Probabilistic inference is the problem of estimating the hidden variables (states or parameters) of a system in an optimal and consistent fashion as a set of noisy or incomplete observations of the system becomes available online. The optimal solution to this problem is given by the recursive Bayesian estimation algorithm which recursively updates the posterior density of the system state as new observations arrive. This posterior density constitutes the complete solution to the probabilistic inference problem, and allows us to calculate any “optimal” estimate of the state. Unfortunately, for most real-world problems, the optimal Bayesian recursion is intractable and approximate solutions must be used. Within the space of approximate solutions, the extended Kalman filter (EKF) has become one of the most widely used algorithms with applications in state, parameter and dual estimation. Unfortunately, the EKF is based on a sub-optimal implementation of the recursive Bayesian estimation framework applied to Gaussian random variables. This can seriously affect the accuracy or even lead to divergence of any inference system that is based on the EKF or that uses the EKF as a component part. Recently a number of related novel, more accurate and theoretically better motivated algorithmic alternatives to the EKF have surfaced in the literature, with specific application to state estimation for automatic control. We have extended these algorithms, all based on derivativeless deterministic sampling based approximations of the relevant Gaussian statistics, to a family of algorithms called Sigma-Point Kalman Filters (SPKF). Furthermore, we successfully expanded the use of this group of algorithms (SPKFs) within the general field of probabilistic inference and machine learning, both as stand-alone filters and as subcomponents of more powerful sequential Monte Carlo methods (particle filters). We have consistently shown that there are large performance benefits to be gained by applying Sigma-Point Kalman filters to areas where EKFs have been used as the de facto standard in the past, as well as in new areas where the use of the EKF is impossible.

/pdf/sigma-point-kalman-filters-for-probabilistic-inference-in-1affrc62j6.pdf

Sigma-point kalman filters for probabilistic inference in dynamic state-space models

https://superfund.berkeley.edu/pdf/92.pdf

Epidemiology of Myelodysplastic Syndromes and Chronic Myeloproliferative Disorders in the United States, 2001-2004, Using Data From the NAACCR and SEER Programs

Diagnosis and treatment of primary myelodysplastic syndromes in adults: recommendations from the European LeukemiaNet

Monads have become very popular for structuring functional programs since Wadler introduced their use in 1990. In particular, libraries of combinators are often based on a monadic type. Such libraries share (in part) a common interface, from which numerous benefits flow, such as the possibility to write generic code which works together with any library. But, several interesting and useful libraries are fundamentally incompatible with the monadic interface. In this paper I propose a generalisation of monads, which I call arrows, with significantly wider applicability. The paper shows how many of the techniques of monadic programming generalise to the new setting, and gives examples to show that the greater generality is useful. In particular, three non-monadic libraries for efficient parsing, building graphical user interfaces, and programming active web pages fit naturally into the new framework.

Generalising monads to arrows

A probabilistic framework, called Sigma-point Kalman Filters (SPKF) was applied to the problem domain addressed by the extended Kalman Filter (EKF). SPKF methods are superior to the standard EKF based estimation approaches, as an SPKF achieves second-order or higher accuracy. The SPKF has also been applied to the integrated navigation problem as it relates to unmanned aerial vehicle (UAV) autonomy. The SPKF-based sensor latency compensation technique is used to demonstrate the lagged GPS measurements.

/pdf/sigma-point-kalman-filters-for-nonlinear-estimation-and-3h5o0hche0.pdf

Sigma-Point Kalman Filters for Nonlinear Estimation and Sensor Fusion: Applications to Integrated Navigation

This paper describes an implementation of a small windowbased graphical user interface toolkit for X Windows written in the lazy functional language LML. By using this toolkit, a H askell or LML programmer can crest e a user interface with menus, buttons and other graphical interface objects, without conforming to more or less imperative programming paradigms imposed if she were to use a traditional (imperative) toolkit. Instead, the power of the abstraction methods provided by Haskell or LML are used. The main abstraction we use is the fr.dget. Fudgets are combined in a hierarchical structure, and they interact by message passing, The current implementation is baaed on a sequential evaluator, but by using non-determinism and oracles, we suggest how the fudgets can evaluate in parallel. We believe that the toolkit can be extended to a full-feathered and practically useful high level graphical toolkit.

FUDGETS: a graphical user interface in a lazy functional language

The striking ability of territory owners to repel intruders has generated a number of theoretical explanations as well as experimental studies in many animal species. However, effects of individual habitat preferences on territorial defence have rarely been studied. From the territory value hypothesis, we predicted that owners of preferred habitats should invest more resources in defence than owners of non-preferred habitats. We tested this prediction with young territorial brown trout in a two-stage experiment. First, trout were allowed to choose individually between gravel and a uniform bright substrate. As expected, they showed a significant (79%) preference for gravel. However, there was considerable variation between individuals in substrate preference, with a few fish preferring the bright substrate. Half of the tested fish were then transferred to a gravel substrate and the rest to a bright substrate, manipulating habitat type in relation to preference. Territory owners were then staged against size-matched intruders whereupon contest aggression was observed and the winner of each contest determined. Overall, owners won most of the contests. Satisfied owners won 86% and owners of less preferred territories, 74% of the contests. Furthermore, more satisfied owners attacked sooner and were more aggressive relative to the intruders. We conclude that brown trout show individual variation in habitat preference, which appears to be linked with their investment in territorial defence. These results suggest that understanding and modelling of animal contests could benefit from considering how territorial defence is influenced by individual habitat preference and specialisation.

Habitat preference increases territorial defence in brown trout (Salmo trutta)

The myelodysplastic syndromes (MDS) have attracted great interest during the last decade. Unfortunately there is not much material published concerning the incidence of these conditions. We present epidemiologic data on MDS based on case registration of patients in a well-defined population and as a comparison similar data on acute myeloid leukaemia (AML). Between the years 1978–1992 we registered 120 cases of MDS and 146 cases of AML. The median age for all the MDS cases was 74.1 years for men and 78.2 years for women. Among haematologists there is a suspicion that the incidence of MDS is rising. Our study does not support this opinion. We have divided the study period into 5-year periods and the crude incidence has been 3.2, 4.1 and 3.5/100 000/year for each period. In the age group over 70 years MDS was more frequent than AML and in the last 5-year period the incidence was 15.0/100 000/year for MDS compared to 10.2/100 000/year for AML. In conclusion MDS is quite common among elderly people and there is no evidence for a rising incidence during the last 15 years.

Incidence of myelodysplastic syndromes in a Swedish population

This paper presents a monadic approach to incremental computation, suitable for purely functional languages such as Haskell. A program that uses incremental computation is able to perform an incremental amount of computation to accommodate for changes in input data. Recently, Acar, Blelloch and Harper presented a small Standard ML library that supports efficient, high-level incremental computations [1]. Here, we present a monadic variant of that library, written in Haskell extended with first-class references. By using monads, not only are we able to provide a purely functional interface to the library, the types also enforce "correct usage" without having to resort to any type-system extension. We also find optimization opportunities based on standard monadic combinators.This is an exercise in putting to work monad transformers with environments, references, and continuations.

Monads for incremental computing

This paper is an initial report on flight experiments with a small, unmanned helicopter using a state dependent Riccati Equation (SDRE) controller for autonomous, agile maneuvering. The control design is based upon a full, 6-DoF, analytic nonlinear dynamic model, which is manipulated into a pseudo-linear form in which system matrices are given explicitly as a function of the current state. A standard Riccati equation is then solved numerically in each frame of a 50 Hz. control loop to design the state feedback control law on-line. Several flights have been flown with the helicopter to evaluate the accuracy of tracking under SDRE control in comparison with simulation results.

Magnus Carlsson

Papers

FUDGETS: a graphical user interface in a lazy functional language

Habitat preference increases territorial defence in brown trout (Salmo trutta)

Incidence of myelodysplastic syndromes in a Swedish population

Monads for incremental computing

State-Dependent Riccati Equation Control of a Small Unmanned Helicopter