Probabilistic feature models for unfolding

Proceedings of the International Workshop on Emotion Representations and Modelling for Companion Technologies, 2015

The manifestation of human emotions evolves over time and space. Most of the work on affective computing research is limited to the association of context-free signal segments, such as utterances and images, to basic emotions. In this paper, we discuss the hypothesis that interpreting emotions requires a conceptual description of their dynamics within the context of their manifestations. We describe the unfolding of emotions through the proposed affective scene framework. Affective scenes are defined in terms of who first expresses the variation in their emotional state in a conversation, how this affects the other speaker's emotional appraisal and response, and which modifications occur from the initial through the final state of the scene. This conceptual framework is applied and evaluated on real human-human conversations drawn from call centers. We show that the automatic classification of affective scenes achieves more than satisfactory results and it benefits from acoustic, lexical and psycholinguistic features of the speech and linguistics signals.

The generalized graded unfolding model (J. Roberts, J. Donoghue, and J. Laughlin, 1998, 1999) is an item response theory model designed to unfold polytomous responses. The model is based on a proximity relation that postulates higher levels of expected agreement with a given statement to the extent that a respondent is located close to the statement on a unidimensional latent continuum. J. Roberts and others (1998) have examined the recovery of item and person parameters from the generalized graded unfolding model. Item and person parameters were estimated. This study used simulation methods to assess the sensitivity of model parameters to the prior distribution used in these estimation procedures. It also examined the effects of the number of quadrature points used in the numerical integration process and the calibration sample on the accuracy of the resulting parameter estimates. The results show that item parameter estimated derived from the marginal maximum likelihood procedure ...

Psychological Methods, 1999

A new model of change is proposed, based on the assumption that cognitive and behavioral processes of change basically follow inverse-U-shaped patterns of variation as smokers move toward effective change: Each process is first increasingly used, up to a maximum value, and then decreases. It is argued that such a model of data is properly dealt with by unfolding models specially designed for those cases. A theoretical foundation for an unfolding model of change is proposed, based on probabilistic reasoning first developed by D. Andrich and G. Luo (1993). An illustrative analysis on responses of 140 French smokers to C. C. DiClemente and J. O. Processes of Change Questionnaire is presented, which yields a very satisfactory unidimensional solution, along which items' locations are in convergence with previous longitudinal studies in the stage-of-change tradition and smokers' locations appear to be a good predictor of actual quitting.

status: published

Contains three published papers and two unpublished manuscripts. The latter are 1) a historic overview of the literature on degeneracies in multidimensional unfolding and 2) a manuscript presenting a combined vector-ideal point model and accompanying algorithm.

Journal of classification, 2024

Synthese, 2012

Figures, unless inserted in the text, are attached as separate files or copied at the end of the document.

Psychometrika, 2005

In this paper, we reconsider the merits of unfolding solutions based on loss functions involving a normalization on the variance per subject. In the literature, solutions based on Stress-2 are often diagnosed to be degenerate in the majority of cases. Here, the focus lies on two frequently occurring types of degeneracies. The first type typically locates some subject points far away from a compact cluster of the other points. In the second type of solution, the object points lie on a circle. In this paper, we argue that these degenerate solutions are well fitting and informative. To reveal the information, we introduce mixtures of plots based on the ideal point model of unfolding, the vector model, and on the signed distance model. In addition to a different representation, we provide a new iterative majorization algorithm to optimize the average squared correlation between the distances in the configuration and the transformed data per individual. It is shown that this approach is equivalent to minimizing Kruskal's Stress-2.

The central theme of BASC 2008 emphasizes the relationships between architecture, sustainability, and human community, and, therefore, by implication, the requirement to satisfy all design constraints. With this focus in mind, let’s dismiss lopsided approaches that do not take us toward holistic progress. First, eliminate ‘development’ as characterized by image-driven, modular, large-scale construction through centralized and monolithic power structures. Second, forget the current ‘star’ culture of professional architects with its requirement for originality at all costs. Third, ignore nostalgic attempts to return to the security of old forms that have lost the cultural and ethical principles that previously infused them with life. Fourth, set aside any single-focus on sustainability which is attainable only at the detriment of other concerns. On offer is a new―untested but promising―approach sketched out in the recent work on unfolding built environments. The inspiration for this new way of thinking is strongly biological, stemming notably from current research in morphogenesis and autopoiesis. Unfolding is sequenced growth. Germane to its dynamic are (a) the co-evolution of part and whole, (b) maintenance of the whole through system repair, (c) the power of fine grain local adaptation, and (d) adaptable, robust, economical forms that can accommodate contradictory demands. Not all, but some building traditions come close to unfolding and are, therefore, worthy of study. The old architecture of the Arab World not only solved the climatic problems and fundamental issues of sustainability through unfolding but did so with a combination of beauty and physical and social functionality. Green builders in the American Southwest, a bio-region similar to the Maghreb, are currently studying these vernacular forms for possible adaptation to the American milieu.

1999

The generalized graded unfolding model (GGUM) (J. Roberts, J. Donoghue, and J. Laughlin, 1998) is an item response theory model designed to analyze binary or graded responses that are based on a proximity relation. The purpose of this study was to assess conditions under which item parameter estimation accuracy increases or decreases, with special attention paid to the influence that a given item parameter value has on the estimation of another item parameter. This assessment was based on a recovery simulation in which the effects of sample size, item location, degree of item discrimination, and extremity of subjective category thresholds were varied. Results indicate that with 750 or more respondents, sample size has negligible effect on all but the estimation of subjective response category thresholds. The true extremity of both item location and item discrimination did affect the estimation of these parameters themselves, and also affected the estimation of other item parameters in the model. However, these effects were modest and had little impact on the estimation of the corresponding item response functions. These results suggest that marginal maximum likelihood estimates of item parameters will provide accurate results across a variety of item parameter configurations when the sample size is at or above the recommended levels. (Contains 1 table, 3 figures, and 14 references.) (Author/SLD)

2005

People express their mental states all the time, even when interacting with machines. These mental states shape the decisions that we make, govern how we communicate with others, and affect our performance. The ability to attribute mental states to others from their behaviour, and to use that knowledge to guide one's own actions and predict those of others is known as theory of mind or mind-reading. The principal contribution of this dissertation is the real time inference of a wide range of mental states from head and facial displays in a video stream. In particular, the focus is on the inference of complex mental states: the affective and cognitive states of mind that are not part of the set of basic emotions. The automated mental state inference system is inspired by and draws on the fundamental role of mind-reading in communication and decision-making. The dissertation describes the design, implementation and validation of a computational model of mind-reading. The design is based on the results of a number of experiments that I have undertaken to analyse the facial signals and dynamics of complex mental states. The resulting model is a multi-level probabilistic graphical model that represents the facial events in a raw video stream at different levels of spatial and temporal abstraction. Dynamic Bayesian Networks model observable head and facial displays, and corresponding hidden mental states over time. The automated mind-reading system implements the model by combining top-down predictions of mental state models with bottom-up vision-based processing of the face. To support intelligent human-computer interaction, the system meets three important criteria. These are: full automation so that no manual preprocessing or segmentation is required, real time execution, and the categorization of mental states early enough after their onset to ensure that the resulting knowledge is current and useful. The system is evaluated in terms of recognition accuracy, generalization and real time performance for six broad classes of complex mental states-agreeing, concentrating, disagreeing, interested, thinking and unsure, on two different corpora. The system successfully classifies and generalizes to new examples of these classes with an accuracy and speed that are comparable to that of human recognition. The research I present here significantly advances the nascent ability of machines to infer cognitive-affective mental states in real time from nonverbal expressions of people. By developing a real time system for the inference of a wide range of mental states beyond the basic emotions, I have widened the scope of human-computer interaction scenarios in which this technology can be integrated. This is an important step towards building socially and emotionally intelligent machines. This dissertation was inspired and shaped through discussions with many people and was only made possible by the support of many more who have shared their love, prayers, time, and know-how with me. Without the support of each and every one of them it would not have been possible to undertake and successfully complete this challenging endeavour. I would like to especially thank my supervisor, Peter Robinson, for giving me the opportunity to be part of this inspirational environment and for being supportive in every imaginable way. I would also like to thank Alan Blackwell for his valuable advice and for involving me with Crucible through which I have learnt tremendously, Neil Dodgson for his technical and moral support and Rosalind Picard and Sean Holden for their insightful comments on the dissertation. Simon Baron-Cohen has and continues to provide the inspiration for many of the ideas in this dissertation. I would like to thank him and his group at the Autism Research Centre, especially Ofer Golan for sharing their thoughts, research and the Mind Reading DVD with us. Tony Manstead, Alf Linney, Alexa Wright and Daren Mc-Donald for fruitful discussions on emotions and facial expressions. Special thanks are due to everyone who has volunteered in taking part in the various studies and experiments which I have undertaken, and especially those who have volunteered with their acting at the CVPR 2004 conference.