论文检索

Deep Learning · Sequential Models, Time series

Zhijian Xu, Ailing Zeng, Qiang Xu

In this paper, we introduce FITS, a lightweight yet powerful model for time series analysis. Unlike existing models that directly process raw time-domain data, FITS operates on the principle that time series can be manipulated through interpolation in the complex frequency domain, achieving performance comparable to state-of-the-art models for time series forecasting and anomaly detection tasks. Notably, FITS accomplishes this with a svelte profile of just about $10k$ parameters, making it ideally suited for edge devices and paving the way for a wide range of applications. The code is available for review at: \url{https://anonymous.4open.science/r/FITS}.

Dongyoon Han, Junsuk Choe, Seonghyeok Chun, John Joon Young Chung, Minsuk Chang, Sangdoo Yun, Jean Y. Song, Seong Joon Oh

Supervised learning of image classifiers distills human knowledge into a parametric model through pairs of images and corresponding labels (X,Y). We argue that this simple and widely used representation of human knowledge neglects rich auxiliary information from the annotation procedure, such as the time-series of mouse traces and clicks left after image selection. Our insight is that such annotation byproducts Z provide approximate human attention that weakly guides the model to focus on the foreground cues, reducing spurious correlations and discouraging shortcut learning. To verify this, we create ImageNet-AB and COCO-AB. They are ImageNet and COCO training sets enriched with sample-wise annotation byproducts, collected by replicating the respective original annotation tasks. We refer to the new paradigm of training models with annotation byproducts as learning using annotation byproducts (LUAB). We show that a simple multitask loss for regressing Z together with Y already improves the generalisability and robustness of the learned models. Compared to the original supervised learning, LUAB does not require extra annotation costs. ImageNet-AB and COCO-AB are at https://github.com/naver-ai/NeglectedFreeLunch.

Bingchen Zhao, Oisin Mac Aodha

We explore the problem of Incremental Generalized Category Discovery (IGCD). This is a challenging category-incremental learning setting where the goal is to develop models that can correctly categorize images from previously seen categories, in addition to discovering novel ones. Learning is performed over a series of time steps where the model obtains new labeled and unlabeled data, and discards old data, at each iteration. The difficulty of the problem is compounded in our generalized setting as the unlabeled data can contain images from categories that may or may not have been observed before. We present a new model for IGCD which combines non-parametric categorization with efficient image sampling to mitigate catastrophic forgetting. To quantify performance, we propose a new benchmark dataset named iNatIGCD that is motivated by a real-world fine-grained visual categorization task. In our experiments we outperform existing related methods.

Enze Ye, Yuhang Wang, Hong Zhang, Yiqin Gao, Huan Wang, He Sun

The dynamics of biomolecules are crucial for our understanding of their functioning in living systems. However, current 3D imaging techniques, such as cryogenic electron microscopy (cryo-EM), require freezing the sample, which limits the observation of their conformational changes in real time. The innovative liquid-phase electron microscopy (liquid-phase EM) technique allows molecules to be placed in the native liquid environment, providing a unique opportunity to observe their dynamics. In this paper, we propose TEMPOR, a Temporal Electron MicroscoPy Object Reconstruction algorithm for liquid-phase EM that leverages an implicit neural representation (INR) and a dynamical variational auto-encoder (DVAE) to recover time series of molecular structures. We demonstrate its advantages in recovering different motion dynamics from two simulated datasets, 7bcq and Cas9. To our knowledge, our work is the first attempt to directly recover 3D structures of a temporally-varying particle from liquid-phase EM movies. It provides a promising new approach for studying molecules' 3D dynamics in structural biology.

Yao Teng, Haisong Liu, Sheng Guo, Limin Wang

Previous object detectors make predictions based on dense grid points or numerous preset anchors. Most of these detectors are trained with one-to-many label assignment strategies. On the contrary, recent query-based object detectors are based a sparse set of learnable queries refined by a series of decoder layers. The one-to-one label assignment is independently applied on each layer for deep supervision during training. Despite the great success of query-based object detection, however, this vanilla one-to-one label assignment strategy requires the detectors to have strong fine-grained discrimination and modeling capacity. In this paper, we propose a new query-based object detector with cross-stage interaction, coined as StageInteractor. During the forward pass, we come up with an efficient way to improve this modeling ability by reusing dynamic operators with lightweight adapters. As for the label assignment, a cross-stage label assigner is designed to improve the one-to-one label assignment. With this assigner, the training target class labels are gathered across stages and then reallocated to proper predictions at each decoder layer. On MS COCO benchmark, our model improves the baseline counterpart by 2.2 AP, and achieves a 44.8 AP with ResNet-50 as backbone, 100 queries and 12 training epochs. With longer training time and 300 queries, StageInteractor achieves 51.3 AP and 52.7 AP with ResNeXt-101-DCN and Swin-S, respectively. The code and models are made available at https://github.com/MCG-NJU/StageInteractor.

Foivos Paraperas Papantoniou, Alexandros Lattas, Stylianos Moschoglou, Stefanos Zafeiriou

Following the remarkable success of diffusion models on image generation, recent works have also demonstrated their impressive ability to address a number of inverse problems in an unsupervised way, by properly constraining the sampling process based on a conditioning input. Motivated by this, in this paper, we present the first approach to use diffusion models as a prior for highly accurate 3D facial BRDF reconstruction from a single image. We start by leveraging a high-quality UV dataset of facial reflectance (diffuse and specular albedo and normals), which we render under varying illumination settings to simulate natural RGB textures and, then, train an unconditional diffusion model on concatenated pairs of rendered textures and reflectance components. At test time, we fit a 3D morphable model to the given image and unwrap the face in a partial UV texture. By sampling from the diffusion model, while retaining the observed texture part intact, the model inpaints not only the self-occluded areas but also the unknown reflectance components, in a single sequence of denoising steps. In contrast to existing methods, we directly acquire the observed texture from the input image, thus, resulting in more faithful and consistent reflectance estimation. Through a series of qualitative and quantitative comparisons, we demonstrate superior performance in both texture completion as well as reflectance reconstruction tasks.

Abdelrahman Shaker, Muhammad Maaz, Hanoona Rasheed, Salman Khan, Ming-Hsuan Yang, Fahad Shahbaz Khan

Self-attention has become a defacto choice for capturing global context in various vision applications. However, its quadratic computational complexity with respect to image resolution limits its use in real-time applications, especially for deployment on resource-constrained mobile devices. Although hybrid approaches have been proposed to combine the advantages of convolutions and self-attention for a better speed-accuracy trade-off, the expensive matrix multiplication operations in self-attention remain a bottleneck. In this work, we introduce a novel efficient additive attention mechanism that effectively replaces the quadratic matrix multiplication operations with linear element-wise multiplications. Our design shows that the key-value interaction can be replaced with a linear layer without sacrificing any accuracy. Unlike previous state-of-the-art methods, our efficient formulation of self-attention enables its usage at all stages of the network. Using our proposed efficient additive attention, we build a series of models called "SwiftFormer" which achieves state-of-the-art performance in terms of both accuracy and mobile inference speed. Our small variant achieves 78.5% top-1 ImageNet-1K accuracy with only 0.8 ms latency on iPhone 14, which is more accurate and 2x faster compared to MobileViT-v2. Our code and models: https://tinyurl.com/5ft8v46w

Favyen Bastani, Piper Wolters, Ritwik Gupta, Joe Ferdinando, Aniruddha Kembhavi

Remote sensing images are useful for a wide variety of planet monitoring applications, from tracking deforestation to tackling illegal fishing. The Earth is extremely diverse---the amount of potential tasks in remote sensing images is massive, and the sizes of features range from several kilometers to just tens of centimeters. However, creating generalizable computer vision methods is a challenge in part due to the lack of a large-scale dataset that captures these diverse features for many tasks. In this paper, we present SatlasPretrain, a remote sensing dataset that is large in both breadth and scale, combining Sentinel-2 and NAIP images with 302M labels under 137 categories and seven label types. We evaluate eight baselines and a proposed method on SatlasPretrain, and find that there is substantial room for improvement in addressing research challenges specific to remote sensing, including processing image time series that consist of images from very different types of sensors, and taking advantage of long-range spatial context. Moreover, we find that pre-training on SatlasPretrain substantially improves performance on downstream tasks, increasing average accuracy by 18% over ImageNet and 6% over the next best baseline. The dataset, pre-trained model weights, and code are available at https://satlas-pretrain.allen.ai/.

Huiyang Shao, Qianqian Xu, Peisong Wen, Peifeng Gao, Zhiyong Yang, Qingming Huang

In this paper, we focus on the mural-restoration task, which aims to detect damaged regions in the mural and repaint them automatically. Different from traditional image restoration tasks like in/out/blind-painting and image renovation, the corrupted mural suffers from more complicated degradation. However, existing mural-restoration methods and datasets still focus on simple degradation like masking. Such a significant gap prevents mural-restoration from being applied to real scenarios. To fill this gap, in this work, we propose a systematic framework to simulate the physical process for damaged murals and provide a new benchmark dataset for mural-restoration. Limited by the simplification of the data synthesis process, the previous mural-restoration methods suffer from poor performance in our proposed dataset. To handle this problem, we propose the Attention Diffusion Framework (ADF) for this challenging task. Within the framework, a damage attention map module is proposed to estimate the damage extent. Facing the diversity of defects, we propose a series of loss functions to choose repair strategies adaptively. Finally, experimental results support the effectiveness of the proposed framework in terms of both mural synthesis and restoration.

Jinhong Wang, Yi Cheng, Jintai Chen, TingTing Chen, Danny Chen, Jian Wu

Ordinal regression refers to classifying object instances into ordinal categories. It has been widely studied in many scenarios, such as medical disease grading and movie rating. Known methods focused only on learning inter-class ordinal relationships, but still incur limitations in distinguishing adjacent categories thus far. In this paper, we propose a simple sequence prediction framework for ordinal regression called Ord2Seq, which, for the first time, transforms each ordinal category label into a special label sequence and thus regards an ordinal regression task as a sequence prediction process. In this way, we decompose an ordinal regression task into a series of recursive binary classification steps, so as to subtly distinguish adjacent categories. Comprehensive experiments show the effectiveness of distinguishing adjacent categories for performance improvement and our new approach exceeds state-of-the-art performances in four different scenarios. Codes are available at https://github.com/wjh892521292/Ord2Seq.

Zhuangzhuang Chen, Jin Zhang, Zhuonan Lai, Guanming Zhu, Zun Liu, Jie Chen, Jianqiang Li

Cracks are usually curve-like structures that are the focus of many computer-vision applications (e.g., road safety inspection and surface inspection of industrial facilities). The existing pixel-based crack segmentation methods rely on time-consuming and costly pixel-level annotations. And the object-based crack detection methods exploit the horizontal box to detect the crack without considering crack orientation, resulting in scale variation and intra-class variation. Considering this, we provide a new perspective for crack detection that models the cracks as a series of sub-cracks with the corresponding orientation. However, the vanilla adaptation of the existing oriented object detection methods to the crack detection tasks will result in limited performance, due to the boundary discontinuity issue and the ambiguities in sub-crack orientation. In this paper, we propose a first-of-its-kind oriented sub-crack detector, dubbed as CrackDet, which is derived from a novel piecewise angle definition, to ease the boundary discontinuity problem. And then, we propose a multi-branch angle regression loss for learning sub-crack orientation and variance together. Since there are no related benchmarks, we construct three fully annotated datasets, namely, ORC, ONPP, and OCCSD, which involve various cracks in road pavement and industrial facilities. Experiments show that our approach outperforms state-of-the-art crack detectors.

Jiye Lee, Hanbyul Joo

Synthesizing interaction-involved human motions has been challenging due to the high complexity of 3D environments and the diversity of possible human behaviors within. We present LAMA, Locomotion-Action-MAnipulation, to synthesize natural and plausible long term human movements in complex indoor environments. The key motivation of LAMA is to build a unified framework to encompass a series of everyday motions including locomotion, scene interaction, and object manipulation. Unlike existing methods that require motion data "paired" with scanned 3D scenes for supervision, we formulate the problem as a test-time optimization by using human motion capture data only for synthesis. LAMA leverages a reinforcement learning framework coupled with motion matching algorithm for optimization, and further exploits a motion editing framework via manifold learning to cover possible variations in interaction and manipulation. Throughout extensive experiments, we demonstrate that LAMA outperforms previous approaches in synthesizing realistic motions in various challenging scenarios.

Kieran Saunders, George Vogiatzis, Luis J. Manso

Current, self-supervised depth estimation architectures rely on clear and sunny weather scenes to train deep neural networks. However, in many locations, this assumption is too strong. For example in the UK (2021), 149 days consisted of rain. For these architectures to be effective in real-world applications, we must create models that can generalise to all weather conditions, times of the day and image qualities. Using a combination of computer graphics and generative models, one can augment existing sunny-weather data in a variety of ways that simulate adverse weather effects. While it is tempting to use such data augmentations for self-supervised depth, in the past this was shown to degrade performance instead of improving it. In this paper, we put forward a method that uses augmentations to remedy this problem. By exploiting the correspondence between unaugmented and augmented data we introduce a pseudo-supervised loss for both depth and pose estimation. This brings back some of the benefits of supervised learning while still not requiring any labels. We also make a series of practical recommendations which collectively offer a reliable, efficient framework for weather-related augmentation of self-supervised depth from monocular video. We present extensive testing to show that our method, Robust-Depth, achieves SotA performance on the KITTI dataset while significantly surpassing SotA on challenging, adverse condition data such as DrivingStereo, Foggy CityScape and NuScenes-Night. The project website can be found at https://kieran514.github.io/Robust-Depth-Project/.

Shijian Mao, Wudong Xi, Lei Yu, Gaotian Lü, Xingxing Xing, Xingchen Zhou, Wei Wan

The Social Media Prediction (SMP) challenge aims to predict the future popularity of online posts by leveraging social media data. Social media data contains multimodal information, such as text, images, time series, etc. Previous methods have proposed many feature extraction and feature construction methods to represent these multimodal information, thereby predicting the popularity of posts. Despite the success of previous methods in extracting features from social media data, these features tend to be predominantly lower-order, posing a challenge in accurately capturing the rich information contained in text and images. In this paper, we propose a more diverse feature mining method and introduce a stacking block module to capture higher-order feature information contained in text and images. "lower-order" refers to the original high-dimensional embedding representation, while "high-order" pertains to the impact on post social popularity captured by tree models from text or image. We conducted massive experiments to evaluate the effectiveness of our proposed method and found that the stacking block module significantly improved performance.

Zheng Zhang 0006, Songling Chen, Mixiao Hou, Guangming Lu 0002

In this paper, we present a multimodal emotion analysis platform, which can flexibly capture, detect and analyze the emotions of video object with multiple modalities under different situations, including offline and online application scenarios. This system can visualize the dynamic effects of different types of emotions from both multimodal and unimodal circumstances. The presented emotion analysis results show instant and time series states in both specific modality and multiple modalities. Our system fills the current research and application gaps in multimodal emotion analysis with an interactive interface. Notably, the constructed system can adaptively process pre-recorded video clips as well as collected real-world data with excellent practicality and interactivity.

Kuan Tian, Yonghang Guan, Jinxi Xiang, Jun Zhang 0018, Xiao Han 0011, Wei Yang 0032

The state-of-the-art neural video codecs have outperformed the most sophisticated traditional codecs in terms of rate-distortion (RD) performance in certain cases. However, utilizing them for practical applications is still challenging for two major reasons. 1) Cross-platform computational errors resulting from floating point operations can lead to inaccurate decoding of the bitstream. 2) The high computational complexity of the encoding and decoding process poses a challenge in achieving real-time performance. In this paper, we propose a real-time cross-platform neural video codec, which is capable of efficiently decoding (25FPS) of 720P video bitstream from other encoding platforms on a consumer-grade GPU (e.g., NVIDIA RTX 2080). First, to solve the problem of inconsistency of codec caused by the uncertainty of floating point calculations across platforms, we design a calibration transmitting system to guarantee the consistent quantization of entropy parameters between the encoding and decoding stages. The parameters that may have transboundary quantization between encoding and decoding are identified in the encoding stage, and their coordinates will be delivered by auxiliary transmitted bitstream. By doing so, these inconsistent parameters can be processed properly in the decoding stage. Furthermore, to reduce the bitrate of the auxiliary bitstream, we rectify the distribution of entropy parameters using a piecewise Gaussian constraint. Second, to match the computational limitations on the decoding side for real-time video codec, we design a lightweight model. A series of efficiency techniques, such as model pruning, motion downsampling, and arithmetic coding skipping, enable our model to achieve 25 FPS decoding speed on NVIDIA RTX 2080 GPU. Experimental results demonstrate that our model can achieve real-time decoding of 720P videos while encoding on another platform. Furthermore, the real-time model brings up to a maximum of 24.2% BD-rate improvement from the perspective of PSNR with the anchor H.265 (medium).

Jinkang Guo, Zhibo Wan, Zhihan Lv

Currently, people are actively seeking new and clean energy sources to replace traditional fossil fuels. As a widely distributed, abundant, and easily accessible renewable energy source, wind energy has multiple applications both domestically and internationally, and has a promising future. In wind power forecasting, this paper proposes a new algorithm LFTformer that combines the Transformer model with linear fuzzy information granulation (LFIG).The model uses the improved LFIG algorithm to extract the semantic information of time series, and divides the original time series into multiple information granules, which are then used as inputs to the Transformer model. Through comparative experiments for wind power prediction over 72 hours, 120 hours, and 168 hours, this research demonstrates that the LFTformer can improve the accuracy of wind power prediction. Finally, a Digital Twins system for wind power prediction is constructed using three-dimensional visualization techniques, displaying the prediction results and key data indicators, providing reliable data support for wind power systems.

Tongjie Pan, Yalan Ye, Hecheng Cai, Shudong Huang, Yang Yang 0002, Guoqing Wang 0001

Multimodal physiological-based emotion recognition is one of the most available but challenging studies due to complexity of emotions and individual differences in physiological signals. However, existing studies mainly combine multimodal data to fuse multimodal information in offline scenarios, ignoring data/modalities correlation among multimodal data and individual differences of non-stationary physiological signals in online scenarios. In this paper, we propose a novel Online Multimodal HyperGraph Learning (OMHGL) method to fuse multimodal information for emotion recognition based on time-series physiological signals. Our method consists of multimodal hypergraph fusion and online hypergraph learning. Specifically, the multimodal hypergraph fusion can fuse multimodal physiological signals to effectively obtain emotionally dependent information via leveraging multimodal information and higher-order correlations among multimodal data/modalities. The online hypergraph learning is designed to learn new information from online data by updating hypergraph projection. As a result, the proposed online emotion recognition model can be more effective for emotion recognition of target subjects when target data arrive in an online manner. Experimental results have demonstrated that the proposed method significantly outperforms the baselines and compared state-of-the-art methods in online emotion recognition tasks.

Wei Liu 0148, Xinlei Yang, Zhenhua Li 0001, Feng Qian 0001

When streaming over the web, correlated videos (e.g., a series of TV episodes) appear to bear considerable redundant clips, mostly included in the intros, outros, recaps, and commercial breaks, leading to a waste of network traffic and playback time. Mainstream video content providers have taken various measures to identify these clips, but often result in unexpected and undesirable user experiences. In this paper, we conduct a large-scale, crowdsourced study to demystify the root causes of poor experiences. Driven by the findings, we propose to reconsider the problem from a novel perspective of scenes without going through the excessive video frames, which pays special attention to how the contents of correlated videos are organized during video production. To enable this idea, we design efficient approaches to the separation of video scenes and the identification of visual redundancy. We build an open-source system to embody our design, which achieves fast (e.g., taking ~38 seconds to process a 45-minute video using a common commodity server) and accurate (incurring only 770-ms deviation on average) redundancy recognition on representative workloads.

Bowei Xu, Hao Chen 0036, Zhan Ma 0001

Optimal adaptive bitrate (ABR) decision depends on a comprehensive characterization of state transitions that involve interrelated modalities over time including environmental observations, returns, and actions. However, state-of-the-art learning-based ABR algorithms solely rely on past observations to decide the next action. This paradigm tends to cause a chain of deviations from optimal action when encountering unfamiliar observations, which consequently undermines the model generalization. This paper presents Karma, an ABR algorithm that utilizes causal sequence modeling to improve generalization by comprehending the interrelated causality among past observations, returns, and actions and timely refining action when deviation occurs. Unlike direct observation-to-action mapping, Karma recurrently maintains a multi-dimensional time series of observations, returns, and actions as input and employs causal sequence modeling via a decision transformer to determine the next action. In the input sequence, Karma uses the maximum cumulative future quality of experience (QoE) (a.k.a, QoE-to-go) as an extended return signal, which is periodically estimated based on current network conditions and playback status. We evaluate Karma through trace-driven simulations and real-world field tests, demonstrating superior performance compared to existing state-of-the-art ABR algorithms, with an average QoE improvement ranging from 10.8% to 18.7% across diverse network conditions. Furthermore, Karma exhibits strong generalization capabilities, showing leading performance under unseen networks in both simulations and real-world tests.