Membership testing, which determines whether an element belongs to a set, is widely used in fields like database systems and network applications. Bloom Filters (BFs) can solve this problem efficiently but suffer from high False Positive Rates (FPRs) and large memory requirements for massive datasets. Learned Bloom Filters (LBFs), combining a learning model with a backup Bloom Filter, mitigate these issues by capturing data distributions. However, the critical problem of memory allocation between the learning model and the backup filter has usually been overlooked, despite its significant impact on LBF performance under constrained budgets.

Irregular multivariate time series (IMTS) is characterized by the lack of synchronized observations across its different channels.In this paper, we point out that this channel-wise asynchrony can lead to poor channel-wise modeling of existing deep learning methods.To overcome this limitation, we propose MTM, a multi-scale token mixing transformer for the classification of IMTS.We find that the channel-wise asynchrony can be alleviated by down-sampling the time series to coarser timescales, and propose to incorporate a masked concat pooling in MTM that gradually down-samples IMTS to enhance the channel-wise attention modules.Meanwhile, we propose a novel channel-wise token mixing mechanism which proactively chooses important tokens from one channel and mixes them with other channels, to further boost the channel-wise learning of our model.Through extensive experiments on real-world datasets and comparison with state-of-the-art methods, we demonstrate that MTM consistently achieves the best performance on all the benchmarks, with improvements of up to 3.8% in AUPRC for classification.

Abstract With the popularity of mobile devices and the development of location technology, there is an increasing amount of text data with spatial and temporal tags generated. Querying with spatial, temporal, and keyword constraints on such data, known as spatio-temporal keyword query (STK query), is of great significance. However, most existing STK query solutions rely on tree-based indexes designed for stand-alone architectures, which struggle to scale for big data. Key-value stores, with the keys as their indexes, are designed for big data scenarios. On one hand, key-value stores can only support one-dimensional indexes initially, which makes them unsuitable for multi-dimensional STK queries. On the other hand, key-value stores put their indexes out of the memory, making it inevitable to trigger many unnecessary disk I/Os and slow down the query efficiency. To this end, based on key-value stores, we provide the first attempt by combining the in-memory index with on-disk index to efficiently support STK queries. Specifically, we design two-layer filters as the in-memory index, which enormously prunes unqualified spatio-temporal keyword combinations. An eviction policy is employed for the in-memory index, allowing it to support an infinite amount of data with limited memory usage. We deploy our solution on both HBase and Redis, conducting extensive experiments with two real and one synthetic datasets. The experimental results demonstrate that our solution achieves approximately twice the query efficiency of the state-of-the-art key-value based solutions, and is much more scalable than the tree-based competitor.

Time series data, including univariate and multivariate ones, are characterized by unique composition and complex multi-scale temporal variations. They often require special consideration of decomposition and multi-scale modeling to analyze. Existing deep learning methods on this best fit to univariate time series only, and have not sufficiently considered sub-series modeling and decomposition completeness. To address these challenges, we propose MSD-Mixer, a M ulti- S cale D ecomposition MLP- Mixer , which learns to explicitly decompose and represent the input time series in its different layers. To handle the multi-scale temporal patterns and multivariate dependencies, we propose a novel temporal patching approach to model the time series as multi-scale patches, and employ MLPs to capture intra- and inter-patch variations and channel-wise correlations. In addition, we propose a novel loss function to constrain both the mean and the autocorrelation of the decomposition residual for better decomposition completeness. Through extensive experiments on various real-world datasets for five common time series analysis tasks, we demonstrate that MSD-Mixer consistently and significantly outperforms other state-of-the-art algorithms with better efficiency.

We study the forecasting problem for traffic with dynamic, possibly periodical, and joint spatial-temporal dependency between regions. Given the aggregated inflow and outflow traffic of regions in a city from time slots 0 to$t - 1$, we predict the traffic at time$t$for any region. Prior arts in the area often considered the spatial and temporal dependencies in a decoupled manner, or were rather computationally intensive in training with a large number of hyper-parameters which needed tuning. We propose ST-TIS, a novel, lightweight and accurateSpatial-TemporalTransformer withinformation fusion and regionsampling for traffic forecasting. ST-TIS extends the canonical Transformer with information fusion and region sampling. The information fusion module captures the complex spatial-temporal dependency between regions. The region sampling module is to improve the efficiency and prediction accuracy, cutting the computation complexity for dependency learning from$O(n^{2})$to$O(n\sqrt{n})$, where$n$is the number of regions. With far fewer parameters than state-of-the-art deep learning models, ST-TIS's offline training is significantly faster in terms of tuning and computation (with a reduction of up to$90\%$on training time and network parameters). Notwithstanding such training efficiency, extensive experiments show that ST-TIS is substantially more accurate in online prediction than state-of-the-art approaches (with an average improvement of$9.5\%$on RMSE, and$12.4\%$on MAPE compared to STDN and DSAN).

Docked bike systems have been widely deployed in many cities around the world. To the service provider, predicting the demand and supply of bikes at any station is crucial to offering the best service quality. The docked bike prediction problem is highly challenging because of the complicated joint spatial-temporal (ST) dependency as bikes are picked up and dropped off, the so-called “flows”, between stations. Prior works often considered the spatial and temporal dependencies separately using sequential network models, and based on locality assumptions. Without sufficiently capturing the joint spatial and temporal features, these approaches are not optimal for attaining the best prediction accuracy. We propose STGNN-DJD, a novel data-driven Spatial-Temporal Graph Neural Network to solve the bike demand and supply prediction problem by unifiedly embedding the Dynamic and Joint ST Dependency in two novel ST graphs. Given station locations and historical rental data on bike flow over the past time slots 0 to$t-1$, we seek to predict online the bike demand and supply at any station at time$t$. To extract joint spatial-temporal dependency, STGNN-DJD employs a graph generator to construct, at the beginning of time$t$, two graphs that embed the flow relationships between stations at various time slots (flow-convoluted graph) and dynamic demand-supply pattern correlation between stations (pattern correlation graph), respectively. Given the two spatial-temporal graphs, STGNN-DJD subsequently employs a graph neural network with novel flow-based and attention-based aggregators to generate embedding of each station for docked bike prediction. We have conducted extensive experiments on two large bike-sharing datasets. Our re-sults confirm the effectiveness of STGNN-DJD as compared with other state-of-the-art approaches, with significant improvement on RMSE and MAE (by 20%-50%). We also provide a case study on dynamic dependencies between stations and demonstrate that the locality assumption does not always hold for a docked bike system.

With the rapid advances and the penetration of the Internet of Things and sensors, a massive amount of trajectory data, given by discrete locations at certain timestamps, have been extracted or collected. Knowing the similarity between trajectories is fundamental to understanding their spatial-temporal correlation, with direct and far-reaching applications in contact tracing, companion detection, personalized marketing, etc. In this work, we consider the general and realistic sensing scenario that the locations of the trajectories may be noisy, and that these trajectories are sporadically sampled with randomness and asynchrony from the underlying continuous paths. Most of the prior work on trajectory similarity has not sufficiently considered the temporal dimension, or the issues of location noise and sporadic sampling, while others have limitations of strong assumptions such as a fixed known speed of users or the availability of a large amount of training data.We propose a novel and effective spatial-temporal measure termed STS (Spatial-Temporal Similarity) to evaluate the spatial-temporal overlap between any two trajectories. In order to account for the location noise and sporadic sampling, STS models each location in a trajectory as an observable outcome drawn from a probability distribution. With that, it efficiently reduces the need for training data by estimating a personalized spatial-temporal probability distribution of the object position from its own trajectory. Based on that, it subsequently computes the co-location probability and hence derives the similarity of any two trajectories. We have conducted extensive experiments to evaluate STS using real large-scale indoor (mall) and outdoor (taxi) datasets. Our results show that STS is substantially more accurate and robust than the state-of-the-art approaches, with an improvement of 63% on precision and 85% on mean rank.

Public transportation is beating heart of a city. Understanding how citizens utilize public transportation can be used to optimize many applications such as traffic planning, crowd flow prediction, and location-based marketing. However, obtaining how citizens used transportation is not a trivial task. It is almost not possible to ask citizens to report their exact location and their transportation mode; moreover, there are usually various public transportation that move along the similar paths. These increase challenges to identify people's transport modes. To address these issues, this paper proposes Public Transport General Framework (PTGF) to identify people's transport modes by their cellular data in both offline and online manners. Regarding the offline phase, given historical cellular data of people and urban transportation networks, PTGF derives cellular data into trajectories, to match each trajectory to public transportation networks to find the most possible transport modes for sub-trajectories of a trajectory. In the online phase, given streaming trajectories, PTGF identifies the transport modes of each location by an LSTM which are trained by historical trajectories with transport modes annotated in the offline phase. Extensive experiments are conducted by using both synthetic and real datasets. The experimental results show that the accuracy of PTGF in offline phase around 80% and that in online phase F1-score around 0.7, which could prove that the effectiveness of the proposed framework PTGF.

Public transportation is essential in people's daily life and it is crucial to understand how people move around the city. Some prior works have exploited GPS, Wi-Fi or bluetooth to collect data, in which extra sensors or devices were needed. Other works utilized data from smart card systems. However, some public transportation systems have their own smart card system and the smart card data cannot include all kinds of transportation modes, which makes it unsuitable for our study.Nowadays, each user has his/her own mobile phones and from the cellular data of mobile phone service providers, it is possible to know the uses' transportation mode and the fine-grained crowd flows. As such, given a set of cellular data, we propose a system for public transportation mode detection, crowd density estimation, and crowd flow estimation. Note that we only have cellular data, no extra sensor data collected from users' mobile phones. In this paper, we refer to some external data sources (e.g., the bus routing networks) to identify transportation modes. Users' cellular data sometimes have uncertainty about user location information. Thus, we propose two approaches for different transportation mode detection considering the cell tower properties, spatial and temporal factors. We demonstrate our system using the data from Chunghwa Telecom, which is the largest telecommunication company in Taiwan, to show the usefulness of our system.