Building Autonomous Sensitive Artificial Listeners

This paper describes a substantial effort to build a real-time interactive multimodal dialogue system with a focus on emotional and non-verbal interaction capabilities. The work is motivated by the aim to provide technology with competences in perceiving and producing the emotional and non-verbal behaviours required to sustain a conversational dialogue. We present the Sensitive Artificial Listener (SAL) scenario as a setting which seems particularly suited for the study of emotional and non- verbal behaviour, since it requires only very limited verbal understanding on the part of the machine. This scenario allows us to concentrate on non-verbal capabilities without having to address at the same time the challenges of spoken language understanding, task modeling etc. We first report on three prototype versions of the SAL scenario, in which the behaviour of the Sensitive Artificial Listener characters was determined by a human operator. These prototypes served the purpose of verifying the effectiveness of the SAL scenario and allowed us to collect data required for building system components for analysing and synthesising the respective behaviours. We then describe the fully autonomous integrated real-time system we created, which combines incremental analysis of user behaviour, dialogue management, and synthesis of speaker and listener behaviour of a SAL character displayed as a virtual agent. We discuss principles that should underlie the evaluation of SAL-type systems. Since the system is designed for modularity and reuse, and since it is publicly available, the SAL system has potential as a joint research tool in the affective computing research community.

Low frequency localized wavepackets in dusty plasmas with opposite charge polarity dust components

The reductive perturbation technique is employed to investigate the modulational instability of dust-acoustic (DA) waves propagating in a four-component dusty plasma. The dusty plasma consists of both positive- and negative-charge dust grains, characterized by a different mass, temperature and density, in addition to a background of Maxwellian electrons and ions. Relying on a multi-fluid plasma model and employing a multiple scales technique, a nonlinear Schrodinger type equation (NLSE) is obtained for the electric potential amplitude perturbation. The occurrence of localized electrostatic wavepackets is shown, in the form of oscillating structures whose modulated envelope is modelled as a soliton (or multi-soliton) solution of the NLSE. The DA wave characteristics, as well as the associated stability thresholds, are studied analytically and numerically. The relevance of these theoretical results with dusty plasmas observed in cosmic and laboratory environments is analysed in detail, by considering realistic multi-component plasma configurations observed in the polar mesosphere, as well as in laboratory experiments.

Molecular dynamics and principal components analysis of human telomeric quadruplex multimers.

Guanine-rich DNA repeat sequences located at the terminal ends of chromosomal DNA can fold in a sequence-dependent manner into G-quadruplex structures, notably the terminal 150–200 nucleotides at the 3' end, which occur as a single-stranded DNA overhang. The crystal structures of quadruplexes with two and four human telomeric repeats show an all-parallel-stranded topology that is readily capable of forming extended stacks of such quadruplex structures, with external TTA loops positioned to potentially interact with other macromolecules. This study reports on possible arrangements for these quadruplex dimers and tetramers, which can be formed from 8 or 16 telomeric DNA repeats, and on a methodology for modeling their interactions with small molecules. A series of computational methods including molecular dynamics, free energy calculations, and principal components analysis have been used to characterize the properties of these higher-order G-quadruplex dimers and tetramers with parallel-stranded topology. The results confirm the stability of the central G-tetrads, the individual quadruplexes, and the resulting multimers. Principal components analysis has been carried out to highlight the dominant motions in these G-quadruplex dimer and multimer structures. The TTA loop is the most flexible part of the model and the overall multimer quadruplex becoming more stable with the addition of further G-tetrads. The addition of a ligand to the model confirms the hypothesis that flat planar chromophores stabilize G-quadruplex structures by making them less flexible.

Predicting revision risk for aseptic loosening of femoral components in total flip arthroplasty in individual patients - A finite element study

Advances in surgical procedure, prosthesis design, and biomaterials performance have considerably increased the longevity of total joint replacements. Preoperative planning is another step in joint replacement that may have the potential to improve clinical outcome for the individual patient, but has remained relatively consistent for a longtime. One means of advancing this aspect of joint replacement surgery may be to include predictive computer simulation into the planning process. In this article, the potential of patient-specific finite element analysis in preoperative assessment is investigated. Seventeen patient-specific finite element models of cemented Charnley reconstructions were created, of which six were early (

Fast automatic two-stage nonlinear model identification based on the extreme learning machine

It is convenient and effective to solve nonlinear problems with a model that has a linear-in-the-parameters (LITP) structure. However, the nonlinear parameters (e.g. the width of Gaussian function) of each model term needs to be pre-determined either from expert experience or through exhaustive search. An alternative approach is to optimize them by a gradient-based technique (e.g. Newton’s method). Unfortunately, all of these methods still need a lot of computations. Recently, the extreme learning machine (ELM) has shown its advantages in terms of fast learning from data, but the sparsity of the constructed model cannot be guaranteed. This paper proposes a novel algorithm for automatic construction of a nonlinear system model based on the extreme learning machine. This is achieved by effectively integrating the ELM and leave-one-out (LOO) cross validation with our two-stage stepwise construction procedure [1]. The main objective is to improve the compactness and generalization capability of the model constructed by the ELM method. Numerical analysis shows that the proposed algorithm only involves about half of the computation of orthogonal least squares (OLS) based method. Simulation examples are included to confirm the efficacy and superiority of the proposed technique.