This paper explores the joint task of machine translation and sentence compression, emphasizing its application in subtitle generation for broadcast and live media for low-resource languages and hardware. We develop CLSC (Cross-Lingual Sentence Compression), a system trained on openly available parallel corpora organized by compression ratios, where the target length is constrained to a fraction of the source sentence length. We present two training methods: 1) Multiple Models (MM), where individual models are trained separately for each compression ratio, and 2) a Controllable Model (CM), a single model per language using a compression token to encode length constraints. We evaluate both subtitle data and transcriptions from the EuroParl corpus. To accommodate low-resource settings, we constrain data sampling for training and show results for transcriptions in French, Hungarian, Lithuanian, and Polish and subtitles in Albanian, Basque, Malay, and Norwegian. Our models preserve high semantic meaning and metric evaluations for compressed contexts.

To address poor working conditions and promote United Nations’ sustainable development goal 8.8, “protect labour rights and promote safe working environments for all workers [...]”, government agencies around the world conduct labour inspections. To carry out these inspections, inspectors traditionally use paper-based checklists as a means to survey individual organisations for working environment violations. Currently, these checklists are created by domain experts, but recent research indicates that machine learning (ML) could be used to generate dynamic checklists to increase inspection efficiency. A drawback with the dynamic checklists is that they are complex and could be difficult to understand for inspectors. They have also never been field-tested. In this paper, we therefore propose user-oriented explanation methods for Context-aware Bayesian Case-Based Reasoning (CBCBR), which is the current state-of-art ML method for generating dynamic checklists. We also introduce a prototype of CBCBR and present a field study where we test it in real-world labour inspections. The results from the study indicate that using the explainable dynamic checklists increases the efficiency of the labour inspections, and inspectors also report that they find the checklists useful. The results also suggest that current ML evaluation methods, where model prediction performance is evaluated on existing data, may not fully reflect the real-world field performance of checklists.

The optimization of Emergency Medical Services is a central issue in modern healthcare systems. With this in focus, we study a data set containing medical emergencies for the years 2015--2019 from Oslo and Akershus, Norway. By developing a discrete trace-based simulation model based on the data set, we compute average response times that are used to optimize ambulance allocations to stations in the region. We study several metaheuristics, specifically genetic, stochastic local search, and memetic algorithms. These metaheuristics are tested using the simulation to optimize ambulance allocations, considering response times. The algorithms are compared against each other and a set of baseline allocation models over different time periods. The main results of our experimental simulation study are that: (i) the metaheuristics generally outperform the simpler baselines, (ii) the best-performing metaheuristic is the genetic algorithm, and (iii) the performance difference between the metaheuristics and the simpler baselines increases in situations with high demand on ambulances. Finally, we present suggestions for future work that may help to further improve upon the current state-of-the-art.

With a focus on both the fitness and cost of subset selection, we study stochastic local search (SLS) heuristics in this paper. In particular, we consider subset selection problems where the cost of fitness function evaluation needs to be accounted for. Here, cost can be fitness evaluation's computation time or energy cost. We propose and study an SLS method, SLS4CFF, tailored to such problems. SLS4CFF ("SLS for costly fitness functions") is an amalgamation of several existing heuristics. We develop a homogeneous Markov chain model that explicitly represents both fitness and cost of subset selection with SLS4CFF. This Markov chain, which can be lumped or compressed for certain fitness and cost functions, enables us to better understand and analyze hyperparameter optimization in a principled manner, via expected hitting times. Studies with synthetic and real-world problems improve the understanding of SLS and demonstrate the importance of cost-awareness.

Every year there are 1.9 million deaths world-wide attributed to occupational health and safety risk factors. To address poor working conditions and fulfill UN's SDG 8, "protect labour rights and promote safe working environments for all workers", governmental agencies conduct labour inspections, using checklists to survey individual organisations for working environment violations. Recent research highlights the benefits of using machine learning for creating checklists. However, the current methods only create static checklists and do not adapt them to new information that surfaces during use. In contrast, we propose a new method called Context-aware Bayesian Case-Based Reasoning (CBCBR) that creates dynamic checklists. These checklists are continuously adapted as the inspections progress, based on how they are answered. Our evaluations show that CBCBR's dynamic checklists outperform static checklists created via the current state-of-the-art methods, increasing the expected number of working environment violations found in the labour inspections.

Among United Nations' 17 Sustainable Development Goals (SDGs), we highlight SDG 8 on Decent Work and Economic Growth. Specifically, we consider how to achieve subgoal 8.8, "protect labour rights and promote safe working environments for all workers [...]", in light of poor health, safety and environment (HSE) conditions being a widespread problem at workplaces. In EU alone, it is estimated that more than 4000 deaths occur each year due to poor working conditions. To handle the problem and achieve SDG 8, governmental agencies conduct labour inspections and it is therefore essential that these are carried out efficiently. Current research suggests that machine learning (ML) can be used to improve labour inspections, for instance by selecting organisations for inspections more effectively. However, the research in this area is very limited, in part due to a lack of publicly available data. Consequently, we introduce a new dataset called the Labour Inspection Checklists Dataset (LICD), which we have made publicly available. LICD consists of 63634 instances where each instance is an inspection conducted by the Norwegian Labour Inspection Authority. LICD has 577 features and labels. The dataset provides several ML research opportunities; we discuss two demonstration experiments. One experiment deals with the problem of selecting a relevant checklist for inspecting a given target organisation. The other experiment concerns the problem of predicting HSE violations, given a specific checklist and a target organisation. Our experimental results, while promising, suggest that achieving good ML classification performance is difficult for both problems. This motivates future research to improve ML performance, inspire other data analysis efforts, and ultimately achieve SDG 8.

We propose a novel framework for structured bandits, which we call an influence diagram bandit. Our framework uses a graphical model to capture complex statistical dependencies between actions, latent variables, and observations; and thus unifies and extends many existing models, such as combinatorial semi-bandits, cascading bandits, and low-rank bandits. We develop novel online learning algorithms that learn to act efficiently in our models. The key idea is to track a structured posterior distribution of model parameters, either exactly or approximately. To act, we sample model parameters from their posterior and then use the structure of the influence diagram to find the most optimistic action under the sampled parameters. We empirically evaluate our algorithms in three structured bandit problems, and show that they perform as well as or better than problem-specific state-of-the-art baselines.

Finding the most likely path to a set of failure states is important to the analysis of safety-critical systems that operate over a sequence of time steps, such as aircraft collision avoidance systems and autonomous cars. In many applications such as autonomous driving, failures cannot be completely eliminated due to the complex stochastic environment in which the system operates. As a result, safety validation is not only concerned about whether a failure can occur, but also discovering which failures are most likely to occur. This article presents adaptive stress testing (AST), a framework for finding the most likely path to a failure event in simulation. We consider a general black box setting for partially observable and continuous-valued systems operating in an environment with stochastic disturbances. We formulate the problem as a Markov decision process and use reinforcement learning to optimize it. The approach is simulation-based and does not require internal knowledge of the system, making it suitable for black-box testing of large systems. We present different formulations depending on whether the state is fully observable or partially observable. In the latter case, we present a modified Monte Carlo tree search algorithm that only requires access to the pseudorandom number generator of the simulator to overcome partial observability. We also present an extension of the framework, called differential adaptive stress testing (DAST), that can find failures that occur in one system but not in another. This type of differential analysis is useful in applications such as regression testing, where we are concerned with finding areas of relative weakness compared to a baseline. We demonstrate the effectiveness of the approach on an aircraft collision avoidance application, where a prototype aircraft collision avoidance system is stress tested to find the most likely scenarios of near mid-air collision.

Dialog response selection is an important step towards natural response generation in conversational agents. Existing work on neural conversational models mainly focuses on offline supervised learning using a large set of context-response pairs. In this paper, we focus on online learning of response selection in retrieval-based dialog systems. We propose a contextual multi-armed bandit model with a nonlinear reward function that uses distributed representation of text for online response selection. A bidirectional LSTM is used to produce the distributed representations of dialog context and responses, which serve as the input to a contextual bandit. In learning the bandit, we propose a customized Thompson sampling method that is applied to a polynomial feature space in approximating the reward. Experimental results on the Ubuntu Dialogue Corpus demonstrate significant performance gains of the proposed method over conventional linear contextual bandits. Moreover, we report encouraging response selection performance of the proposed neural bandit model using the Recall@k metric for a small set of online training samples.

Soil moisture is critical to agriculture, ecology, and certain natural disasters. Existing soil moisture models often fail to predict soil moisture accurately for time periods greater than a few hours. To tackle this problem, we introduce in this paper two novel models, the Naive Accumulative Representation (NAR) and the Additive Exponential Accumulative Representation (AEAR). The parameters in these models reflect hydrological redistribution processes of gravity and suction. We validate our models using soil moisture and rainfall time series data collected from a steep gradient post-wildfire site in Southern California. Data analysis is challenging, since rapid landscape change in steep, burned hillslopes is typically observed in response to even small to moderate rain events. We found that the AEAR model fits the data well for three distinct soil textures at different depths below the ground surface (at 5cm, 15cm, and 30cm). Similar strong results are demonstrated in controlled soil moisture experiments. Our recommended AEAR model has been validated as effective and useful by earth scientists, giving better forecasts than existing models for time horizons of 10 to 24 hours.

The goal of image captioning via machine learning is to automatically learn to provide a free-form description of an image, while focusing on the significant objects in an image. Inspired by recent work on attention in image captioning, we study in this paper different attention mechanisms within a deep learning setting. In contrast to previous research on attention models which focus on applying attention to the image modality, we introduce three language-based attention models. These language-based attention models, which we developed iteratively from simpler RNN-and LSTM-based baseline models, consist of two sub-networks: a deep recurrent neural network for the language modality and a convolutional neural network for the image modality. The language-based attention models learn a joint representation of the language and image modalities, given the image and the previous words in the caption. At test time, novel captions are produced from this learned distribution. We provide a comparative quantitative and qualitative analysis of our three language-based attention models, which outperform the simple baseline models. We validate the effectiveness of our attention models with state-of-the-art performance on the Flickr8k dataset.

Deep neural networks, including recurrent networks, have been successfully applied to human activity recognition. Unfortunately, the final representation learned by recurrent networks might encode some noise (irrelevant signal components, unimportant sensor modalities, etc.). Besides, it is difficult to interpret the recurrent networks to gain insight into the models' behavior. To address these issues, we propose two attention models for human activity recognition: temporal attention and sensor attention. These two mechanisms adaptively focus on important signals and sensor modalities. To further improve the understandability and mean Fl score, we add continuity constraints, considering that continuous sensor signals are more robust than discrete ones. We evaluate the approaches on three datasets and obtain state-of-the-art results. Furthermore, qualitative analysis shows that the attention learned by the models agree well with human intuition.

Understanding heterogeneous multivariate time series data is important in many applications ranging from smart homes to aviation. Learning models of heterogeneous multivariate time series that are also human-interpretable is challenging and not adequately addressed by the existing literature. We propose grammar-based decision trees (GBDTs) and an algorithm for learning them. GBDTs extend decision trees with a grammar framework. Logical expressions derived from a context-free grammar are used for branching in place of simple thresholds on attributes. The added expressivity enables support for a wide range of data types while retaining the interpretability of decision trees. In particular, when a grammar based on temporal logic is used, we show that GBDTs can be used for the interpretable classification of high-dimensional and heterogeneous time series data. Furthermore, we show how GBDTs can also be used for categorization, which is a combination of clustering and generating interpretable explanations for each cluster. We apply GBDTs to analyze the classic Australian Sign Language dataset as well as data on near mid-air collisions (NMACs). The NMAC data comes from aircraft simulations used in the development of the next-generation Airborne Collision Avoidance System (ACAS X).

Labeled data used for training activity recognition classifiers are usually limited in terms of size and diversity. Thus, the learned model may not generalize well when used in real-world use cases. Semi-supervised learning augments labeled examples with unlabeled examples, often resulting in improved performance. However, the semi-supervised methods studied in the activity recognition literatures assume that feature engineering is already done. In this paper, we lift this assumption and present two semi-supervised methods based on convolutional neural networks (CNNs) to learn discriminative hidden features. Our semi-supervised CNNs learn from both labeled and unlabeled data while also performing feature learning on raw sensor data. In experiments on three real world datasets, we show that our CNNs outperform supervised methods and traditional semi-supervised learning methods by up to 18% in mean F1-score (Fm).

Deep neural networks (DNNs) are popular in diverse fields such as computer vision and natural language processing. DNN inference tasks are emerging as a service provided by cloud computing environments. However, cloud-hosted DNN inference faces new challenges in workload scheduling for the best Quality of Service (QoS), due to dependence on batch size, model complexity and resource allocation. This paper represents the QoS metric as a utility function of response delay and inference accuracy. We first propose a simple and effective heuristic approach that keeps low response delay and satisfies the requirement on processing throughput. Then we describe an advanced deep reinforcement learning (RL) approach that learns to schedule from experience. The RL scheduler is trained to maximize QoS, using a set of system statuses as the input to the RL policy model. Our approach performs scheduling actions only when there are free GPUs, thus reduces scheduling overhead over common RL schedulers that run at every continuous time step. We evaluate the schedulers on a simulation platform and demonstrate the advantages of RL over heuristics.

This work considers that multiple mobile clients offload various continuous sensing applications with end-to-end delay constraints, to a cluster of machines as the server. Contention for shared computing resources on a server can result in delay degradation and application malfunction. We present ATOMS (Accurate Timing prediction and Offloading for Mobile Systems), a framework to mitigate multi-tenant resource contention and to improve delay using a two-phase Plan-Schedule approach. The planning phase includes methods to predict future workloads from all clients, to estimate contention, and to devise offloading schedule to reduce contention. The scheduling phase dispatches arriving offloaded workload to the server machine that minimizes contention, based on the running workloads on each machine.

Soil moisture plays a crucial part in earth science, with impact on agriculture, ecology, hydrology, landslides, and water resources. Extremes in soil moisture, which we denote as peaks and valleys, caused by heavy rainfalls and subsequent dry weather, are very important when predicting future soil moisture or even landslides. Existing methods, like moving averages, have limitations when it comes to smoothing time series data while preserving peaks and valleys. In this work, we propose a novel method, HyperSTL, for extrema-preserving smoothing of soil moisture time series. The method optimizes an existing time series decomposition technique, Seasonal Decomposition of Time Series by Loess (STL). HyperSTL optimizes STL's control parameters, which we call hyperparameters, using an objective function over the decomposed components. We demonstrate in experiments with nine soil moisture datasets that using HyperSTL generally results in improved predictions compared to using other smoothing methods.

Similarity-based recommender systems can be used to pre-compute distance between item pairs, and then to quickly recommend similar items to users. The content-based approach to similarity uses the item's description, which in movies could mean genre, director or cast. These similarity methods are often built with unsupervised learning, which means the notion of similarity is defined by those who write the method. This notion may not match that of the users, resulting in poor user experience. In this paper we used user-contributed labels representing perceived similarity between movies to build a supervised content-based (CB) model for movie recommendations. Our user study shows that the CB method with human perception factored in was significantly preferred over the CB model without.

Mobile devices, such as smartphones and tablets, have become prevalent given their ample functionality brought by a variety of applications. Unfortunately, these devices face security and privacy threats due to unauthorized access. Ordinary protection mechanisms such as passcode and fingerprint verification are widely employed to mitigate the threats. To achieve strong security without sacrificing usability, extensive research efforts have been devoted to continuous authentication through passive sensing and behavior modeling. Nowadays, more and more users own multiple devices. This trend presents opportunities for further optimization of authentication across devices. In this paper, we conduct an empirical study on how a behavioral model created on one device can be transferred to other devices to bootstrap continuous authentication. To pursue this goal, we collect 160 sets of usage data on multiple mobile devices and perform a proof-of-concept experiment. The results demonstrate that we can leverage the similarity between user behaviors on different devices to enable cross-device authentication and anomaly detection.

Recommender systems play a key role in personalizing service experiences by recommending relevant items to users. One popular technique for producing such personalization at scale is collaborative filtering via Matrix Factorization (MF). The essence of MF is to train a model by factorizing a sparse rating matrix consisting of users' ratings of item. Unfortunately, existing MF methods require model Learning from Scratch when new data (for users, items, or user ratings) arrive. Learning large models from scratch incurs significant computation cost and typically also results in stale recommendations. With increasing amounts of data and a need for real-time recommendations, incremental learning is desirable. In this paper, we develop a novel but simple method for incremental learning of MF models, called One-sided Least Squares, and demonstrate its parallel implementation via Apache Spark. We also describe how to integrate it with batch learning via Alternating Least Squares (ALS). Unlike previous incremental learning methods, we study our method's approximation of the results of ALS, while significantly reducing compute and storage costs. Our theoretical analysis and experimental results on three real-world datasets suggest that One-sided Least Squares achieves prediction accuracy close to Learning from Scratch with ALS at substantially faster learning speeds. This fast and accurate method for incremental learning enables improved Web-scale recommender systems.

In Infrastructure-as-a-Service cloud data centers, services are provided to cloud customers in the form of virtual machines. Cloud customers can place restrictions on these services by specifying affinity and anti-affinity constraints. Load imbalance is one key issue that cloud data centers regularly face when running these services. Load imbalance arises when existing services are stopped either by the cloud customers or in the event of host power cycling. One way to achieve load balance in such situations is to perform load rebalancing. Load rebalancing is a process of migrating services among hosts to ensure uniform resource distribution. By doing load rebalancing, SLA violations due to resource shortages on over-utilized hosts can be mitigated. The benefits of load rebalancing come at the expense of migration cost. The presence of affinity and anti-affinity constraints make the load rebalancing challenging. In this paper, we focus on load rebalancing of services with affinity and anti-affinity constraints by applying a novel genetic algorithm. Our objective function aims at reducing the number of migrations and variation of available resources in the hosts. Experimental results show that our algorithm achieves a good resource balance while being computationally efficient.

Many optimization problems are multi-modal. In certain cases, we are interested in finding multiple locally optimal solutions rather than just a single optimum as is computed by traditional genetic algorithms (GAs). Several niching techniques have been developed that seek to find multiple such local optima. These techniques, which include sharing and crowding, are clearly powerful and useful. But they do not explicitly let the user control the number of local optima being computed, which we believe to be an important capability. In this paper, we develop a method that provides, as an input parameter to niching, the desired number of local optima. Our method integrates techniques from feedback control, includes a sensor based on clustering, and utilizes a scaling parameter in Generalized Crowding to control the number of niches being explored. The resulting Feedback Control GA (FCGA) is tested in several experiments and found to perform well compared to previous approaches. Overall, the integration of feedback control and Generalized Crowding is shown to effectively guide the search for multiple local optima in a more controlled fashion. We believe this novel capability has the potential to impact future applications as well as other evolutionary algorithms.

The junction tree approach, with applications in artificial intelligence, computer vision, machine learning, and statistics, is often used for computing posterior distributions in probabilistic graphical models. One of the key challenges associated with junction trees is computational, and several parallel computing technologies - including many-core processors - have been investigated to meet this challenge. Many-core processors (including GPUs) are now programmable, unfortunately their complexities make it hard to manually tune their parameters in order to optimize software performance. In this paper, we investigate a machine learning approach to minimize the execution time of parallel junction tree algorithms implemented on a GPU. By carefully allocating a GPU's threads to different parallel computing opportunities in a junction tree, and treating this thread allocation problem as a machine learning problem, we find in experiments that regression - specifically support vector regression - can substantially outperform manual optimization.

Genetic algorithms typically use crossover, which relies on mating a set of selected parents. As part of crossover, random mating is often carried out. A novel approach to parent mating is presented in this work. Our novel approach can be applied in combination with a traditional similarity-based criterion to measure distance between individuals or with a fitness-based criterion. We introduce a parameter called the mating index that allows different mating strategies to be developed within a uniform framework: an exploitative strategy called best-first, an explorative strategy called best-last, and an adaptive strategy called self-adaptive. Self-adaptive mating is defined in the context of the novel algorithm, and aims to achieve a balance between exploitation and exploration in a domain-independent manner. The present work formally defines the novel mating approach, analyzes its behavior, and conducts an extensive experimental study to quantitatively determine its benefits. In the domain of real function optimization, the experiments show that, as the degree of multimodality of the function at hand grows, increasing the mating index improves performance. In the case of the self-adaptive mating strategy, the experiments give strong results for several case studies.

For hard computational problems, stochastic local search has proven to be a competitive approach to finding optimal or approximately optimal problem solutions. Two key research questions for stochastic local search algorithms are: Which algorithms are effective for initialization? When should the search process be restarted? In the present work, we investigate these research questions in the context of approximate computation of most probable explanations (MPEs) in Bayesian networks (BNs). We introduce a novel approach, based on the Viterbi algorithm, to explanation initialization in BNs. While the Viterbi algorithm works on sequences and trees, our approach works on BNs with arbitrary topologies. We also give a novel formalization of stochastic local search, with focus on initialization and restart, using probability theory and mixture models. Experimentally, we apply our methods to the problem of MPE computation, using a stochastic local search algorithm known as Stochastic Greedy Search. By carefully optimizing both initialization and restart, we reduce the MPE search time for application BNs by several orders of magnitude compared to using uniform at random initialization without restart. On several BNs from applications, the performance of Stochastic Greedy Search is competitive with clique tree clustering, a state-of-the-art exact algorithm used for MPE computation in BNs.

Crowding is a technique used in genetic algorithms to preserve diversity in the population and to prevent premature convergence to local optima. It consists of pairing each offspring with a similar individual in the current population (pairing phase) and deciding which of the two will remain in the population (replacement phase). The present work focuses on the replacement phase of crowding, which usually has been carried out by one of the following three approaches: Deterministic, Probabilistic, and Simulated Annealing. These approaches present some limitations regarding the way replacement is conducted. On the one hand, the first two apply the same selective pressure regardless of the problem being solved or the stage of the genetic algorithm. On the other hand, the third does not apply a uniform selective pressure over all the individuals in the population, which makes the control of selective pressure over the generations somewhat difficult. This work presents a Generalized Crowding approach that allows selective pressure to be controlled in a simple way in the replacement phase of crowding, thus overcoming limitations of the other approaches. Furthermore, the understanding of existing approaches is greatly improved, since both Deterministic and Probabilistic Crowding turn out to be special cases of Generalized Crowding. In addition, the temperature parameter used in Simulated Annealing is replaced by a parameter called scaling factor that controls the selective pressure applied. Theoretical analysis using Markov chains and empirical evaluation using Bayesian networks demonstrate the potential of this novel Generalized Crowding approach.

We present in this paper a case study of the probabilistic approach to model-based diagnosis. Here, the diagnosed system is a real-world electrical power system (EPS), i.e., the Advanced Diagnostic and Prognostic Testbed (ADAPT) located at the NASA Ames Research Center. Our probabilistic approach is formally well founded and based on Bayesian networks (BNs) and arithmetic circuits (ACs). We pay special attention to meeting two of the main challenges often associated with real-world application of model-based diagnosis technologies: model development and real-time reasoning. To address the challenge of model development, we develop a systematic approach to representing EPSs as BNs, supported by an easy-to-use specification language. To address the real-time reasoning challenge, we compile BNs into ACs. AC evaluation (ACE) supports real-time diagnosis by being predictable, fast, and exact. In experiments with the ADAPT BN, which contains 503 discrete nodes and 579 edges and produces accurate results, the time taken to compute the most probable explanation using ACs has a mean of 0.2625 ms and a standard deviation of 0.2028 ms. In comparative experiments, we found that, while the variable elimination and join tree propagation algorithms also perform very well in the ADAPT setting, ACE was an order of magnitude or more faster.

Constraints occur in many application areas of interest to evolutionary computation. The area considered here is Bayesian networks (BNs), which is a probability-based method for representing and reasoning with uncertain knowledge. This work deals with constraints in BNs and investigates how tournament selection can be adapted to better process such constraints in the context of abductive inference. Abductive inference in BNs consists of finding the most probable explanation given some evidence. Since exact abductive inference is NP-hard, several approximate approaches to this inference task have been developed. One of them applies evolutionary techniques in order to find optimal or close-to-optimal explanations. A problem with the traditional evolutionary approach is this: As the number of constraints determined by the zeros in the conditional probability tables grows, performance deteriorates because the number of explanations whose probability is greater than zero decreases. To minimize this problem, this paper presents and analyzes a new evolutionary approach to abductive inference in BNs. By considering abductive inference as a constraint optimization problem, the novel approach improves performance dramatically when a BN's conditional probability tables contain a significant number of zeros. Experimental results are presented comparing the performances of the traditional evolutionary approach and the approach introduced in this work. The results show that the new approach significantly outperforms the traditional one.

A wide range of niching techniques have been investigated in evolutionary and genetic algorithms. In this article, we focus on niching using crowding techniques in the context of what we call local tournament algorithms. In addition to deterministic and probabilistic crowding, the family of local tournament algorithms includes the Metropolis algorithm, simulated annealing, restricted tournament selection, and parallel recombinative simulated annealing. We describe an algorithmic and analytical framework which is applicable to a wide range of crowding algorithms. As an example of utilizing this framework, we present and analyze the probabilistic crowding niching algorithm. Like the closely related deterministic crowding approach, probabilistic crowding is fast, simple, and requires no parameters beyond those of classical genetic algorithms. In probabilistic crowding, subpopulations are maintained reliably, and we show that it is possible to analyze and predict how this maintenance takes place. We also provide novel results for deterministic crowding, show how different crowding replacement rules can be combined in portfolios, and discuss population sizing. Our analysis is backed up by experiments that further increase the understanding of probabilistic crowding.

Weather is a complex, dynamic process with tremendous impact on aviation. While pilots often have access to large amounts of aviation weather data, they find it difficult and time-consuming to identify weather hazards, due to the sheer amount and cryptic formatting of the data. To address this challenge, we have developed information filtering concepts based on a unified Bayesian network model, integrating text and graphical weather data in the context of specific mission, equipment and personal profiles. Based on these concepts, we have implemented three applications, all of which were to existing technology. Using one of the applications, the AWARE Preflight system, pilots found significantly more hazards in about half the time compared to using the current technology

Describes a model-based design method to develop CoRaven, a decision support tool that is intended to assist military intelligence analysts in managing and interpreting large quantities of battlefield information. In this method, we use observations of practitioners solving specific tasks in order to understand and model how they use information. We use this model of the task to help identify user needs that the tool must support, and, during initial prototyping, to guide usability analyses. We have found task models to be an important consideration in the decision support tool design process that can help to constrain the design space and reduce the time required to develop an effective decision support tool prototype.

Intelligence analysis is one of the major functions performed by an Army staff in battlefield management. In particular, intelligence analysts develop intelligence requirements based on the commander's information requirements, develop a collection plan, and then monitor messages from the battlefield with respect to the commander's information requirements. The goal of the CoRAVEN project is to develop an intelligent collaborative multimedia system to support intelligence analysts. Key ingredients of our design approach include: (1) significant knowledge engineering activities with domain experts, (2) representation of an explicit model of reasoning and activity to drive design, (3) the use of Bayesian belief networks as a way to structure inferences that relate observable data to the commander's information requirements, (4) collaborative graphical user interfaces to provide flexible support for the multiple tasks in which analysts are engaged, (5) sonification of data streams and alarms to support enhanced situation awareness, (6) detailed psychological studies of reasoning and judgment under uncertainty, and (7) iterative prototyping of candidate designs with domain experts for both formative and summative evaluation. The paper discusses our current progress on all these fronts.