Multimodal Interaction for Presentation and Knowledge Capture regarding 3D Design
Heriot-Watt University invites applications for one fully funded three-year PhD studentship from EU and UK students.
Key phrases: multimodal interaction, iPads, virtual reality, gestures, machine learning, search, topic modelling, 3D modelling, 3D visualisation.
Date of issue: 13/2/2012
Project Description
Goals: The goals of this project are to enable designers to (i) use multimodal methods to quickly search, locate and replay relevant sections of design meeting interactions, and (ii) to analyse these multimodal data to automatically encode and formalise the rationale behind design decisions.
Motivation: UK companies lose millions every year because designers do not have time to record the reasoning behind crucial decisions made in design reviews. Consequently they lose time repeating blind alleys, performing rework, repeating previous mistakes or simply discussing previous design decisions .
Approach: With recent advances in pervasive and virtual reality computing (iPads, data pens, virtual reality suits, data gloves and haptics) it now possible to record gigabytes of data about all aspects of a meeting: including how individuals manipulate 3D designs, where they look on the screen, what they write, draw and say, how they feel and what gestures they use. However, such data is useless unless efficient and intuitive ways of storing, browsing and searching these data are provided.
This project would use advanced machine learning together with speech, gesture and eye-movement analysis on time-synchronised multimodal data in order to extract information concerning the rationale and argument behind design decisions.
It would aim to provide two levels of support to designers: (i) it would help them to quickly the intuitively locate appropriate streaming data that would remind them of the detail of previous discussions, and (ii) it would provide automatic encoding of, and production of formal design documentation on, the reasoning behind major design decisions.
Qualifications: first or upper second class honors degree (or 2II + good MSc) in computing or related subject
Eligibility: EPSRC rules apply (see details here). It means this post is restricted to individuals who basically:
- have ‘Settled status in the UK’ (meaning they have no restrictions on how long they can stay) and
- have been ‘normally residing in the UK’ for the last 3 years.
Fees: PhD registration fees will be paid in full
Stipend: £13,590 per year tax free
Latest Start date: 1 October 2012
Duration: 3 – 3.5 years .
Location: Heriot-Watt Edinburgh Campus
Closing date: 2 March 2012
Application: A CV and short covering letter explaining why this post is of interest should sent by email with the subject line of ‘Knowledge Capture PhD vacancy” to:
| Prof. Mike Chantler School of Mathematical & Computer Sciences Heriot-Watt University Riccarton Edinburgh Scotland EH14 4AS M.J.Chantler@hw.ac.uk |
Links: Texture lab: http://www.macs.hw.ac.uk/texturelab/