2014 Intelligent Sensing Summer School
Location: FB 1.13, Mile End Campus, Queen Mary University of London
2014 competition: briefing presentation and rules
2013 Intelligent Sensing
videos of the talks
what people said?
|26 August 2014 - DAY 1|
|9.00-9.10||Opening - Andrea Cavallaro|
|9.10-9.30||Introduction: developing an impact pitch - Adam Daykin|
|9.30-10.45||Preparation of a pitch||Tech Transfer training
|10.45-11.00||Final pitch practice|
|11.30-12.00||Groups present pitches||Tech Transfer training
|14.00-15.30||Your Tech Transfer idea||Tech Transfer
|16.00-17.00||Your elevator pitch - Judges:
|27 August 2014 - DAY 2|
|9.00-10.00||Announcement of the pitches selected for the QTech Bootcamp (*) - Andrea Cavallaro||Tech Transfer|
Preparing an investment pitch
|Adam Daykin||Tech Transfer talks|
|10.30-11.00||The optimum business model
Microsoft charge. Google do it for free. Choosing the optimum business model is an essential part of successfully commercialising software. This talk will examine the available options, and offer guidance on selecting the right model by asking where the value is.
Introduction to IP
A brief introduction to intellectual property and patents in particular, focussing on the sensitive issues surrounding the protection of computer implemented inventions. In particular, when is it appropriate to protect inventions implemented in software through the patent system?
|Peter Thorniley||Tech Transfer talks|
External funding sources by Careers & Enterprise
From basic academic research to high-tech start-up
This talk tells the story of the founding of MixGenius, a high-tech start-up company based around music technology research here at Queen Mary University of London. We will describe the journey from academic research to commercialisation, describe how this journey took some unconventional paths, and highlight some of the lessons learned along the way.
|Josh Reiss||Start-up talks
Chair: Peter McOwan
Stuck in the chasm: the less glamorous alternative to failing fast
When you hear about a startup, you are either treated to an inspirational rag-to-riches story or, more likely, to a cautionary tale of earnest efforts and personal savings dissolving into nothingness. There is in fact a third option, severely underrepresented: the company sort-of works, the product sells, you even have an US office, but investors are not queuing up outside your door and internal growth is stagnant: you're stuck in the chasm. I'm not sure how to fix the situation, but I can certainly share the story on how we got there!
|15.00-15.30||Round table with entrepreneurs - Chair: Peter McOwan||Tech Transfer panel discussion|
|16.00-17.00||PhD advice from post-docs and senior PhD students to 1st and 2nd year PhD students
Panel moderators: Xiatian Zhu and Fabio Poiesi
|PhD panel discussion|
|28 August 2014 - DAY 3|
|09.00-09.15||Introduction of the day - Andrea Cavallaro|
From teeth to scrolls to film
What happens when we want to look inside a tooth without cutting it open, or to read a scroll that cannot be unrolled, or to view an old film that's disintegrating and cannot be unreeled? With the help of X-rays, mathematics and a bit of old fashioned luck we tackle these problems on the boundaries of possibility.
|Graham R. Davis||Research talk
Chair: Akram Alomainy
Feature detection and description in non-linear scale spaces
The easily implementable Gaussian scale space does not respect natural boundaries of objects. Using non-linear diffusion filtering reduces noise while retaining object boundaries via locally adaptive blurring.
|Pablo Alcantarilla||Industry talk
(Toshiba Research Cambridge)
Performing graph computations at scale
Graphs are invaluable tools for researchers, easing the modelling of many problems. However, as graphs grow large, there are numerous challenges for efficiently process them. This talk will cover the main characteristics of distributed graph processing systems, and discuss performance factors and research challenges.
|Félix Cuadrado||Research talks
Chair: Akram Alomainy
Transductive multi-view embedding for zero-shot learning
Humans can visually recognise an object which they have never seen before but only read about. Can a computer do the same and recognise unseen categories of objects? In this talk, I will introduce a framework based on embedding multiple semantic representations of objects for zero-shot learning, that is recognition without any labelled training data.
|14.30-16.15||QTech Bootcamp: public pitches and voting - Chair: Adam Daykin, Judges:
|Dragon's Den session|
|16.15-16.30||Winners announced and closing (followed by drinks & nibbles)|
|12.15-14.00||Informal feedback session - Adam Daykin||Tech Transfer|
|14.00-17.00||1-to-1 sessions with Tech Transfer experts||QTech Bootcamp|
|9.00-13.00||Meeting mentors for 1-to-1||QTech Bootcamp|