Dialog state tracking is difficult because automatic speech recognition (ASR) and spoken language understanding (SLU) errors are common and can cause the system to misunderstand the user. At the same time, state tracking is crucial because the system relies on the estimated dialog state to choose actions -- for example, which restaurants to suggest. Figure 1 shows an illustration of the dialog state tracking task. Historically dialog state tracking has been done with handcrafted rules. More recently, statistical methods have been found to be superior by effectively overcoming some SLU errors, resulting in better dialogs. Despite this progress, direct comparisons between methods have not been possible because past studies use different domains, system components, and evaluation measures, hindering progresss.

This literature review covers AI techniques used for real-time strategy video games, focusing specifically on StarCraft. It finds that the main areas of current academic research are in tactical and strategic decision-making, plan recognition, and learning, and it outlines the research contributions in each of these areas. The paper then contrasts the use of game AI in academia and industry, finding the academic research heavily focused on creating game-winning agents, while the indus- try aims to maximise player enjoyment. It finds the industry adoption of academic research is low because it is either in- applicable or too time-consuming and risky to implement in a new game, which highlights an area for potential investi- gation: bridging the gap between academia and industry. Fi- nally, the areas of spatial reasoning, multi-scale AI, and co- operation are found to require future work, and standardised evaluation methods are proposed to produce comparable re- sults between studies.

As planetary rovers expand their capabilities, traveling longer distances, deploying complex tools, and collecting voluminous scientific data, the requirements for intelligent guidance and control also grow. This, coupled with limited bandwidth and latencies, motivates onboard autonomy that ensures the quality of the science data return. Increasing quality of the data involves better sample selection, data validation, and data reduction. Robotic studies in Mars-like desert terrain have advanced autonomy for long distance exploration and seeded technologies for planetary rover missions. In these field experiments the remote science team uses a novel control strategy that intersperses preplanned activities with autonomous decision making. The robot performs automatic data collection, interpretation, and response at multiple spatial scales. Specific capabilities include instrument calibration, visual targeting of selected features, an onboard database of collected data, and a long range path planner that guides the robot using analysis of current surface and prior satellite data. Field experiments in the Atacama Desert of Chile over the past decade demonstrate these capabilities and illustrate current challenges and future directions.

The challenging timeline for DARPA’s Orbital Express mission demanded a flexible, responsive, and (above all) safe approach to mission planning. Mission planning for space is challenging because of the mixture of goals and constraints. Every space mission tries to squeeze all of the capacity possible out of the spacecraft. For Orbital Express, this means performing as many experiments as possible, while still keeping the spacecraft safe. Keeping the spacecraft safe can be very challenging because we need to maintain the correct thermal environment (or batteries might freeze), we need to avoid pointing cameras and sensitive sensors at the sun, we need to keep the spacecraft batteries charged, and we need to keep the two spacecraft from colliding... made more difficult as only one of the spacecraft had thrusters. Because the mission was a technology demonstration, pertinent planning information was learned during actual mission execution. For example, we didn’t know for certain how long it would take to transfer propellant from one spacecraft to the other, although this was a primary mission goal. The only way to find out was to perform the task and monitor how long it actually took. This information led to amendments to procedures, which led to changes in the mission plan. In general, we used the ASPEN planner scheduler to generate and validate the mission plans. ASPEN is a planning system that allows us to enter all of the spacecraft constraints, the resources, the communications windows, and our objectives. ASPEN then could automatically plan our day. We enhanced ASPEN to enable it to reason about uncertainty. We also developed a model generator that would read the text of a procedure and translate it into an ASPEN model. Note that a model is the input to ASPEN that describes constraints, resources, and activities. These technologies had a significant impact on the success of the Orbital Express mission. Finally, we formulated a technique for converting procedural information to declarative information by transforming procedures into models of hierarchical task networks (HTNs). The impact of this effort on the mission was a significant reduction in (1) the execution time of the mission, (2) the daily staff required to produce plans, and (3) planning errors. Not a single miss-configured command was sent during operations.

This article describes the DSN scheduling wngine (DSE) component of a new scheduling system being deployed for NASA's deep space network. The DSE provides core automation functionality for scheduling the network, including the interpretation of scheduling requirements expressed by users, their elaboration into tracking passes, and the resolution of conflicts and constraint violations. The DSE incorporates both systematic search and repair-based algorithms, used for different phases and purposes in the overall system. It has been integrated with a web application which provides DSE functionality to all DSN users through a standard web browser, as part of a peer-to-peer schedule negotiation process for the entire network. The system has been deployed operationally and is in routine use, and is in the process of being extended to support long-range planning and forecasting, and near-real-time scheduling.

We are pleased to introduce the space application issue articles in this issue of AI Magazine. The exploration of space is a testament to human curiosity and the desire to understand the universe that we inhabit. As many space agencies around the world design and deploy missions, it is apparent that there is a need for intelligent, exploring systems that can make decisions on their own in remote, potentially hostile environments. At the same time, the monetary cost of operating missions, combined with the growing complexity of the instruments and vehicles being deployed, make it apparent that substantial improvements can be made by the judicious use of automation in mission operations.

Representing the information from the recent past as transient activity distributed over a network has been actively researched in biophysical as well as purely computational domains [1, 2]. It is understood that recurrent connections in the network can keep the information from distant past alive so that it can be recovered from the current state. The memory capacity of these networks are generally measured in terms of the accuracy of recovery of the past information [2-4]. Although the memory capacity strongly depends on the network's topology and sparsity [5-8], it can be significantly increased by exploiting any prior knowledge of the underlying structure of the encoded signal [9, 10]. Our approach to encoding memory stems from a focus on its utility for future prediction, rather than on the accuracy of recovering the past. In particular we are interested in encoding time varying signals from the natural world into memory so as to optimize future prediction. It is well known that most natural signals exhibit scale free long range correlations [11-13]. By exploiting this intrinsic structure underlying natural signals, prior work has shown that the predictive information contained in a finite sized memory system can be maximized if the past is encoded in a scale-invariantly coarse grained fashion [14]. Each node in such a memory system would represent a coarse grained average around a specific past moment, and the time window of coarse graining linearly scales with the past timescale.

We propose a high dimensional classification method that involves nonparametric feature augmentation. Knowing that marginal density ratios are the most powerful univariate classifiers, we use the ratio estimates to transform the original feature measurements. Subsequently, penalized logistic regression is invoked, taking as input the newly transformed or augmented features. This procedure trains models equipped with local complexity and global simplicity, thereby avoiding the curse of dimensionality while creating a flexible nonlinear decision boundary. The resulting method is called Feature Augmentation via Nonparametrics and Selection (FANS). We motivate FANS by generalizing the Naive Bayes model, writing the log ratio of joint densities as a linear combination of those of marginal densities. It is related to generalized additive models, but has better interpretability and computability. Risk bounds are developed for FANS. In numerical analysis, FANS is compared with competing methods, so as to provide a guideline on its best application domain. Real data analysis demonstrates that FANS performs very competitively on benchmark email spam and gene expression data sets. Moreover, FANS is implemented by an extremely fast algorithm through parallel computing.

We propose to learn a kernel-based message operator which takes as input all expectation propagation (EP) incoming messages to a factor node and produces an outgoing message. In ordinary EP, computing an outgoing message involves estimating a multivariate integral which may not have an analytic expression. Learning such an operator allows one to bypass the expensive computation of the integral during inference by directly mapping all incoming messages into an outgoing message. The operator can be learned from training data (examples of input and output messages) which allows automated inference to be made on any kind of factor that can be sampled.

We consider distributed convex optimization problems originated from sample average approximation of stochastic optimization, or empirical risk minimization in machine learning. We assume that each machine in the distributed computing system has access to a local empirical loss function, constructed with i.i.d. data sampled from a common distribution. We propose a communication-efficient distributed algorithm to minimize the overall empirical loss, which is the average of the local empirical losses. The algorithm is based on an inexact damped Newton method, where the inexact Newton steps are computed by a distributed preconditioned conjugate gradient method. We analyze its iteration complexity and communication efficiency for minimizing self-concordant empirical loss functions, and discuss the results for distributed ridge regression, logistic regression and binary classification with a smoothed hinge loss. In a standard setting for supervised learning, the required number of communication rounds of the algorithm does not increase with the sample size, and only grows slowly with the number of machines.