IWSR 2015 - International Workshop on Speech Robotics (IWSR)

Speech robotics is a newly feasible and exciting frontier in speech research. Speech robotics generally involves the intersection of three distinct fields:
Mechanical and articulatory speech synthesis
An ultimate goal is to develop mechanical vocal tracts that produce intelligible and appealing speech that may be useful to people. Among various applications these tracts may be used as prosthetic communication tools and for human-robot interaction. A mechanical vocal tract has the potential to synthesize the most natural sounding speech in the history of artificial speech synthesis because such a robotic-controlled mechanical vocal tract will compute the physics for free of the aerodynamics of speech. This general feld of research includes work in phonetics and voice quality, speech and hearing sciences, as well as mechanical engineering and materials science. The ongoing development of digital articulatory synthesizers and of other articulatory-hybrid forms of speech synthesis also falls into this category.
Speech motor learning, representation, planning, and control
Philosophers have been addressing the nature of thought, communication, langauge, and speech motor control since the ancient times. Though human speech is by no means a solved mystery, modern times presents us with many new insights. The analysis-by-synthesis methodology offered by speech robotics promises many new insights still. This general field of research includes contributions from psycholinguistics as well as contributions from computational models, and robotics and neural models for motor control. This general research area also includes insights and theory from developmental psychology (speech acquisition, statistical pattern learning), phonology, and philosophy of mind and semantics.
Human-robot speech communication, interaction, and co-vocalization systems
This research field emphasizes methodologies for the study of speech cognition and dialog. Here, ‘speech cognition’ is a superset of linguistics that includes laughter, grunts, affect, sarcasm, and any other communicative cues that are generally not captured by text or traditional linguistic notations. Furthermore, human-robot speech interaction may involve vision, shared attention, and other co-vocalization systems and methodologies. This general research category also involves contextualized automatic speech recognition, gesture recognition systems, and the development of interactive dialog systems.
The overarching goal of the workshop is to bring together experts from these three general areas to share ideas that may facilitate one-another’s work.