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.

A Framework for the Evaluation of Digital Musical Instruments

At the outset of a discussion of evaluating digital musical instruments, that is to say instruments whose sound generators are digital and separable though not necessarily separate from their control interfaces (Malloch, 2006), it is reasonable to ask what the term evaluation in this context really means. After all, there may be many perspectives from which to view the effectiveness or otherwise of the instruments we build. For most performers, performance on an instrument becomes a means of evaluating how well it functions in the context of live music making, and their measure of success is the response of the audience to their performance. Audiences evaluate performances on the basis of how engaged they feel they have been by what they have seen and heard. When questioned, they are likely to describe good performances as “exciting,” “skillful,” “musical.” Bad performances are “boring,” and those which are marred by technical malfunction are often dismissed out of hand. If performance is considered to be a valid means of evaluating a musical instrument, then it follows that, for the field of DMI design, a much broader definition of the term “evaluation” than that typically used in human-computer interaction (HCI) is required to reflect the fact that there are a number of stakeholders involved in the design and evaluation of DMIs. In addition to players and audiences, there are also composers, instrument builders, component manufacturers, and perhaps even customers, each of whom will have a different concept of what is meant by “evaluation.”

The Belfast Induced Natural Emotion Database

For many years psychological research on facial expression of emotion has relied heavily on a recognition paradigm based on posed static photographs. There is growing evidence that there may be fundamental differences between the expressions depicted in such stimuli and the emotional expressions present in everyday life. Affective computing, with its pragmatic emphasis on realism, needs examples of natural emotion. This paper describes a unique database containing recordings of mild to moderate emotionally coloured responses to a series of laboratory based emotion induction tasks. The recordings are accompanied by information on self-report of emotion and intensity, continuous trace-style ratings of valence and intensity, the sex of the participant, the sex of the experimenter, the active or passive nature of the induction task and it gives researchers the opportunity to compare expressions from people from more than one culture.

Evaluation of Four Designed Virtual Agent Personalities

Convincing conversational agents require a coherent set of behavioral responses that can be interpreted by a human observer as indicative of a personality. This paper discusses the continued development and subsequent evaluation of virtual agents based on sound psychological principles. We use Eysenck's theoretical basis to explain aspects of the characterization of our agents, and we describe an architecture where personality affects the agent's global behavior quality as well as their back-channel productions. Drawing on psychological research, we evaluate perception of our agents' personalities and credibility by human viewers (N = 187). Our results suggest that we succeeded in validating theoretically grounded indicators of personality in our virtual agents, and that it is feasible to place our characters on Eysenck's scales. A key finding is that the presence of behavioral characteristics reinforces the prescribed personality profiles that are already emerging from the still images. Our long-term goal is to enhance agents' ability to sustain realistic interaction with human users, and we discuss how this preliminary work may be further developed to include more systematic variation of Eysenck's personality scales. © 2012 IEEE.


--------------------------------------------------------------------------------

Reaxys Database Information|

--------------------------------------------------------------------------------

Visualising Textual Knowledge about Risks to Aid Risk Communication

This paper demonstrates a potential application for latent semantic analysis and similar techniques in visualising the differences between two levels of knowledge about a risk issue. The HIV/AIDS risk issue will be examined and the semantic clusters of key words in a technical corpora derived from specific literature about HIV/AIDS will be compared with the semantic clusters of those in more general corpora. It is hoped that these comparisons will create a fast and efficient complementary approach to the articulation of mental models of risk issues that could be used to target possible inconsistencies between expert and lay mental models.

Amplitude-modulation characteristics of power-combining Class-E amplifier with finite choke

Power back-off performances of a new variant power-combining Class-E amplifier under different amplitude-modulation schemes such as continuous wave (CW), envelope elimination and restoration (EER), envelope tracking (ET) and outphasing are for the first time investigated in this study. Finite DC-feed inductances rather than massive RF chokes as used in the classic single-ended Class-E power amplifier (PA) resulted from the approximate yet effective frequency-domain circuit analysis provide the wherewithal to increase modulation bandwidth up to 80% higher than the classic single-ended Class-E PA. This increased modulation bandwidth is required for the linearity improvement in the EER/ET transmitters. The modified output load network of the power-combining Class-E amplifier adopting three-harmonic terminations technique relaxes the design specifications for the additional filtering block typically required at the output stage of the transmitter chain. Qualitative agreements between simulation and measurement results for all four schemes were achieved where the ET technique was proven superior to the other schemes. When the PA is used within the ET scheme, an increase of average drain efficiency of as high as 40% with respect to the CW excitation was obtained for a multi-carrier input signal with 12 dB peak-to-average power ratio. © 2011 The Institution of Engineering and Technology.

Robust Multimodal Person Identification With Limited Training Data

This paper presents a novel method of audio-visual feature-level fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there are limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new multimodal feature representation and a modified cosine similarity are introduced to combine and compare bimodal features with limited training data, as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal dataset created from the SPIDRE speaker recognition database and AR face recognition database with variable noise corruption of speech and occlusion in the face images. The system's speaker identification performance on the SPIDRE database, and facial identification performance on the AR database, is comparable with the literature. Combining both modalities using the new method of multimodal fusion leads to significantly improved accuracy over the unimodal systems, even when both modalities have been corrupted. The new method also shows improved identification accuracy compared with the bimodal systems based on multicondition model training or missing-feature decoding alone.

Effect of membrane stiffness and cytoskeletal element density on mechanical stimuli within cells: an analysis of the consequences of ageing in cells

A finite element model of a single cell was created and used to investigate the effects of ageing on biophysical stimuli generated within a cell. Major cellular components were incorporated in the model: the membrane, cytoplasm, nucleus, microtubules, actin filaments, intermediate filaments, nuclear lamina, and chromatin. The model used multiple sets of tensegrity structures. Viscoelastic properties were assigned to the continuum components. To corroborate the model, a simulation of Atomic Force Microscopy (AFM) indentation was performed and results showed a force/indentation simulation with the range of experimental results.

Ageing was simulated by both increasing membrane stiffness (thereby modelling membrane peroxidation with age) and decreasing density of cytoskeletal elements (thereby modelling reduced actin density with age). Comparing normal and aged cells under indentation predicts that aged cells have a lower membrane area subjected to high strain compared to young cells, but the difference, surprisingly, is very small and would not be measurable experimentally. Ageing is predicted to have more significant effect on strain deep in the nucleus. These results show that computation of biophysical stimuli within cells are achievable with single-cell computational models whose force/displacement behaviour is within experimentally observed ranges. the models suggest only small, though possibly physiologically-significant, differences in internal biophysical stimuli between normal and aged cells.