Partially observable Markov decision processes (POMDPs) are a general mathematical model for sequential decision-making in stochastic environments under state uncertainty. POMDPs are often solved online, which enables the algorithm to adapt to new information in real time. Online solvers typically use bootstrap particle filters based on importance resampling for updating the belief distribution. Since directly sampling from the ideal state distribution given the latest observation and previous state is infeasible, particle filters approximate the posterior belief distribution by propagating states and adjusting weights through prediction and resampling steps. However, in practice, the importance resampling technique often leads to particle degeneracy and sample impoverishment when the state transition model poorly aligns with the posterior belief distribution, especially when the received observation is noisy. We propose an approach that constructs a sequence of bridge distributions between the state-transition and optimal distributions through iterative Monte Carlo steps, better accommodating noisy observations in online POMDP solvers. Our algorithm demonstrates significantly superior performance compared to state-of-the-art methods when evaluated across multiple challenging POMDP domains.

Sequential decision-making in high-dimensional continuous action spaces, particularly in stochastic environments, faces significant computational challenges. We explore this challenge in the traditional offline RL setting, where an agent must learn how to make decisions based on data collected through a stochastic behavior policy. We present \textit{Latent Macro Action Planner} (L-MAP), which addresses this challenge by learning a set of temporally extended macro-actions through a state-conditional Vector Quantized Variational Autoencoder (VQ-VAE), effectively reducing action dimensionality. L-MAP employs a (separate) learned prior model that acts as a latent transition model and allows efficient sampling of plausible actions. During planning, our approach accounts for stochasticity in both the environment and the behavior policy by using Monte Carlo tree search (MCTS). In offline RL settings, including stochastic continuous control tasks, L-MAP efficiently searches over discrete latent actions to yield high expected returns. Empirical results demonstrate that L-MAP maintains low decision latency despite increased action dimensionality. Notably, across tasks ranging from continuous control with inherently stochastic dynamics to high-dimensional robotic hand manipulation, L-MAP significantly outperforms existing model-based methods and performs on par with strong model-free actor-critic baselines, highlighting the effectiveness of the proposed approach in planning in complex and stochastic environments with high-dimensional action spaces.

Many real-world applications require decision-making where the environmental dynamics evolve over time. These non-stationary environments pose significant challenges to traditional decision-making models, which typically assume stationary dynamics. Non-stationary Markov decision processes (NS-MDPs) offer a framework to model and solve decision problems under such changing conditions. However, there are no standardized simulation frameworks for NS-MDPs, as opposed to widely popular frameworks for stationary problems. We present NS-Gym, the first simulation toolkit designed explicitly for NS-MDPs, integrated within the popular Gymnasium framework. In NS-Gym, we segregate the evolution of the environmental parameters that characterize non-stationarity from the agent’s decision-making module, allowing for modular and flexible adaptations to dynamic environments. We review prior work in this domain and present a toolkit encapsulating key problem characteristics and types in NS-MDPs. This toolkit is the first effort to develop a set of standardized interfaces and benchmark problems to enable consistent and reproducible evaluation of algorithms under non-stationary conditions. We also benchmark several algorithmic approaches from prior work on NS-MDPs using NS-Gym. We envision that NS-Gym will enable researchers to study decision-making under non-stationarity by providing standardized interfaces, simulation frameworks, and benchmark problems.

In Partially Observable Markov Decision Processes (POMDPs), maintaining and updating belief distributions over possible underlying states provides a principled way to summarize action-observation history for effective decision-making under uncertainty. As environments grow more realistic, belief distributions develop complexity that standard mathematical models cannot accurately capture, creating a fundamental challenge in maintaining representational accuracy. Despite advances in deep learning and probabilistic modeling, existing POMDP belief approximation methods fail to accurately represent complex uncertainty structures such as high-dimensional, multi-modal belief distributions, resulting in estimation errors that lead to suboptimal agent behaviors. To address this challenge, we present ESCORT (Efficient Stein-variational and sliced Consistency-Optimized Representation for Temporal beliefs), a particle-based framework for capturing complex, multi-modal distributions in high-dimensional belief spaces. ESCORT extends SVGD with two key innovations: correlation-aware projections that model dependencies between state dimensions, and temporal consistency constraints that stabilize updates while preserving correlation structures. This approach retains SVGD's attractive-repulsive particle dynamics while enabling accurate modeling of intricate correlation patterns. Unlike particle filters prone to degeneracy or parametric methods with fixed representational capacity, ESCORT dynamically adapts to belief landscape complexity without resampling or restrictive distributional assumptions. We demonstrate ESCORT's effectiveness through extensive evaluations on both POMDP domains and synthetic multi-modal distributions of varying dimensionality, where it consistently outperforms state-of-the-art methods in terms of belief approximation accuracy and downstream decision quality.

Effective traffic incident management is critical for road safety and operational efficiency. Yet, many transportation agencies rely on reactionary methods, where incidents are reported by human agents and managed through rule-based frameworks like traditional Traffic Incident Management (TIM) systems. However, these are vulnerable to human error, oversight, and delays during high-stress conditions. Although recent initiatives incorporating real-time sensor data for corridor monitoring and enhanced roadway information systems represent strides toward modernization, these systems often still require substantial human intervention. Recent advancements in graph-based deep learning models offer promising potential for addressing the limitations of traditional methods. While state-of-the-art models exist, the complexities of incident localization within dynamic and interconnected road networks, along with limited availability of high-quality labeled data and variability in real-time traffic measurements, are still open challenges. To address these, we propose the Traffic Response Anomaly Capture Engine (TRACE), a novel approach that combines graph neural networks, transformers, and probabilistic normalizing flows to accurately detect and localize traffic anomalies in real time. TRACE captures spatial-temporal dependencies, manages data uncertainty, and enhances automation, supporting more precise and timely incident localization. Our approach is validated on real-world traffic data and improved incident localization by 0.6 miles (17%) than SOTA methods while maintaining similar incident detection accuracy and mean detection delay.

Smart cities operate on computational predictive frameworks that collect aggregate and utilize data from large-scale sensor networks. However these frameworks are prone to multiple sources of data and algorithmic bias which often lead to unfair prediction results. In this work we first demonstrate that bias persists at a micro-level both temporally and spatially by studying real city data from Chattanooga TN. To alleviate the issue of such bias we introduce FairGuard a micro-level temporal logic-based approach for fair smart city policy adjustment and generation in complex temporal-spatial domains. The FairGuard framework consists of two phases. First we develop a static generator that is able to reduce data bias based on temporal logic conditions by minimizing correlations between selected attributes. Second to ensure fairness in predictive algorithms we design a dynamic component to regulate prediction results and generate future fair predictions by harnessing logic rules. To navigate potential conflicts among these single fairness rules including logical contradictions and data interference we formulate detection strategies grounded in Satisfiability Modulo Theories (SMT) across both logic and data levels. Furthermore acknowledging the limitations of fairness rules focused on a single attribute we enhance the Static FairGuard to accommodate heterogeneous fairness rules that simultaneously consider multiple protected attributes. In addition we develop an interactive online visualizer that displays the adjustments made to correct unfair city states thereby improving fairness alongside the prediction outcomes from the dynamic component. Evaluations showcase that logic-enabled Static FairGuard can effectively reduce the biased correlations while Dynamic FairGuard can guarantee fairness on protected groups at runtime with minimal impact on overall performance.

Public bus transit systems provide critical transportation services for large sections of modern communities. On-time performance and maintaining the reliable quality of service is therefore very important. Unfortunately, disruptions caused by overcrowding, vehicular failures, and road accidents often lead to service performance degradation. Though transit agencies keep a limited number of vehicles in reserve and dispatch them to relieve the affected routes during disruptions, the procedure is often ad-hoc and has to rely on human experience and intuition to allocate resources (vehicles) to affected trips under uncertainty. In this article, we describe a principled approach using non-myopic sequential decision procedures to solve the problem and decide (a) if it is advantageous to anticipate problems and proactively station transit buses near areas with high-likelihood of disruptions and (b) decide if and which vehicle to dispatch to a particular problem. Our approach was developed in partnership WeGo Public Transit, a public transportation agency based in Nashville, Tennessee and models the system as a semi-Markov decision problem (solved as a Monte-Carlo tree search procedure) and shows that it is possible to obtain an answer to these two coupled decision problems in a way that maximizes the overall reward (number of people served). We sample many possible futures from generative models; each is assigned to a tree and processed using root parallelization. We validate our approach with both real-world and scaled-up data from two agencies in Tennessee. Our experiments show that the proposed framework serves 2% more passengers while reducing deadhead miles by 40%. Finally, we introduce Vectura, a dashboard providing transit dispatchers a complete view of the transit system at a glance along with access to our developed tools.

More than 5 million children under five years die from largely preventable or treatable medical conditions every year, with an overwhelmingly large proportion of deaths occurring in underdeveloped countries with low vaccination uptake. One of the United Nations' sustainable development goals (SDG 3) aims to end preventable deaths of newborns and children under five years of age. We focus on Nigeria, where the rate of infant mortality is appalling. In particular, low vaccination uptake in Nigeria is a major driver of more than 2,000 daily deaths of children under the age of five years. In this paper, we describe our collaboration with government partners in Nigeria to deploy ADVISER: AI-Driven Vaccination Intervention Optimiser. The framework, based on an integer linear program that seeks to maximize the cumulative probability of successful vaccination, is the first successful deployment of an AI-enabled toolchain for optimizing the allocation of health interventions in Nigeria. In this paper, we provide a background of the ADVISER framework and present results, lessons, and success stories of deploying ADVISER to more than 13,000 families in the state of Oyo, Nigeria.

Monte Carlo tree search (MCTS) is one of the most capable online search algorithms for sequential planning tasks, with significant applications in areas such as resource allocation and transit planning. Despite its strong performance in real-world deployment, the inherent complexity of MCTS makes it challenging to understand for users without technical background. This paper considers the use of MCTS in transportation routing services, where the algorithm is integrated to develop optimized route plans. These plans are required to meet a range of constraints and requirements simultaneously, further complicating the task of explaining the algorithm’s operation in real-world contexts. To address this critical research gap, we introduce a novel computation tree logic-based explainer for MCTS. Our framework begins by taking user-defined requirements and translating them into rigorous logic specifications through the use of language templates. Then, our explainer incorporates a logic verification and quantitative evaluation module that validates the states and actions traversed by the MCTS algorithm. The outcomes of this analysis are then rendered into human-readable descriptive text using a second set of language templates. The user satisfaction of our approach was assessed through a survey with 82 participants. The results indicated that our explanatory approach significantly outperforms other baselines in user preference.

An emergency responder management (ERM) system dispatches responders, such as ambulances, when it receives requests for medical aid. ERM systems can also proactively reposition responders between predesignated waiting locations to cover any gaps that arise due to the prior dispatch of responders or significant changes in the distribution of anticipated requests. Optimal repositioning is computationally challenging due to the exponential number of ways to allocate responders between locations and the uncertainty in future requests. The state-of-the-art approach in proactive repositioning is a hierarchical approach based on spatial decomposition and online Monte Carlo tree search, which may require minutes of computation for each decision in a domain where seconds can save lives. We address the issue of long decision times by introducing a novel reinforcement learning (RL) approach, based on the same hierarchical decomposition, but replacing online search with learning. To address the computational challenges posed by large, variable-dimensional, and discrete state and action spaces, we propose: (1) actor-critic based agents that incorporate transformers to handle variable-dimensional states and actions, (2) projections to fixed-dimensional observations to handle complex states, and (3) combinatorial techniques to map continuous actions to discrete allocations. We evaluate our approach using real-world data from two U.S. cities, Nashville, TN and Seattle, WA. Our experiments show that compared to the state of the art, our approach reduces computation time per decision by three orders of magnitude, while also slightly reducing average ambulance response time by 5 seconds.

Public transit systems are paramount in lowering carbon emissions and reducing urban congestion for environmental sustainability. However, overcrowding has adverse effects on the quality of service, passenger experience, and overall efficiency of public transit causing a decline in the usage of public transit systems. Therefore, it is crucial to identify and forecast potential windows of overcrowding to improve passenger experience and encourage higher ridership. Predicting ridership is a complex task, due to the inherent noise of collected data and the sparsity of overcrowding events. Existing studies in predicting public transit ridership consider only a static depiction of bus networks. We address these issues by first applying a data processing pipeline that cleans noisy data and engineers several features for training. Then, we address sparsity by converting the network to a dynamic graph and using a graph convolutional network, incorporating temporal, spatial, and auto-regressive features, to learn generalizable patterns for each route. Finally, since conventional loss functions like categorical cross-entropy have limitations in addressing class imbalance inherent in ridership data, our proposed approach uses focal loss to refine the prediction focus on less frequent yet task-critical overcrowding instances. Our experiments, using real-world data from our partner agency, show that the proposed approach outperforms existing state-of-the-art baselines in terms of accuracy and robustness.

The increasing popularity of electronic vehicles has spurred a demand for EV charging infrastructure. In the United States alone, over 160,000 public and private charging ports have been installed. This has stoked fear of potential grid issues in the future. Meanwhile, companies, specifically building owners are also seeing the opportunity to leverage EV batteries as energy stores to serve as buffers against the electric grid. The main idea is to influence and control charging behavior to provide a certain level of energy resiliency and demand responsiveness to the building from grid events while ensuring that they meet the demands of EV users. However, managing and co-optimizing energy requirements of EVs and cost-saving measures of building owners is a difficult task. First, user behavior and grid uncertainty contribute greatly to the potential effectiveness of different policies. Second, different charger configurations can have drastically different effects on the cost. Therefore, we propose a complete end-to-end discrete event simulator for vehicle-to-building charging optimization. This software is aimed at building owners and EV manufacturers such as Nissan, looking to deploy their charging stations with state-of-the-art optimization algorithms. We provide a complete solution that allows the owners to train, evaluate, introduce uncertainty, and benchmark policies on their datasets. Lastly, we discuss the potential for extending our work with other vehicle-to-grid deployments.

This workshop will bring together researchers and practitioners across different strands of data science research and a wide range of important real-world application domains. The objective is to share the current state of research and practice, explore future work directions, and create collaboration opportunities. In addition, the workshop will emphasize highlighting data science approaches for tackling the United Nations Sustainable Development Goals (see preliminary agenda below). The organizers believe that data science research has an important role to play in providing unique insights about critical challenges faced by marginalized communities around the world; we encourage submissions from both data science researchers as well as social workers, agronomists, epidemiologists, health policy researchers, and other domain experts who are interested in engaging with the SIGKDD community.

Timely and accurate detection of events affecting the stability and reliability of power transmission systems is crucial for safe grid operation. This paper presents an efficient unsupervised machine-learning algorithm for event detection using a combination of discrete wavelet transform (DWT) and convolutional autoencoders (CAE) with synchrophasor phasor measurements. These measurements are collected from a hardware-in-the-loop testbed setup equipped with a digital real-time simulator. Using DWT, the detail coefficients of measurements are obtained. Next, the decomposed data is then fed into the CAE that captures the underlying structure of the transformed data. Anomalies are identified when significant errors are detected between input samples and their reconstructed outputs. We demonstrate our approach on the IEEE-14 bus system considering different events such as generator faults, line-to-line faults, line-to-ground faults, load shedding, and line outages simulated on a real-time digital simulator (RTDS). The proposed implementation achieves a classification accuracy of 97.7%, precision of 98.0%, recall of 99.5%, F1 Score of 98.7%, and proves to be efficient in both time and space requirements compared to baseline approaches.

Public transit is a vital mode of transportation in urban areas, and its efficiency is crucial for the daily commute of millions of people. To improve the reliability and predictability of transit systems, researchers have developed separate single-task learning models to predict the occupancy and delay of buses at the stop or route level. However, these models provide a narrow view of delay and occupancy at each stop and do not account for the correlation between the two. We propose a novel approach that leverages broader generalizable patterns governing delay and occupancy for improved prediction. We introduce a multitask learning toolchain that takes into account General Transit Feed Specification feeds, Automatic Passenger Counter data, and contextual temporal and spatial information. The toolchain predicts transit delay and occupancy at the stop level, improving the accuracy of the predictions of these two features of a trip given sparse and noisy data. We also show that our toolchain can adapt to fewer samples of new transit data once it has been trained on previous routes/trips as compared to state-of-the-art methods. Finally, we use actual data from Chattanooga, Tennessee, to validate our approach. We compare our approach against the state-of-the-art methods and we show that treating occupancy and delay as related problems improves the accuracy of the predictions. We show that our approach improves delay prediction significantly by as much as 4% in F1 scores while producing equivalent or better results for occupancy.

The ability to accurately predict public transit ridership demand benefits passengers and transit agencies. Agencies will be able to reallocate buses to handle under or over-utilized bus routes, improving resource utilization, and passengers will be able to adjust and plan their schedules to avoid overcrowded buses and maintain a certain level of comfort. However, accurately predicting occupancy is a non-trivial task. Various reasons such as heterogeneity, evolving ridership patterns, exogenous events like weather, and other stochastic variables, make the task much more challenging. With the progress of big data, transit authorities now have access to real-time passenger occupancy information for their vehicles. The amount of data generated is staggering. While there is no shortage in data, it must still be cleaned, processed, augmented, and merged before any useful information can be generated. In this paper, we propose the use and fusion of data from multiple sources, cleaned, processed, and merged together, for use in training machine learning models to predict transit ridership. We use data that spans a 2-year period (2020-2022) incorporating transit, weather, traffic, and calendar data. The resulting data, which equates to 17 million observations, is used to train separate models for the trip and stop level prediction. We evaluate our approach on real-world transit data provided by the public transit agency of Nashville, TN. We demonstrate that the trip level model based on Xgboost and the stop level model based on LSTM outperform the baseline statistical model across the entire transit service day.

While cloud computing is the current standard for outsourcing computation, it can be prohibitively expensive for cities and infrastructure operators to deploy services. At the same time, there are underutilized computing resources within cities and local edge-computing deployments. Using these slack resources may enable significantly lower pricing than comparable cloud computing; such resources would incur minimal marginal expenditure since their deployment and operation are mostly sunk costs. However, there are challenges associated with using these resources. First, they are not effectively aggregated or provisioned. Second, there is a lack of trust between customers and suppliers of computing resources, given that they are distinct stakeholders and behave according to their own interests. Third, delays in processing inputs may diminish the value of the applications. To resolve these chal-lenges, we introduce an architecture combining a distributed trusted computing mechanism, such as a blockchain, with an efficient messaging system like Apache Pulsar. Using this architecture, we design a decentralized computation market where customers and suppliers make offers to deploy and host applications. The proposed architecture can be realized using any trusted computing mechanism that supports smart contracts, and any messaging framework with the necessary features. This combination ensures that the market is robust without incurring the input processing delays that limit other blockchain based solutions. We evaluate the market protocol using game-theoretic analysis to show that deviation from the protocol is discouraged. Finally, we assess the performance of a prototype implementation based on experiments with a streaming computer-vision application.

More than 5 million children under five years die from largely preventable or treatable medical conditions every year, with an overwhelmingly large proportion of deaths occurring in under-developed countries with low vaccination uptake. One of the United Nations' sustainable development goals (SDG 3) aims to end preventable deaths of newborns and children under five years of age. We focus on Nigeria, where the rate of infant mortality is appalling. We collaborate with HelpMum, a large non-profit organization in Nigeria, to design and optimize the allocation of heterogeneous health interventions under uncertainty to increase vaccination uptake, the first such collaboration in Nigeria. Our framework, ADVISER: AI-Driven Vaccination Intervention Optimiser, is based on an integer linear program that seeks to maximize the cumulative probability of successful vaccination. Our optimization formulation is intractable in practice. We present a heuristic approach that enables us to solve the problem for real-world use-cases. We also present theoretical bounds for the heuristic method. Finally, we show that the proposed approach outperforms baseline methods in terms of vaccination uptake through experimental evaluation. HelpMum is currently planning a pilot program based on our approach to be deployed in the largest city of Nigeria, which would be the first deployment of an AI-driven vaccination uptake program in the country and hopefully, pave the way for other data-driven programs to improve health outcomes in Nigeria.

Vehicle routing problems (VRPs) can be divided into two major categories: offline VRPs, which consider a given set of trip requests to be served, and online VRPs, which consider requests as they arrive in real-time. Based on discussions with public transit agencies, we identify a real-world problem that is not addressed by existing formulations: booking trips with flexible pickup windows (e.g., 3 hours) in advance (e.g., the day before) and confirming tight pickup windows (e.g., 30 minutes) at the time of booking. Such a service model is often required in paratransit service settings, where passengers typically book trips for the next day over the phone. To address this gap between offline and online problems, we introduce a novel formulation, the offline vehicle routing problem with online bookings. This problem is very challenging computationally since it faces the complexity of considering large sets of requests—similar to offline VRPs—but must abide by strict constraints on running time—similar to online VRPs. To solve this problem, we propose a novel computational approach, which combines an anytime algorithm with a learning-based policy for real-time decisions. Based on a paratransit dataset obtained from the public transit agency of Chattanooga, TN, we demonstrate that our novel formulation and computational approach lead to significantly better outcomes in this setting than existing algorithms.

Resource allocation under uncertainty is a classic problem in city-scale cyber-physical systems. Consider emergency response, where urban planners and first responders optimize the location of ambulances to minimize expected response times to incidents such as road accidents. Typically, such problems involve sequential decision making under uncertainty and can be modeled as Markov (or semi-Markov) decision processes. The goal of the decision maker is to learn a mapping from states to actions that can maximize expected rewards. While online, offline, and decentralized approaches have been proposed to tackle such problems, scalability remains a challenge for real world use cases. We present a general approach to hierarchical planning that leverages structure in city level CPS problems for resource allocation. We use emergency response as a case study and show how a large resource allocation problem can be split into smaller problems. We then use Monte Carlo planning for solving the smaller problems and managing the interaction between them. Finally, we use data from Nashville, Tennessee, a major metropolitan area in the United States, to validate our approach. Our experiments show that the proposed approach outperforms state-of-the-art approaches used in the field of emergency response.

Pre-processing spatial data for machine learning applications often includes combining different datasets into a form usable by the machine learning algorithms. Spatial data is generally available in two representations, raster and vector. The best data science and machine learning applications need to combine multiple datasets of both representations which is a data and compute intensive problem. This paper proposes a formal raster-vector join operator, Raptor Join, that can bridge the gap between raster and vector data. It is modeled as a relational join operator in Spark that can be easily combined with other operators, while also offering the advantage of in-situ processing. To implement the Raptor join operator efficiently, we propose a novel Flash index that has a low memory requirement and can process the entire operation with one data scan. We run an extensive experimental evaluation on large scale satellite data with up-to a trillion pixels, and big vector data with up-to hundreds of millions of segments and billions of points, and show that the proposed method can scale to big data with up-to three orders of magnitude performance gain over baselines.

Spatial data is traditionally represented using two data models, raster and vector. Raster data refers to satellite imagery while vector data includes GPS data, Tweets, and regional boundaries. While there are many real-world applications that need to process both raster and vector data concurrently, state-of-the-art systems are limited to processing one of these two representations while converting the other one which limits their scalability. This paper draws the attention of the research community to the research problems that emerge from the concurrent processing of raster and vector data. It describes three real-world applications and explains their computation and access patterns for raster and vector data. Additionally, it runs an extensive experimental evaluation using state-of-the-art big spatial data systems with raster data of up-to a trillion pixels, and vector data with up-to hundreds of millions of edges. The results show that while most systems can analyze raster and vector concurrently, but they have limited scalability for large-scale data.

Emergency response is highly dependent on the time of incident reporting. Unfortunately, the traditional approach to receiving incident reports (e.g., calling 911 in the USA) has time delays. Crowdsourcing platforms such as Waze provide an opportunity for early identification of incidents. However, detecting incidents from crowdsourced data streams is difficult due to the challenges of noise and uncertainty associated with such data. Further, simply optimizing over detection accuracy can compromise spatial-temporal localization of the inference, thereby making such approaches infeasible for real-world deployment. This paper presents a novel problem formulation and solution approach for practitioner-centered incident detection using crowdsourced data by using emergency response management as a case-study. The proposed approach CROME (Crowdsourced Multi-objective Event Detection) quantifies the relationship between the performance metrics of incident classification (e.g., F1 score) and the requirements of model practitioners (e.g., 1 km. radius for incident detection). First, we show how crowdsourced reports, ground-truth historical data, and other relevant determinants such as traffic and weather can be used together in a Convolutional Neural Network (CNN) architecture for early detection of emergency incidents. Then, we use a Pareto optimization-based approach to optimize the output of the CNN in tandem with practitioner-centric parameters to balance detection accuracy and spatial-temporal localization. Finally, we demonstrate the applicability of this approach using crowdsourced data from Waze and traffic accident reports from Nashville, TN, USA. Our experiments demonstrate that the proposed approach outperforms existing approaches in incident detection while simultaneously optimizing the needs for real-world deployment and usability.

Emergency Response Management (ERM) necessitates the use of models capable of predicting the spatial-temporal likelihood of incident occurrence. These models are used for proactive stationing in order to reduce overall response time. Traditional methods simply aggregate past incidents over space and time; such approaches fail to make useful short-term predictions when the spatial region is large and focused on fine-grained spatial entities like interstate highway networks. This is partially due to the sparsity of incidents with respect to space and time. Further, accidents are affected by several covariates. Collecting, cleaning, and managing multiple streams of data from various sources is challenging for large spatial areas. In this paper, we highlight how this problem is being solved in collaboration with the Tennessee Department of Transportation (TDOT) to improve ERM in the state of Tennessee. Our pipeline, based on a combination of synthetic resampling, clustering, and data mining techniques, can efficiently forecast the spatio-temporal dynamics of accident occurrence, even under sparse conditions. Our pipeline uses data related to roadway geometry, weather, historical accidents, and traffic to aid accident forecasting. To understand how our forecasting model can affect allocation and dispatch, we improve and employ a classical resource allocation approach. Experimental results show that our approach can noticeably reduce response times and the number of unattended incidents in comparison to current approaches followed by first responders. The developed pipeline is efficacious, applicable in practice, and open-source.

Spatio-temporal incident prediction is a central issue in law enforcement, with applications in fighting crimes like poaching, human trafficking, illegal fishing, burglaries and smuggling. However, state of the art approaches fail to account for evasion in response to predictive models, a common form of which is spatial shift in incident occurrence. We present a general approach for incident forecasting that is robust to spatial shifts. We propose two techniques for solving the resulting robust optimization problem: first, a constraint generation method guaranteed to yield an optimal solution, and second, a more scalable gradient-based approach. We then apply these techniques to both discrete-time and continuous-time robust incident forecasting. We evaluate our algorithms on two different real-world datasets, demonstrating that our approach is significantly more robust than conventional methods.