Combining CNNs or ViTs, with RNNs for spatiotemporal forecasting, has yielded unparalleled results in predicting temporal and spatial dynamics. However, modeling extensive global information remains a formidable challenge; CNNs are limited by their narrow receptive fields, and ViTs struggle with the intensive computational demands of their attention mechanisms. The emergence of recent Mamba-based architectures has been met with enthusiasm for their exceptional long-sequence modeling capabilities, surpassing established vision models in efficiency and accuracy, which motivates us to develop an innovative architecture tailored for spatiotemporal forecasting. In this paper, we propose the VMRNN cell, a new recurrent unit that integrates the strengths of Vision Mamba blocks with LSTM. We construct a network centered on VMRNN cells to tackle spatiotemporal prediction tasks effectively. Our extensive evaluations show that our proposed approach secures competitive results on a variety of tasks while maintaining a smaller model size. Our code is available at https://github.com/yyyujintang/VMRNN-PyTorch.
We study the problem of unsupervised domain adaptation for egocentric videos. We propose a transformer-based model to learn class-discriminative and domain-invariant feature representations. It consists of two novel designs. The first module is called Generative Adversarial Domain Alignment Network with the aim of learning domain-invariant representations. It simultaneously learns a mask generator and a domain-invariant encoder in an adversarial way. The domain-invariant encoder is trained to minimize the distance between the source and target domain. The masking generator, conversely, aims at producing challenging masks by maximizing the domain distance. The second is a Masked Consistency Learning module to learn class-discriminative representations. It enforces the prediction consistency between the masked target videos and their full forms. To better evaluate the effectiveness of domain adaptation methods, we construct a more challenging benchmark for egocentric videos, U-Ego4D. Our method achieves state-of-the-art performance on the Epic-Kitchen and the proposed U-Ego4D benchmark.
We study zero-shot instance navigation, in which the agent navigates to a specific object without using object annotations for training. Previous object navigation approaches apply the image-goal navigation (ImageNav) task (go to the location of an image) for pretraining, and transfer the agent to achieve object goals using a vision-language model. However, these approaches lead to issues of semantic neglect, where the model fails to learn meaningful semantic alignments. In this paper, we propose a Prioritized Semantic Learning (PSL) method to improve the semantic understanding ability of navigation agents. Specifically, a semantic-enhanced PSL agent is proposed and a prioritized semantic training strategy is introduced to select goal images that exhibit clear semantic supervision and relax the reward function from strict exact view matching. At inference time, a semantic expansion inference scheme is designed to preserve the same granularity level of the goal-semantic as training. Furthermore, for the popular HM3D environment, we present an Instance Navigation (InstanceNav) task that requires going to a specific object instance with detailed descriptions, as opposed to the Object Navigation (ObjectNav) task where the goal is defined merely by the object category. Our PSL agent outperforms the previous state-of-the-art by 66% on zero-shot ObjectNav in terms of success rate and is also superior on the new InstanceNav task. Code will be released at https://anonymous.4open. science/r/PSL/.
Zero-shot action recognition (ZSAR) aims to learn an alignment model between videos and class descriptions of seen actions that is transferable to unseen actions. The text queries (class descriptions) used in existing ZSAR works, however, are often short action names that fail to capture the rich semantics in the videos, leading to misalignment. With the intuition that video content descriptions (e.g., video captions) can provide rich contextual information of visual concepts in videos, we propose to utilize human annotated video descriptions to enrich the semantics of the class descriptions of each action. However, all existing action video description datasets are limited in terms of the number of actions, the semantics of video descriptions, etc. To this end, we collect a large-scale action video descriptions dataset named ActionHub, which covers a total of 1,211 common actions and provides 3.6 million action video descriptions. With the proposed ActionHub dataset, we further propose a novel Cross-modality and Cross-action Modeling (CoCo) framework for ZSAR, which consists of a Dual Cross-modality Alignment module and a Cross-action Invariance Mining module. Specifically, the Dual Cross-modality Alignment module utilizes both action labels and video descriptions from ActionHub to obtain rich class semantic features for feature alignment. The Cross-action Invariance Mining module exploits a cycle-reconstruction process between the class semantic feature spaces of seen actions and unseen actions, aiming to guide the model to learn cross-action invariant representations. Extensive experimental results demonstrate that our CoCo framework significantly outperforms the state-of-the-art on three popular ZSAR benchmarks (i.e., Kinetics-ZSAR, UCF101 and HMDB51) under two different learning protocols in ZSAR. We will release our code, models, and the proposed ActionHub dataset.
Vision transformers have recently emerged as an effective alternative to convolutional networks for action recognition. However, vision transformers still struggle with geometric variations prevalent in video data. This paper proposes a novel approach, GeoDeformer, designed to capture the variations inherent in action video by integrating geometric comprehension directly into the ViT architecture. Specifically, at the core of GeoDeformer is the Geometric Deformation Predictor, a module designed to identify and quantify potential spatial and temporal geometric deformations within the given video. Spatial deformations adjust the geometry within individual frames, while temporal deformations capture the cross-frame geometric dynamics, reflecting motion and temporal progression. To demonstrate the effectiveness of our approach, we incorporate it into the established MViTv2 framework, replacing the standard self-attention blocks with GeoDeformer blocks. Our experiments at UCF101, HMDB51, and Mini-K200 achieve significant increases in both Top-1 and Top-5 accuracy, establishing new state-of-the-art results with only a marginal increase in computational cost. Additionally, visualizations affirm that GeoDeformer effectively manifests explicit geometric deformations and minimizes geometric variations. Codes and checkpoints will be released.
Developing end-to-end models for long-video action understanding tasks presents significant computational and memory challenges. Existing works generally build models on long-video features extracted by off-the-shelf action recognition models, which are trained on short-video datasets in different domains, making the extracted features suffer domain discrepancy. To avoid this, action recognition models can be end-to-end trained on clips, which are trimmed from long videos and labeled using action interval annotations. Such fully supervised annotations are expensive to collect. Thus, a weakly supervised method is needed for long-video action understanding at scale. Under the weak supervision setting, action labels are provided for the whole video without precise start and end times of the action clip. To this end, we propose an AdaFocus framework. AdaFocus estimates the spike-actionness and temporal positions of actions, enabling it to adaptively focus on action clips that facilitate better training without the need for precise annotations. Experiments on three long-video datasets show its effectiveness. Remarkably, on two of datasets, models trained with AdaFocus under weak supervision outperform those trained under full supervision. Furthermore, we form a weakly supervised feature extraction pipeline with our AdaFocus, which enables significant improvements on three long-video action understanding tasks.
Accurate precipitation forecasting is a vital challenge of both scientific and societal importance. Data-driven approaches have emerged as a widely used solution for addressing this challenge. However, solely relying on data-driven approaches has limitations in modeling the underlying physics, making accurate predictions difficult. Coupling AI-based post-processing techniques with traditional Numerical Weather Prediction (NWP) methods offers a more effective solution for improving forecasting accuracy. Despite previous post-processing efforts, accurately predicting heavy rainfall remains challenging due to the imbalanced precipitation data across locations and complex relationships between multiple meteorological variables. To address these limitations, we introduce the PostRainBench, a comprehensive multi-variable NWP post-processing benchmark consisting of three datasets for NWP post-processing-based precipitation forecasting. We propose CAMT, a simple yet effective Channel Attention Enhanced Multi-task Learning framework with a specially designed weighted loss function. Its flexible design allows for easy plug-and-play integration with various backbones. Extensive experimental results on the proposed benchmark show that our method outperforms state-of-the-art methods by 6.3%, 4.7%, and 26.8% in rain CSI on the three datasets respectively. Most notably, our model is the first deep learning-based method to outperform traditional Numerical Weather Prediction (NWP) approaches in extreme precipitation conditions. It shows improvements of 15.6%, 17.4%, and 31.8% over NWP predictions in heavy rain CSI on respective datasets. These results highlight the potential impact of our model in reducing the severe consequences of extreme weather events.
Although the Transformer has been the dominant architecture for time series forecasting tasks in recent years, a fundamental challenge remains: the permutation-invariant self-attention mechanism within Transformers leads to a loss of temporal information. To tackle these challenges, we propose PatchMixer, a novel CNN-based model. It introduces a permutation-variant convolutional structure to preserve temporal information. Diverging from conventional CNNs in this field, which often employ multiple scales or numerous branches, our method relies exclusively on depthwise separable convolutions. This allows us to extract both local features and global correlations using a single-scale architecture. Furthermore, we employ dual forecasting heads that encompass both linear and nonlinear components to better model future curve trends and details. Our experimental results on seven time-series forecasting benchmarks indicate that compared with the state-of-the-art method and the best-performing CNN, PatchMixer yields $3.9\%$ and $21.2\%$ relative improvements, respectively, while being 2-3x faster than the most advanced method. We will release our code and model.
LiDAR semantic segmentation plays a crucial role in enabling autonomous driving and robots to understand their surroundings accurately and robustly. There are different types of methods, such as point-based, range-image-based, polar-based, and hybrid methods. Among these, range-image-based methods are widely used due to their efficiency. However, they face a significant challenge known as the ``many-to-one'' problem caused by the range image's limited horizontal and vertical angular resolution. As a result, around 20\% of the 3D points can be occluded. In this paper, we present TFNet, a range-image-based LiDAR semantic segmentation method that utilizes temporal information to address this issue. Specifically, we incorporate a temporal fusion layer to extract useful information from previous scans and integrate it with the current scan. We then design a max-voting-based post-processing technique to correct false predictions, particularly those caused by the ``many-to-one'' issue. We evaluated the approach on two benchmarks and demonstrate that the post-processing technique is generic and can be applied to various networks. We will release our code and models.