The currently most prominent algorithm to train keyword spotting (KWS) models with deep neural networks (DNNs) requires strong supervision i.e., precise knowledge of the spoken keyword location in time. Thus, most KWS approaches treat the presence of redundant data, such as noise, within their training set as an obstacle. A common training paradigm to deal with data redundancies is to use temporally weakly supervised learning, which only requires providing labels on a coarse scale. This study explores the limits of DNN training using temporally weak labeling with applications in KWS. We train a simple end-to-end classifier on the common Google Speech Commands dataset with increased difficulty by randomly appending and adding noise to the training dataset. Our results indicate that temporally weak labeling can achieve comparable results to strongly supervised baselines while having a less stringent labeling requirement. In the presence of noise, weakly supervised models are capable to localize and extract target keywords without explicit supervision, leading to a performance increase compared to strongly supervised approaches.
Transformers have emerged as a prominent model framework for audio tagging (AT), boasting state-of-the-art (SOTA) performance on the widely-used Audioset dataset. However, their impressive performance often comes at the cost of high memory usage, slow inference speed, and considerable model delay, rendering them impractical for real-world AT applications. In this study, we introduce streaming audio transformers (SAT) that combine the vision transformer (ViT) architecture with Transformer-Xl-like chunk processing, enabling efficient processing of long-range audio signals. Our proposed SAT is benchmarked against other transformer-based SOTA methods, achieving significant improvements in terms of mean average precision (mAP) at a delay of 2s and 1s, while also exhibiting significantly lower memory usage and computational overhead. Checkpoints are publicly available https://github.com/RicherMans/SAT.
Spatio-Temporal prediction plays a critical role in smart city construction. Jointly modeling multiple spatio-temporal tasks can further promote an intelligent city life by integrating their inseparable relationship. However, existing studies fail to address this joint learning problem well, which generally solve tasks individually or a fixed task combination. The challenges lie in the tangled relation between different properties, the demand for supporting flexible combinations of tasks and the complex spatio-temporal dependency. To cope with the problems above, we propose an Automated Spatio-Temporal multi-task Learning (AutoSTL) method to handle multiple spatio-temporal tasks jointly. Firstly, we propose a scalable architecture consisting of advanced spatio-temporal operations to exploit the complicated dependency. Shared modules and feature fusion mechanism are incorporated to further capture the intrinsic relationship between tasks. Furthermore, our model automatically allocates the operations and fusion weight. Extensive experiments on benchmark datasets verified that our model achieves state-of-the-art performance. As we can know, AutoSTL is the first automated spatio-temporal multi-task learning method.
With the development of sophisticated sensors and large database technologies, more and more spatio-temporal data in urban systems are recorded and stored. Predictive learning for the evolution patterns of these spatio-temporal data is a basic but important loop in urban computing, which can better support urban intelligent management decisions, especially in the fields of transportation, environment, security, public health, etc. Since traditional statistical learning and deep learning methods can hardly capture the complex correlations in the urban spatio-temporal data, the framework of spatio-temporal graph neural network (STGNN) has been proposed in recent years. STGNNs enable the extraction of complex spatio-temporal dependencies by integrating graph neural networks (GNNs) and various temporal learning methods. However, for different predictive learning tasks, it is a challenging problem to effectively design the spatial dependencies learning modules, temporal dependencies learning modules and spatio-temporal dependencies fusion methods in STGNN framework. In this paper, we provide a comprehensive survey on recent progress on STGNN technologies for predictive learning in urban computing. We first briefly introduce the construction methods of spatio-temporal graph data and popular deep learning models that are employed in STGNNs. Then we sort out the main application domains and specific predictive learning tasks from the existing literature. Next we analyze the design approaches of STGNN framework and the combination with some advanced technologies in recent years. Finally, we conclude the limitations of the existing research and propose some potential directions.
Training a 3D scene understanding model requires complicated human annotations, which are laborious to collect and result in a model only encoding close-set object semantics. In contrast, vision-language pre-training models (e.g., CLIP) have shown remarkable open-world reasoning properties. To this end, we propose directly transferring CLIP's feature space to 3D scene understanding model without any form of supervision. We first modify CLIP's input and forwarding process so that it can be adapted to extract dense pixel features for 3D scene contents. We then project multi-view image features to the point cloud and train a 3D scene understanding model with feature distillation. Without any annotations or additional training, our model achieves promising annotation-free semantic segmentation results on open-vocabulary semantics and long-tailed concepts. Besides, serving as a cross-modal pre-training framework, our method can be used to improve data efficiency during fine-tuning. Our model outperforms previous SOTA methods in various zero-shot and data-efficient learning benchmarks. Most importantly, our model successfully inherits CLIP's rich-structured knowledge, allowing 3D scene understanding models to recognize not only object concepts but also open-world semantics.
Keyword spotting (KWS) is a core human-machine-interaction front-end task for most modern intelligent assistants. Recently, a unified (UniKW-AT) framework has been proposed that adds additional capabilities in the form of audio tagging (AT) to a KWS model. However, previous work did not consider the real-world deployment of a UniKW-AT model, where factors such as model size and inference speed are more important than performance alone. This work introduces three mobile-device deployable models named Unified Transformers (UiT). Our best model achieves an mAP of 34.09 on Audioset, and an accuracy of 97.76 on the public Google Speech Commands V1 dataset. Further, we benchmark our proposed approaches on four mobile platforms, revealing that the proposed UiT models can achieve a speedup of 2 - 6 times against a competitive MobileNetV2.
The success of deep learning heavily relies on large-scale data with comprehensive labels, which is more expensive and time-consuming to fetch in 3D compared to 2D images or natural languages. This promotes the potential of utilizing models pretrained with data more than 3D as teachers for cross-modal knowledge transferring. In this paper, we revisit masked modeling in a unified fashion of knowledge distillation, and we show that foundational Transformers pretrained with 2D images or natural languages can help self-supervised 3D representation learning through training Autoencoders as Cross-Modal Teachers (ACT). The pretrained Transformers are transferred as cross-modal 3D teachers using discrete variational autoencoding self-supervision, during which the Transformers are frozen with prompt tuning for better knowledge inheritance. The latent features encoded by the 3D teachers are used as the target of masked point modeling, wherein the dark knowledge is distilled to the 3D Transformer students as foundational geometry understanding. Our ACT pretrained 3D learner achieves state-of-the-art generalization capacity across various downstream benchmarks, e.g., 88.21% overall accuracy on ScanObjectNN. Codes will be released at https://github.com/RunpeiDong/ACT.
Learning descriptive 3D features is crucial for understanding 3D scenes with diverse objects and complex structures. However, it is usually unknown whether important geometric attributes and scene context obtain enough emphasis in an end-to-end trained 3D scene understanding network. To guide 3D feature learning toward important geometric attributes and scene context, we explore the help of textual scene descriptions. Given some free-form descriptions paired with 3D scenes, we extract the knowledge regarding the object relationships and object attributes. We then inject the knowledge to 3D feature learning through three classification-based auxiliary tasks. This language-assisted training can be combined with modern object detection and instance segmentation methods to promote 3D semantic scene understanding, especially in a label-deficient regime. Moreover, the 3D feature learned with language assistance is better aligned with the language features, which can benefit various 3D-language multimodal tasks. Experiments on several benchmarks of 3D-only and 3D-language tasks demonstrate the effectiveness of our language-assisted 3D feature learning. Code is available at https://github.com/Asterisci/Language-Assisted-3D.
Robust prediction of citywide traffic flows at different time periods plays a crucial role in intelligent transportation systems. While previous work has made great efforts to model spatio-temporal correlations, existing methods still suffer from two key limitations: i) Most models collectively predict all regions' flows without accounting for spatial heterogeneity, i.e., different regions may have skewed traffic flow distributions. ii) These models fail to capture the temporal heterogeneity induced by time-varying traffic patterns, as they typically model temporal correlations with a shared parameterized space for all time periods. To tackle these challenges, we propose a novel Spatio-Temporal Self-Supervised Learning (ST-SSL) traffic prediction framework which enhances the traffic pattern representations to be reflective of both spatial and temporal heterogeneity, with auxiliary self-supervised learning paradigms. Specifically, our ST-SSL is built over an integrated module with temporal and spatial convolutions for encoding the information across space and time. To achieve the adaptive spatio-temporal self-supervised learning, our ST-SSL first performs the adaptive augmentation over the traffic flow graph data at both attribute- and structure-levels. On top of the augmented traffic graph, two SSL auxiliary tasks are constructed to supplement the main traffic prediction task with spatial and temporal heterogeneity-aware augmentation. Experiments on four benchmark datasets demonstrate that ST-SSL consistently outperforms various state-of-the-art baselines. Since spatio-temporal heterogeneity widely exists in practical datasets, the proposed framework may also cast light on other spatial-temporal applications. Model implementation is available at https://github.com/Echo-Ji/ST-SSL.