A model of story generation recently proposed by Riedl and Young casts it as planning, with the additional condition that story characters behave intentionally. This means that characters have perceivable motivation for the actions they take. I show that this condition can be compiled away (in more ways than one) to produce a classical planning problem that can be solved by an off-the-shelf classical planner, more efficiently than by Riedl and Young's specialised planner.

We present a joint message passing approach that combines belief propagation and the mean field approximation. Our analysis is based on the region-based free energy approximation method proposed by Yedidia et al. We show that the message passing fixed-point equations obtained with this combination correspond to stationary points of a constrained region-based free energy approximation. Moreover, we present a convergent implementation of these message passing fixedpoint equations provided that the underlying factor graph fulfills certain technical conditions. In addition, we show how to include hard constraints in the part of the factor graph corresponding to belief propagation. Finally, we demonstrate an application of our method to iterative channel estimation and decoding in an orthogonal frequency division multiplexing (OFDM) system.

Efficient use of multiple batteries is a practical problem with wide and growing application. The problem can be cast as a planning problem under uncertainty. We describe the approach we have adopted to modelling and solving this problem, seen as a Markov Decision Problem, building effective policies for battery switching in the face of stochastic load profiles. Our solution exploits and adapts several existing techniques: planning for deterministic mixed discrete-continuous problems and Monte Carlo sampling for policy learning. The paper describes the development of planning techniques to allow solution of the non-linear continuous dynamic models capturing the battery behaviours. This approach depends on carefully handled discretisation of the temporal dimension. The construction of policies is performed using a classification approach and this idea offers opportunities for wider exploitation in other problems. The approach and its generality are described in the paper. Application of the approach leads to construction of policies that, in simulation, significantly outperform those that are currently in use and the best published solutions to the battery management problem. We achieve solutions that achieve more than 99% efficiency in simulation compared with the theoretical limit and do so with far fewer battery switches than existing policies. Behaviour of physical batteries does not exactly match the simulated models for many reasons, so to confirm that our theoretical results can lead to real measured improvements in performance we also conduct and report experiments using a physical test system. These results demonstrate that we can obtain 5%-15% improvement in lifetimes in the case of a two battery system.

We develop dependent hierarchical normalized random measures and apply them to dynamic topic modeling. The dependency arises via superposition, subsampling and point transition on the underlying Poisson processes of these measures. The measures used include normalised generalised Gamma processes that demonstrate power law properties, unlike Dirichlet processes used previously in dynamic topic modeling. Inference for the model includes adapting a recently developed slice sampler to directly manipulate the underlying Poisson process. Experiments performed on news, blogs, academic and Twitter collections demonstrate the technique gives superior perplexity over a number of previous models.

In this paper we propose a simple yet powerful method for learning representations in supervised learning scenarios where each original input datapoint is described by a set of vectors and their associated outputs may be given by soft labels indicating, for example, class probabilities. We represent an input datapoint as a mixture of probabilities over the corresponding set of feature vectors where each probability indicates how likely each vector is to belong to an unknown prototype pattern. We propose a probabilistic model that parameterizes these prototype patterns in terms of hidden variables and therefore it can be trained with conventional approaches based on likelihood maximization. More importantly, both the model parameters and the prototype patterns can be learned from data in a discriminative way. We show that our model can be seen as a probabilistic generalization of learning vector quantization (LVQ). We apply our method to the problems of shape classification, hyperspectral imaging classification and people's work class categorization, showing the superior performance of our method compared to the standard prototype-based classification approach and other competitive benchmark methods.

We show that the variational representations for f-divergences currently used in the literature can be tightened. This has implications to a number of methods recently proposed based on this representation. As an example application we use our tighter representation to derive a general f-divergence estimator based on two i.i.d. samples and derive the dual program for this estimator that performs well empirically. We also point out a connection between our estimator and MMD.

We consider composite loss functions for multiclass prediction comprising a proper (i.e., Fisher-consistent) loss over probability distributions and an inverse link function. We establish conditions for their (strong) convexity and explore the implications. We also show how the separation of concerns afforded by using this composite representation allows for the design of families of losses with the same Bayes risk.

The paper illustrates a novel approach to conformant planning using classical planners. The approach relies on two core ideas developed to deal with incomplete information in the initial situation: the use of a classical planner to solve non-classical planning problems, and the reduction of the size of the initial belief state. Differently from previous uses of classical planners to solve non-classical planning problems, the approach proposed in this paper creates a valid plan from a possible plan---by inserting actions into the possible plan and maintaining only one level of non-deterministic choice (i.e., the initial plan being modified). The algorithm can be instantiated with different classical planners---the paper presents the GC[LAMA] implementation, whose classical planner is LAMA. We investigate properties of the approach, including conditions for completeness. GC[LAMA] is empirically evaluated against state-of-the-art conformant planners, using benchmarks from the literature. The experimental results show that GC[LAMA] is superior to other planners, in both performance and scalability. GC[LAMA] is the only planner that can solve the largest instances from several domains. The paper investigates the reasons behind the good performance and the challenges encountered in GC[LAMA].

Heuristics based on the delete relaxation are at the forefront of modern domain-independent planning techniques. Here we introduce a principled and flexible technique for augmenting delete-relaxed tasks with a limited amount of delete information, by introducing special fluents that explicitly represent conjunctions of fluents in the original planning task. Differently from previous work in this direction, conditional effects are used to limit the growth of the task to be linear, rather than exponential, in the number of conjunctions that are introduced, making its use for obtaining heuristic functions feasible. We discuss how to obtain an informative set of conjunctions to be represented explicitly, and analyze and extend existing methods for relaxed planning in the presence of conditional effects. The resulting heuristics are empirically evaluated, and shown to be sometimes much more informative than standard delete-relaxation heuristics.

Existing ship scheduling approaches either ignore constraints on ship draft (distance between the waterline and the keel), or model these in very simple ways, such as a constant draft limit that does not change with time. However, in most ports the draft restriction changes over time due to variation in environmental conditions. More accurate consideration of draft constraints would allow more cargo to be scheduled for transport on the same set of ships. We present constraint programming (CP) and mixed integer programming (MIP) models for the problem of scheduling ships at a port with time-varying draft constraints so as to optimise cargo throughput at the port. We also investigate the effect of several variations to the CP model, including a model containing sequence variables, and a model with ordered inputs. Our model allows us to solve realistic instances of the problem to optimality in a very short time, and produces better schedules than both scheduling with constant draft, and manual scheduling approaches used in practice at ports.