize the dynamic channel gains among UAV nodes, which are then interpreted with fuzzy numbers. Instead of the traditional observation space where the channel capacity is a crisp reward, we implement the learning and decision process in a mapped fuzzy space. This allows the system to achieve a smoother and more robust performance by optimizing in an alternate space. To this end, we design a fuzzy payoffs function (FPF) to describe the fluctuated utility, and the problem of POCs assignment is formulated as a fuzzy payoffs game (FPG). Assisted by an attractive property of fuzzy bi-matrix games, the existence of fuzzy Nash equilibrium (FNE) for our formulated FPG is proved. Our robust fuzzy-learning algorithm could reach the equilibrium solution via a least-deviation method. Finally, numerical simulations are provided to demonstrate the advantages of our new scheme over the existing scheme.

Bin Li, Yueheng Lan, Weisi Guo et al.

Detecting anomalies and discovering driving signals is an essential component of scientific research and industrial practice. Often the underlying mechanism is highly complex, involving hidden evolving nonlinear dynamics and noise contamination. When representative physical models and large labeled data sets are unavailable, as is the case with most real-world applications, model-dependent Bayesian approaches would yield misleading results, and most supervised learning machines would also fail to reliably resolve the intricately evolving systems. Here, we propose an unsupervised machine-learning approach that operates in a well-constructed function space, whereby the evolving nonlinear dynamics are captured through a linear functional representation determined by the Koopman operator. This breakthrough leverages on the time-feature embedding and the ensuing reconstruction of a phase-space representation of the dynamics, thereby permitting the reliable identification of critical global signatures from the whole trajectory. This dramatically improves over commonly used static local features, which are vulnerable to unknown transitions or noise. Thanks to its data-driven nature, our method excludes any prior models and training corpus. We benchmark the astonishing accuracy of our method on three diverse and challenging problems in: biology, medicine, and engineering. In all cases, it outperforms existing state-of-the-art methods. As a new unsupervised information processing paradigm, it is suitable for ubiquitous nonlinear dynamical systems or end-users with little expertise, which permits an unbiased excavation of underlying working principles or intrinsic correlations submerged in unlabeled data flows.