Generating realistic and controllable vehicle trajectories remains a key challenge in traffic simulation due to the dynamic and stochastic nature of road traffic. Existing deep learning approaches often rely solely on raw sensory inputs, limiting their ability to capture both fine-grained spatiotemporal dependencies and highlevel semantic context. To address this gap, we propose a generalizable framework with two complementary modules. The first performs low-level numerical modeling using a dynamic Spatial-Temporal Graph Neural Network and visual encoder to capture spatial-temporal dynamics and environmental context. The second applies perception-driven semantic modeling via a Multi-Modal Large Language Model pipeline to extract human-like interpretations of traffic scenarios. Outputs from both modules condition a diffusion-based generative model to produce behaviorally feasible, controllable trajectories. Experiments on a real-world dataset show significant improvements over state-of-the-art baselines across multiple metrics, validating the effectiveness of our approach.

In the current social media landscape, the study of influence propagation and consensus formation has gained prominence. While user interactions like retweeting are apparent, the underlying pathways of influence often remain hidden and complex. This study proposes a novel network called Latent Influence Network (LIN), which advances the analysis of influence on social media. LIN's architecture and the process of parameter selection are meticulously discussed within the comprehensive Latent Influence Detection Framework (LIDET). Based on the user's behavior label, LIN identifies the optimal network configuration, revealing more accurate influence patterns. We applied the LIDET framework to four diverse datasets, each demonstrating substantial improvements in influence pattern recognition over traditional network models. Specifically, in a case study on a COVID-19 dataset, LIN achieved a classification accuracy of 99%, significantly outperforming conventional methods. These findings underscore the utility of LIN in capturing the dynamics of influence and enhancing our understanding of opinion formation on social media.

The rise of social media has been accompanied by a dark side with the ease of creating fake accounts and disseminating misinformation through coordinated attacks. Existing methods to identify such attacks often rely on thematic similarities or network-based approaches, overlooking the intricate causal relationships that underlie coordinated actions. This work introduces a novel approach for detecting coordinated attacks using Convergent Cross Mapping (CCM), a technique that infers causality from temporal relationships between user activity. We build on the theoretical framework of CCM by incorporating topic modelling as a basis for further optimizing its performance. We apply CCM to real-world data from the infamous IRA attack on US elections, achieving F1 scores up to 75.3% in identifying coordinated accounts. Furthermore, we analyse the output of our model to identify the most influential users in a community and uncover leader-follower dynamics based on inferred causal relationships. We also demonstrate how our method reveals coordinated behaviour across different time periods, including campaigns predating the 2016 elections. We apply our model to a case study involving COVID-19 anti-vax related discussions on Twitter. Our results demonstrate the effectiveness of our model in uncovering causal structures of coordinated behaviour, offering a promising avenue for mitigating the threat of malicious campaigns on social media platforms.

The increasing availability of urban data offers new opportunities for learning region representations, which can be used as input to machine learning models for downstream tasks such as check-in or crime prediction. While existing solutions have produced promising results, an issue is their fixed formation of regions and fixed input region features, which may not suit the needs of different downstream tasks. To address this limitation, we propose a model named FlexiReg for urban region representation learning that is flexible with both the formation of urban regions and the input region features. FlexiReg is based on a spatial grid partitioning over the spatial area of interest. It learns representations for the grid cells, leveraging publicly accessible data, including POI, land use, satellite imagery, and street view imagery. We propose adaptive aggregation to fuse the cell representations and prompt learning techniques to tailor the representations towards different tasks, addressing the needs of varying formations of urban regions and downstream tasks. Extensive experiments on five real-world datasets demonstrate that FlexiReg outperforms state-of-the-art models by up to 202% in term of the accuracy of four diverse downstream tasks using the produced urban region representations.

An increasing number of related urban data sources have brought forth novel opportunities for learning urban region representations, i.e., embeddings. The embeddings describe latent features of urban regions and enable discovering similar regions for urban planning applications. Existing methods learn an embedding for a region using every different type of region feature data, and subsequently fuse all learned embeddings of a region to generate a unified region embedding. However, these studies often overlook the significance of the fusion process. The typical fusion methods rely on simple aggregation, such as summation and concatenation, thereby disregarding correlations within the fused region embeddings. To address this limitation, we propose a novel model named HAFusion. Our model is powered by a dual-feature attentive fusion module named DAFusion, which fuses embeddings from different region features to learn higher-order correlations be-tween the regions as well as between the different types of region features. DAFusion is generic - it can be integrated into existing models to enhance their fusion process. Further, motivated by the effective fusion capability of an attentive module, we propose a hybrid attentive feature learning module named HALearning to enhance the embedding learning from each individual type of region features. Extensive experiments on three real-world datasets demonstrate that our model HAFusion outperforms state-of-the-art models across three different prediction tasks. Using our learned region embeddings leads to consistent and up to 31 % improvements in the prediction accuracy.

The emergence of social media as one of the main platforms for people to access news has enabled the wide dissemination of fake news, having serious impacts on society. Thus, it is really important to identify fake news with high confidence in a timely manner, which is not feasible using manual analysis. This has motivated numerous studies on automating fake news detection. Most of these approaches are supervised, which requires extensive time and labour to build a labelled dataset. Although there have been limited attempts at unsupervised fake news detection, their performance suffers due to not exploiting the knowledge from various modalities related to news records and due to the presence of various latent biases in the existing news datasets (e.g., unrealistic real and fake news distributions). To address these limitations, this work proposes an effective framework for unsupervised fake news detection, which first embeds the knowledge available in four modalities (i.e., source credibility, textual content, propagation speed, and user credibility) in news records and then proposes$(UMD)^{2}$, a novel noise-robust self-supervised learning technique, to identify the veracity of news records from the multi-modal embeddings. Also, we propose a novel technique to construct news datasets minimizing the latent biases in existing news datasets. Following the proposed approach for dataset construction, we produce a Large-scale Unlabelled News Dataset consisting 419,351 news articles related to COVID-19, acronymed asLUND-COVID. We trained the proposed unsupervised framework usingLUND-COVIDto exploit the potential of large datasets, and evaluate it using a set of existing labelled datasets. Our results show that the proposed unsupervised framework largely outperforms existing unsupervised baselines for different tasks such as multi-modal fake news detection, fake news early detection and few-shot fake news detection, while yielding notable improvements for unseen domains during training.

Abstract With the increasing connectedness of vehicles, real-time spatio-temporal data can be collected from citywide road networks. Innovative data management solutions can process the collected data for the purpose of reducing travel time. However, a majority of the existing solutions have missed the opportunity to better manage the collected data for improving road safety at the network level. We propose an efficient data management framework that uses network-level data to improve road safety for citywide applications. Our framework uses a graph-based data structure to maintain real-time network-level traffic data. Based on the graph, the framework uses a novel technique to generate driving instructions for individual vehicles. By following the instructions, inter-vehicular spacing can be increased, leading to an improvement of road safety. Experimental results show that our framework improves road safety, measured based on the time to collision between vehicles, from the state-of-the-art traffic data management solutions by a large margin while achieving lower travel times compared with the solutions. The framework is also readily deployable for large-scale real-time applications due to its low computation costs.

Dynamic road network optimization has been used for improving traffic flow in an infrequent and localized manner. The development of intelligent systems and technology provides an opportunity to improve the frequency and scale of dynamic road network optimization. However, such improvements are hindered by the high computational complexity of the existing algorithms that generate the optimization plans. We present a novel solution that integrates machine learning and road network optimization. Our solution consists of two complementary parts. The first part is an efficient algorithm that uses reinforcement learning to find the best road network configurations at real-time. The second part is a dynamic routing mechanism, which helps connected vehicles adapt to the change of the road network. Our extensive experimental results demonstrate that the proposed solution can substantially reduce the average travel time in a variety of scenarios, whilst being computationally efficient and hence applicable to real-life situations.

Intelligent traffic management solutions that leverage machine learning have gained a lot of interest in recent years. These techniques, however, cannot be deployed in real-world settings at a desirable pace due to technological barriers. Thus, easily customizable, realistic simulation environments are needed to train and verify the effectiveness of machine learning algorithms for traffic control. We propose an easily extendable traffic simulation system named e-SMARTS to allow researchers to experiment with novel data-driven traffic management algorithms in a setup that mimics real-world traffic conditions. We demonstrate the flexibility of e-SMARTS using widely researched traffic management solutions for Autonomous Intersection Management (AIM). In the demonstration, we present several pluggable algorithms for AIM and show that these computationally efficient algorithms can achieve effective and safe results.

Graph combinatorial optimization (CO) is a widely studied problem with use-cases stemming from many fields. Typically, in real-world applications, the features of a graph tend to change over time (e.g. traffic congestion, or travel time), thus, finding a solution to the dynamic graph CO problem is critical. In recent years, using deep learning techniques to find heuristic solutions for NP-hard CO problems has gained much interest as these learned heuristics can find near-optimal solutions efficiently. However, most of the existing methods for learning heuristics focus on static CO problems. The dynamic nature makes NP-hard CO problems much more challenging to learn, and the existing methods fail to find reasonable solutions. We propose a novel architecture named Graph Temporal Attention with Reinforcement Learning (GTA-RL) to learn heuristic solutions for dynamic versions of graph CO problems. We then extend our architecture to learn heuristics for the real-time version of CO problems where all input features of a problem are not known a priori, but rather learned in real-time. A detailed experimental evaluation against several state-of-the-art learning-based algorithms and optimal solvers demonstrates the efficiency and effectiveness of our approach.

With increasing connectivity and autonomy in traffic eco-systems, Autonomous Intersection Management (AIM) has attracted strong attention from the research community. AIM helps optimize traffic by coordinating the trajectory of connected vehicles around intersections. Most of the existing AIM solutions are developed for single-objective optimization problems that are focused on improving traffic flow. A complete AIM solution needs to perform bi-objective optimization that considers both traffic flow and safety. However, the computational complexity for achieving both objectives is significantly high with the existing solutions, especially when traffic demand is stochastic. We address the limitations of the existing solutions using deep reinforcement learning (deep RL) that helps solve complex problems efficiently. Our solution uses two types of RL agents. The first type is intersection-level agents, which generate theoretically sound trajectory plans for individual vehicles approaching intersections. The second type is vehicle-level agents that control vehicles' actual trajectories around the intersections based on the plans. Both agents incorporate traffic flow and safety constraints into their decision making. Our experimental results show that our solution achieves a high safety level with a minimum impact on travel time.

Due to the importance of road freight, there is a significant cost of delaying freight vehicles on the road. In this work, we focus on freight vehicle optimization by reducing delays at intersections. Our simulation study evaluates the effectiveness of an autonomous intersection management strategy that prioritizes connected freight vehicles using intelligent traffic lights. We simulate a wide range of traffic scenarios on our microscopic traffic simulator. Our results show that the strategy can help reduce the delay of freight vehicles with a minimal impact on other vehicles in a real road network. Our simulations also reveal the scenarios where the strategy works best and where it should be avoided. Effects of individual parameters are also measured through simulations.

In agile iterative development, an agile team needs to analyze documented information for effort estimation and sprint planning. While documentation can be changed, the documentation changes after sprint planning may invalidate the estimated effort and sprint plan. Hence, to help the team be aware of the potential documentation changes, we developed DocWarn to estimate the probability that a work item will have documentation changes. We developed three variations of DocWarn, which are based on the characteristics extracted from the work items (DocWarn-C), the natural language text (DocWarn-T), and both inputs (DocWarn-H).

Traffic forecasting plays a vital role in traffic management systems. Recently, deep learning models have been applied to citywide traffic forecasting. However, the existing work models and predicts traffic at a single (dense) resolution, making it challenging to capture long-range spatial dependencies or high-level traffic dynamics. This shortcoming limits the accuracy of prediction and results in computationally expensive models. We propose a traffic forecasting model based on deep convolutional networks to improve the accuracy of citywide traffic forecasting. Our model uses a hierarchical architecture that captures traffic dynamics at multiple spatial resolutions. Based on this architecture, we apply a multi-task learning scheme, which trains the model to predict traffic at different resolutions. Our model helps provide a coherent understanding of traffic dynamics by capturing spatial dependencies between different regions of a city. Experimental results on multiple real datasets show that our model can achieve competitive results compared to complex state-of-the-art approaches while being more computationally efficient.

Many learning tasks involve multi-modal data streams, where continuous data from different modes convey a comprehensive description about objects. A major challenge in this context is how to efficiently interpret multi-modal information in complex environments. This has motivated numerous studies on learning unsupervised representations from multi-modal data streams. These studies aim to understand higher-level contextual information (e.g., a Twitter message) by jointly learning embeddings for the lower-level semantic units in different modalities (e.g., text, user, and location of a Twitter message). However, these methods directly associate each low-level semantic unit with a continuous embedding vector, which results in high memory requirements. Hence, deploying and continuously learning such models in low-memory devices (e.g., mobile devices) becomes a problem. To address this problem, we present METEOR, a novel MEmory and Time Efficient Online Representation learning technique, which: (1) learns compact representations for multi-modal data by sharing parameters within semantically meaningful groups and preserves the domain-agnostic semantics; (2) can be accelerated using parallel processes to accommodate different stream rates while capturing the temporal changes of the units; and (3) can be easily extended to capture implicit/explicit external knowledge related to multi-modal data streams. We evaluate METEOR using two types of multi-modal data streams (i.e., social media streams and shopping transaction streams) to demonstrate its ability to adapt to different domains. Our results show that METEOR preserves the quality of the representations while reducing memory usage by around 80% compared to the conventional memory-intensive embeddings.

We study the problem of tracking the movement of groups using sparse trajectory data extracted from Location Based Social Networks (LBSNs). Tracking group movement using LBSN data is challenging because the data may contain a large amount of noise due to the lack of stability in group entity, spatial extent and posting time. We propose a first-of-its-kind solution, Group Kalman Filter (GKF), which aims to improve the effectiveness of group tracking by predicting the spatial properties of groups with a group movement model. Our experiments with real LBSN data and synthetic LBSN data show that GKF can detect groups and predict group movement with a high level of accuracy and efficiency.

Each year, millions of people either die or get injured due to road incidents. Thus, integrating safety optimization techniques into future traffic systems is of utmost importance. Emerging connected vehicle technologies have enabled ways to manage traffic networks with optimization goals such as travel time efficiency, fuel efficiency. However, these existing studies have focused less on maximizing traffic safety. Increasing space between vehicles in the road network with an acceptable travel time increase will help to improve the safety of the system. We propose the Platooning Graph, which is capable of modelling the inter-vehicular spacing optimization problem and we provide a fast and readily deployable algorithm to find a good approximate solution. Using microscopic traffic simulations, we demonstrate how the proposed method can improve safety, with minimal impact on travel time.

We present a new unsupervised dimensionality reduction technique, called LN-SNE, for anomaly detection. LN-SNE generates a parametric embedding by means of Restricted Boltzmann Machines and uses a heavy-tail distribution to project data to a lower dimensional space such that dissimilarities between normal data and anomalies are preserved or strengthened. We compare LN-SNE to several benchmark dimensionality reduction methods on real datasets. The results suggest that LN-SNE for anomaly detection is less sensitive to the dimension of the latent space than the other methods and outperforms them in terms of accuracy. We empirically show that our technique scales near-linearly with respect to the number of dimensions and data size.

Compared with traditional news media, social media nowadays provides a richer and more timely source of news. We are interested in multi-spatial level event detection from geo-tagged tweet streams. Specifically, in this paper we (1) examine the statistical characteristic for the time series of the number of geo-tagged tweets posted from specific regions during a short time interval, e.g., one minute; (2) verify from over thirty datasets that while almost all such time series exhibit self-similarity, those that correspond to events, especially short-term and unplanned outbursts, follow a power-law distribution; (3) demonstrate that these findings can be applied to facilitate event detection from tweet streams. We propose two algorithms-Power-law basic and Power-law advanced, where Power-law basic only checks the existence of power-law distributions in the time series from tweet streams at multi-spatial scales, without looking into the content of each tweet, and Power-law advanced integrates power-law verification with semantic analysis via word embedding. Our experiments on multiple datasets show that when combined with a Quad-tree, the seemingly naive algorithm of Power-law basic achieves comparable results with more advanced event detection methods, while the semantic analysis enhanced version, Power-law advanced, can significantly increase both the precision and the recall.

Identifying events happening in a specific locality is important as an early warning for accidents, protests, elections or breaking news. However, this location-specific event detection is challenging as the locations and types of events are not known beforehand. To address this problem, we propose an online spatio-temporal event detection system using social media that is able to detect events at different time and space resolutions. First, we exploit a quad-tree method to split the geographical space into multiscale regions based on the density of social media data. Then, we implement a statistical unsupervised approach using Poisson distribution and a smoothing method for highlighting regions with unexpected density of social posts. Further, event duration is estimated by merging events happening in the same region at consecutive time intervals. A post processing stage is introduced to filter out events that are spam, fake or wrong. Finally, we incorporate simple semantics by using social media entities to assess the integrity, and accuracy of detected events. The proposed method is evaluated using Twitter and Flickr for the city of Melbourne based on recall and precision measures. We also propose a new quality measure named strength index, which automatically measures how accurate the reported event is.

Currently, the enormous span of social media usage - while providing valuable resources for linguistic behavior analysis - makes tracking and understanding these multilingual discussions a challenging task. We have undertaken a multidisciplinary comprehensive study of multilingual discussions via the development of specialized data collection techniques that discover and track multilingual users of social media, and their associated discussions, within a defined geographical region. To facilitate automatic discussion analysis of large numbers of discussions we generated a machine learning model based on ground truth data obtained from Amazon Turk. Our approach goes beyond analyzing social media posts in isolation, by analyzing them in the context of the discussion in which they appear. We show a selection of example discussions found using our approach which reveals a number of interesting socio-linguistic interactions in the communities that we sampled, in support of approach as a general methodology for multilingual community analysis.

Building spatiotemporal activity models for people's activities in urban spaces is important for understanding the ever-increasing complexity of urban dynamics. With the emergence of Geo-Tagged Social Media (GTSM) records, previous studies demonstrate the potential of GTSM records for spatiotemporal activity modeling. State-of-the-art methods for this task embed different modalities (location, time, and text) of GTSM records into a single embedding space. However, they ignore Non-GeoTagged Social Media (NGTSM) records, which generally account for the majority of posts (e.g., more than 95% in Twitter), and could represent a great source of information to alleviate the sparsity of GTSM records. Furthermore, in the current spatiotemporal embedding techniques, less focus has been given to the users, who exhibit spatially motivated behaviors. To bridge this research gap, this work proposes USTAR, a novel online learning method for User-guided SpatioTemporal Activity Representation, which (1) embeds locations, time, and text along with users into the same embedding space to capture their correlations; (2) uses a novel collaborative filtering approach to incorporate both NGTSM and GTSM records in learning; and (3) introduces a novel sampling technique to learn spatiotemporal representations in an online fashion to accommodate recent information into the embedding space, while avoiding overfitting to recent records and frequently appearing units in social media streams. Our results show that USTAR substantially improves the state-of-the-art for region retrieval and keyword retrieval and its potential to be applied to other downstream applications such as local event detection.

Real-time processing of large-scale, time-varying graphs is becoming essential in bigdata streaming applications that require high-throughput and low-latency. Streaming applications typically perform graph computations on periodic snapshots of the graph, using windowing. For large-scale graphs, that contains millions of edges, distributed parallel processing is essential to meet the throughput and latency needs of streaming applications. Apache Storm, a widely used stream processing system, provides a topology-based processing architecture which supports distribution and pipelining to meet the latency and throughput needs; a typical Storm topology contains a fixed number of stages, each stage with a single operator supporting a single computation step, which can be parallelized through horizontal scaling (bolt instances). One of the limitations of such a topology, is the fixed number of stages, which makes it not suitable for graph algorithms that require iterative computations, such as the PageRank algorithm, because the number of iterations, hence the number of stages in the topology, is unknown a-priori. In this paper we explore how iterative graph algorithms can be executed efficiently in a distributed manner for scalability (with the graph distributed across a set of Storm bolts), using a single stage of a Storm topology, using Storm's cyclic communication (S-CC) mechanism. We show that S-CC approach outperforms alternative approaches that can be used for parallel-iterative computations in Storm, such as off-loading to a third-party graph processing system such as Giraph or using a distributed memory architecture, both in terms of scalability and performance.

Stream processing systems are used for a plethora of low-latency applications that deal with high volumes and varieties of data. In general, streaming applications are formulated as a fixed number of inter-connected operators, where the operator graph indicates the sequence of computations that apply to the streaming data in motion. Typically an iterative operation in a streaming application is done by embedding the iteration in a single operator (which excludes multiple operator instances and hence not scalable) or by using a sequence of operators, one for each iteration (which requires knowledge of the number of iterations a priori). However, the number of iterations of some applications such as k-means clustering cannot be determined at the creation of the application as it depends on the properties of the data received at runtime and the convergence criterion. Hence only a limited number of iterative computations can be executed using this approach. In this paper, we propose a communication model to support intra-operator communication so that iterative computations, including those with an arbitrary number of iterations, can be efficiently executed in streaming applications. We show that the proposed model can support different iterative algorithms that have complex communication patterns. Finally, through evaluating a number of parallel iterative algorithms using large-scale datasets, we demonstrate the scalability and performance of our proposed communication model and compare it to the existing approaches used for constructing iterative streaming applications.

There are many real world applications that require to identify movement of users such as identifying movement corridors, most popular paths, and nearest neighbours. If one is not given trajectories mapping to movement of people but rather sporadic location data, such as location based social network data, finding movement related information becomes difficult. Rather than processing all points in a data set given a query, a clever approach is to construct a graph, based on user locations, and query this graph for all queries. One example is the shortest path graph. However the shortest path graph can be inefficient and ineffective analysing movement, as it calculates the graph considering all points in a data set. We propose the stepping stone graph, which calculates graph considering point pairs rather than all points, that focuses on local possible movement, making it both efficient and effective for location based social network related queries. We demonstrate its uses by applying it in the aforementioned domain and comparing with the shortest path graph. We also compare its properties to a range of other graphs.

Microscopic traffic simulators play a major role to carry research on transportation problems. Microscopic traffic simulation is powerful because it enables efficient analysis of complex traffic problems to the highest level of detail. We developed Scalable Microscopic Adaptive Road Traffic Simulator (SMARTS) [14] that can perform large-scale simulations at a high speed by utilizing distributed computing resources. Previous results show that SMARTS can run 1.14 times faster than real time when simulating one million vehicles for the city of Melbourne on 30 distributed processors, while producing highly accurate simulation results. SMARTS' pluggable architecture allows it to be easily extended to simulate specific scenarios of interest to users. In this demonstration we show how SMARTS can be used to simulate an intersection design, the P-turn, in a major intersection of Melbourne. Our simulation shows the impact of the design on the traffic flow, confirming the justification for introduction of the particular intersection. The demo can be used as a template for future use of the simulator for other traffic problems.

With the proliferation of sensors and the increased connectivity of citizens, many global cities are increasingly adopting Smart City initiatives. Such initiatives provide real-time monitoring capabilities, and effective modelling techniques allow the prediction of future states in a city. For example, urban electricity smart meter data can be utilised to predict future demand in order to facilitate capacity planning. However, the accuracy of this foresight is often marred by low quality and missing sensor data in real-world systems. In this work, we focus on the problem of reliable forecasting by mitigating the effect of missing data on forecast accuracy. In order to mitigate the effects of missing data, we develop a multi-task learning scheme to jointly learn Gaussian Process Regression models between highly correlated sensors. We demonstrate that our methods are robust in a variety of error generation scenarios. We validate our methods based on publicly available and real-world datasets related to electricity smart meters in a university campus and pedestrian counts in a global city, where we achieve significant improvement over competitive baselines and other effective forecasting methods.

Twitter is a popular microblogging service, where users frequently engage in discussions about various topics of interest, ranging from popular topics (e.g., music) to niche topics (e.g., politics). With the large amount of tweets, a key challenge is to automatically model and determine the discussion topics without having prior knowledge of the types and number of topics, or requiring the technical expertise to define various algorithmic parameters. For this purpose, we propose the Clustering-based Topic Modelling (ClusTop) algorithm that constructs various types of word network and automatically determines the discussion topics using community detection approaches. Unlike traditional topic models, ClusTop is able to automatically determine the appropriate number of topics and does not require numerous parameters to be set. The ClusTop algorithm is also able to capture the syntactic meaning in tweets via the use of bigrams, trigrams and other word combinations in constructing the word network graph. Using three Twitter datasets with labelled crises and events as topics, ClusTop has been shown to outperform various baselines in terms of topic coherence, pointwise mutual information, precision, recall and F-score.

Topic modelling is a well-studied field that aims to identify topics from traditional documents such as news articles and reports. More recently, Latent Dirichlet Allocation (LDA) and its variants, have been applied on social media platforms to model and study topics relating to sports, politics and companies. While these applications were able to successfully identify the general topics, we posit that standard LDA can be augmented with spatial and temporal considerations based on the geo-coordinates and timestamps of social media posts. Towards this effort, we propose a spatial and temporal variant of LDA to better detect more specific topics, such as a particular art exhibit held at a museum or a security incident happening on a particular day. We validate our approach on a Twitter dataset and find that the detected topics are well-aligned to real-life events happening on the specific days and locations.

Traffic congestion is a serious problem that is only expected to get worse in the future. Statistics shows that half of traffic congestion is caused by temporary disruptions like accidents. These events have dramatic impact on road network availability and cause huge delays for commuters. Also, they are usually unexpected and hard to manage by traffic authorities. State-of-the-art navigation systems started to provide real-time information about traffic conditions to help users make better routing decisions. However, traffic in the road network changes rapidly and the advice calculated now may not be valid after few minutes. This is especially critical in the presence of traffic incidents, where the impact of the incident could cause traffic to propagate to nearby roads. Thus, it is important for navigation systems to consider the evolution and future impact of traffic events. In this work, we present a navigation system that uses faster than realtime simulations to predict the evolution of traffic events and help drivers proactively avoid congestion caused by events. The system can subscribe to real-time traffic information and forecast the traffic conditions using fast simulations. We evaluate our approach through extensive experiments to test the performance and accuracy of the simulator with real data obtained from TomTom Traffic API. Also, we test the quality of navigation advice in realistic settings and show that our solution is able to help drivers avoid congested areas in cases where even real-time update methods lead drivers to congested routes.

Distributed spatial simulations commonly employ Bulk Synchronous Parallel model (BSP) implementation. However, implementations using BSP are usually fraught with the straggler problem, where the delay of any worker slows down the entire system. Random stragglers commonly occur due to many reasons: imbalanced workload, operating system scheduling, or communication delays. The straggler problem is further exasperated with increasing parallelism. To reduce the straggler problem and preserve simplicity and scalability advantages of the BSP model, we propose a new parallel model, which we call Priority Asynchronous Parallel (PAP) model. PAP exploits data dependencies of parallel processes to be computed and synchronized based on data priority to the other workers. For further computational improvement, we develop a load balancing and partitioning method, called GridGraph that utilizes the spatial and connectivity properties of the simulation space to reduce the size of exchanged data in addition to balancing the workload among workers. The proposed schemes are implemented and evaluated in a microscopic traffic simulator. Running traffic simulation for Melbourne, Beijing, and New York cities on 80 workers, the simulation achieves a performance speedup of around 47.4% for Melbourne, 52.18% for Beijing, and 65.84% for New York, using PAP model combined with GridGraph partitioning compared to BSP model.

Personalized itinerary recommendation is a complex and time-consuming problem, due to the need to recommend popular attractions that are aligned to the interest preferences of a tourist, and to plan these attraction visits as an itinerary that has to be completed within a specific time limit. Furthermore, many existing itinerary recommendation systems do not automatically determine and consider queuing times at attractions in the recommended itinerary, which varies based on the time of visit to the attraction, e.g., longer queuing times at peak hours. To solve these challenges, we propose the PersQ algorithm for recommending personalized itineraries that take into consideration attraction popularity, user interests and queuing times. We also implement a framework that utilizes geo-tagged photos to derive attraction popularity, user interests and queuing times, which PersQ uses to recommend personalized and queue-aware itineraries. We demonstrate the effectiveness of PersQ in the context of five major theme parks, based on a Flickr dataset spanning nine years. Experimental results show that PersQ outperforms various state-of-the-art baselines, in terms of various queuing-time related metrics, itinerary popularity, user interest alignment, recall, precision and F1-score.

Microscopic traffic simulators are important tools for studying transportation systems as they describe the evolution of traffic to the highest level of detail. A major challenge to microscopic simulators is the slow simulation speed due to the complexity of traffic models. We have developed the Scalable Microscopic Adaptive Road Traffic Simulator (SMARTS), a distributed microscopic traffic simulator that can utilize multiple independent processes in parallel. SMARTS can perform fast large-scale simulations. For example, when simulating 1 million vehicles in an area the size of Melbourne, the system runs 1.14 times faster than real time with 30 computing nodes and 0.2s simulation timestep. SMARTS supports various driver models and traffic rules, such as the car-following model and lane-changing model, which can be driver dependent. It can simulate multiple vehicle types, including bus and tram. The simulator is equipped with a wide range of features that help to customize, calibrate, and monitor simulations. Simulations are accurate and confirm with real traffic behaviours. For example, it achieves 79.1% accuracy in predicting traffic on a 10km freeway 90 minutes into the future. The simulator can be used for predictive traffic advisories as well as traffic management decisions as simulations complete well ahead of real time. SMARTS can be easily deployed to different operating systems as it is developed with the standard Java libraries.

Major navigation companies have resorted to crowdsourcing for obtaining traffic data to improve their services. In fact, a new generation of navigation systems are emerging for the sole purpose of connecting drivers on the road (Waze.com) so that they benefit from each other. The focus of these efforts is on real-time traffic estimation. While obtaining real-time data is desirable and has great benefits, it is still limited, given the potential of the detailed individual data from drivers. For example, commuters in a city have valuable knowledge about their city as a result of their daily commute over many years. Using drivers as mere traffic sensors and aggregating their data desposes off any human knowledge that exists in their choices. In this work, we demonstrate a social navigation system that can help drivers avoid road incidents using expert advice in real-time without explicitly consulting drivers but through an automated system built on top of a location-based social network. In our system, we do not aggregate data from individual users and we instead outsource the task of finding a route that best avoids incidents by looking up drivers' old choices.

The tracking of microblog discussion, on a given topic, is useful for a wide range of higher level applications. Microblog services like Twitter provide a simple keyword based tracking capability, where any tweet containing a keyword is returned. Due to the short length of microblog posts, using a small number of topic specific query words for tracking, would impact recall. Use of a larger number of keywords (compared to regular document retrieval) is generally required in order to obtain good recall, but this would result in a large number of off-topic posts, resulting in low precision. In our work, we consider the scenario of using a large number of query terms to maintain high recall, for automated tracking of a microblog streams. The challenge we address is how to score each of the returned microblogs, with respect to the query, on-line, in an unsupervised manner, so as to identify those that are on topic. To this end, we proposed a new term-scoring expression, which we call Adjusted Information Gain (AIG), and we compare this to other term-scoring expressions: inverse document frequency, Dice, Jaccard and keyword frequency. Our comparisons consider a selection of document-scoring functions applied to roughly 40 million tweets collects over a 20 day period for each of two topics. Our results show significant improvements (from 8%-40% of the area under the ROC curves) to existing term-scoring expressions, depending on topic and specificity, and provide insight into further work in query expansion techniques.

Wireless Sensor Networks (WSNs) provide a low cost option for gathering spatially dense data from different environments. However, WSNs have limited energy resources that hinder the dissemination of the raw data over the network to a central location. This has stimulated research into efficient data mining approaches, which can exploit the restricted computational capabilities of the sensors to model their normal behavior. Having a normal model of the network, sensors can then forward anomalous measurements to the base station. Most of the current data modeling approaches proposed for WSNs require a fixed offline training period and use batch training in contrast to the real streaming nature of data in these networks. In addition they usually work in stationary environments. In this paper we present an efficient online model construction algorithm that captures the normal behavior of the system. Our model is capable of tracking changes in the data distribution in the monitored environment. We illustrate the proposed algorithm with numerical results on both real-life and simulated data sets, which demonstrate the efficiency and accuracy of our approach compared to existing methods.