Informatics Newsletter August 2018

Issue 17 of our School newsletter for students and staff.

Announcement

A thank you to Johanna Moore

As many of you will know, Prof Johanna Moore has recently stepped down in her capacity as Head of School after 4 years in the post, and this month we welcomed Prof Jane Hillston as the new Head of the School!

We would like to thank Johanna, for all that she has accomplished during her time as the Head, and express how her work and dedication have been appreciated by us all in Informatics.

During her term there have been many changes and successes within informatics, including the renewal of the Athena Silver Award in 2017 (following the first and successful submission led by Prof Jane Hillston), an award that encourages and recognises commitment to advancing the careers of women in science and technology in higher education and research. As a School, we were very proud of securing this award for the second time in a row, proving our commitment in developing opportunities for our students and working towards a more equal workplace for our academics and professional services.

Over the past 5 years, the School also celebrated successful results in the latest REF assessment, placing informatics amongst the best in the UK for the breadth and quality of its research. The launch of the Blockchain Technology Laboratory, the first days of Edinburgh Centre for Robotics and getting the School’s three CDTs off to a good start, are just a few initiatives which Johanna was involved in, while leading the School during a time of significant growth.

This is not farewell – everyone will still be able to see Johanna in the Forum and possibly share some dance moves with her at the Jamboree.

We would also like to formally welcome Jane to her new role in the School and look forward to discussing her plans and ambitions for the School of Informatics in the forthcoming years.

I truly enjoyed my time as Head of School. It was an honour and a pleasure to represent such an exciting and talented group of people. I sincerely wish Jane all the best, and I am sure she will lead Informatics to an even brighter future.

Professor Johanna Moore

New Staff

Academic

Tariq Elahi - Lecturer in Security and the Internet of Things – started on 22^nd August 2018
Heather Yorston - University Teacher – started on 1^st August 2018

Professional Services

Craig Skeldon - Business Development Manager – started on 18^th August 2018
Agapi Stylianidou - IGS Administrative Assistant – started on 16^th August 2018

Research

Carol Tiseo - Research Associate – started on 27^th August 2018
Sean Tull - Research Associate– started on 7^th August 2018
Abu Farha Ibrahim Jamal Ibrahim - Research Assistant– started on 11^th August 2018
James Vaughan - Research Assistant– started on 20^th August 2018

Promotions

Congratulations to the following staff who were promoted on 1st August:

Murray Cole - Personal Chair of Patterned Parallel Computing
Sharon Goldwater - Personal Chair of Computational Language Learning
Amos Storkey - Personal Chair of Machine Learning
Myrto Arapinis - Reader
Boris Grot - Reader
Sam Lindley - Senior Research Fellow
Wei Chen - Senior Researcher
Petros Papapanagiotou - Senior Researcher
Pavlos Petoumenos - Senior Researcher
Cassia Valentini-Botinhao - Senior Researcher
Oliver Watts - Senior Researcher

Petros Papapanagiotou has been appointed to a Chancellor’s Fellowship from January 2019
Pavlos Petoumenos has been awarded a Royal Academy of Engineering Fellowship to be held in the School

Senior Academic Posts

Jane Hillston – Head of School
Stuart Anderson – Deputy Head of School, Director of Teaching
Nigel Topham – Deputy Head of School, Director of Graduate School
Alex Lascarides – Deputy Director of Graduate School
Chris Williams – Director of Research
Ian Simpson – Deputy Director of Research
Bob Fisher – Director of Equality and Diversity
Dorota Glowacka – Deputy Director of Equality and Diversity

Research successes

Dr Paul Patras has been elevated to the grade of Senior Member of Institute of Electrical and Electronics Engineers (IEEE). Senior Members is the highest professional grade of IEEE for which a member may apply. It requires extensive experience and reflects professional accomplishment and maturity. Only 10% out of more than 400.000 members have achieved this level. IEEE is the world’s largest technical professional organization dedicated to advancing technology for the benefit of humanity. Its roots go back to 1884 when electricity began to become a major influence in society. In the spring that year, a group of individuals in the electrical professions formed a new organization to support professionals in their field — the American Institute of Electrical Engineers (AIEE). In 1912a new organisation the Institute of Radio Engineers was established, modelled on the AIEE but was devoted to radio, and then broadly to electronics. On 1 January 1963, the AIEE and the IRE merged to form the Institute of Electrical and Electronics Engineers, or IEEE.

The UDRC Phase 3 “Signal Processing in the Information Age” project is funding a consortium of Universities to develop underpinning signal processing and machine learning methodologies that are likely to be valuable in defence applications. Partners include Edinburgh University, Heriot-Watt University, University of Strathclyde and Queen’s University Belfast. Edinburgh Informatics with Tim Hospedales as a PI is leading the sub-project on Verifiable Deep Learning. Deep Neural Networks (DNNs) are deployed in increasingly many and increasingly mission-critical applications due to their impressive performance in a variety of areas. It is thus increasingly important to be confident that their outputs can be relied upon for decision-making. However as relatively “black box” systems, they have so far been hard to analyse in order to provide any kind of guarantee that they will always function as as expected, particularly when exposed to new, unexpected, and potentially adversarial situations and inputs. In this project we will develop methods to certify and verify that DNNs are fit for purpose, even when extrapolating in the presence of novel inputs. This will involve developing methods that can provide limited mathematical guarantees on DNNs' behaviour, despite their black box nature; and also methods for explaining their reasoning process so that their decisions can be manually checked for validity. This research should allow us to benefit from the applications being enabled by DNN's high performance, while relaxing in the confidence that they are doing the “right thing”.

Professor Wenfei Fan received the Royal Society Wolfson Merit Award for his Querying Big Data with Limited Resources project. Big data is the next frontier for innovation, competition and productivity. However, big data analytics is often prohibitively costly. It can take days to combine information from tables with 10 million tuples. In other words, computation problems that are traditionally considered "tractable" in the classical theory of computation may nevertheless be impossible in practice. Wenfei’s project will develop a new query evaluation paradigm that starts from the observation that with right auxiliary structures, many queries can be answered from a small subset of the data -- one whose size can be bounded in advance of evaluation. Researchers can hence reduce such ``boundedly evaluable'' queries on big data to computations on small data. For queries that are not bounded, they will introduce a scheme to compute approximate answers with accuracy guarantees, again by accessing a bounded amount of data. The characterisations of boundedly evaluable queries and the new approximation scheme extend the theory of scalable querying of big data that was proposed earlier. Big data analysis is currently the province of organisations that have massive computational resources at their disposal. This project has a potential to make it feasible for small businesses and researchers.

Pavlos Petoumenos secured funding for his project "Deep Learning For Easier Compiler Analysis and Optimisation". The project will use deep learning to understand and analyse computer code to enable novel optimisations for high performing software. Computing platforms are becoming more powerful but also more complex and harder to program. Producing code that uses the hardware efficiently requires time and expertise. Most developers lack either or both of them, so applications are inefficient, wasting processing power and electricity. Through deep neural networks, this project will remove human expertise from this process, making it easy, fast, and low cost to generate programs adapted to their hardware.

Lexi Birch secured funding for her project MTStretch: Low-resource Machine Translation. Neural machine translation (NMT) has recently made major advances in translation quality, however high performing neural models require many millions of human translated sentences for training. For many real-world applications, there is not enough data to build useful MT systems. In this project researchers plan to stretch the resources and capabilities that they have, in order to develop robust MT technologies which can be deployed for low-resource language pairs and for highly specialised low-resource domains. They will investigate making translation significantly more robust by using the intuition that translated (or parallel) corpora contain enormous redundancies, and are an inefficient way to learn to translate. Inspired by human learning, they will study Bayesian models which build up meaning compositionally and are able to learn to learn, thus creating models which only need a few training examples. They will also develop machine learning techniques, such as transfer learning and data augmentation, to extract knowledge from monolingual and parallel resources from other languages and domains. These advances will allow their partners, BBC World Service and BBC Monitoring, to cover under-resourced languages.

Hakan Bilen secured funding for his project Joint Inference Multi-task Networks. When we look at an image that contains a number of desks, chairs and a whiteboard, we can effortlessly predict that it is a classroom. Surprisingly even this simple problem requires us to coherently solve multiple tasks such as categorising the objects (e.g. chair, desk, whiteboard), estimating the depth and geometry of the room, spatial relations between the objects. During this process, our perception benefits from interactions among these tasks such that solving one would render solving another one easier (e.g. knowing the floor plane can help to find chairs, knowing the chair locations can conversely help to detect the floor). Similar to us machines would also divide the same problem into smaller bits however solve each subproblem disjointly and ignore their interactions. This paradigm prevents machines to develop a holistic understanding of visual world and thus to produce accurate predictions. This project will address such limitations by introducing collaborative problem solving strategies for machine perception and boost the performance of machine vision.

News from around the School

Other news

Strategy Committee

Minutes from August 22nd 2018

Bayes Centre

At the end of August, we've finally taken occupancy of the new Bayes Centre building. After EPCC, DataLab, Mathematics/ICMS, and Informatics staff and students move in, we will be welcoming from Design Informatics and the Alan Turing Institute in late September, as well as several external organisations including the startups participating in the Wayra accelerator. Bayes is one of the five hubs the University is establishing to deliver the objectives of the £300-million Data-Driven Innovation Programme funded by the Edinburgh and South East Scotland City and Region Deal, and it is hosted by the College of Science and Engineering. Much of its activity will build on the strength and excellence of Informatics, and Informatics is certainly the largest "member" of Bayes in terms of data science and AI capability. These are exciting times as we are establishing a new way of collaborating with each other and with a host of external partners in situ at the University. While Bayes is the result of over three years of hard work of many people at Informatics and across the University - especially that of the late Jon Oberlander, whose vision and enthusiasm was a key driver behind the initiative - we are still only at the beginning of the journey. It is part of an ambitious plan to turn our research, education, and innovation expertise into real-world impact and economic growth for Edinburgh and the region - and, ultimately, to demonstrate how data technology can benefit society around the world.

As the Centre takes starts up its activities over the next few months, there will be frequent updates to the (still rudimentary) website - make sure you sign up to the mailing list or follow the Bayes Centre on social media to stay in touch.

Bayes Centre website

Closing the student feedback loop interviews

Professor Stuart Anderson and Neil Heatley took part in interviews to share good practice among the University community.

Watch the interviews

Student news

ICSA student Rui Li won Brendan Murphy prize

This prize is in memory of Brendan Murphy, who was an outstanding researcher and mountaineer known to many in the communications and distributed systems research community, and a regular participant at MSN since its inception.

IPAB student Kunkun Pang won Piero Zamperoni Best Student Paper Award at ICPR

The primary purpose of this award is to acknowledge and encourage excellence in pattern recognition research by students, and to help assure the future livelihood of the field. The award also honours the memory of Dr. Piero Zamperoni, an outstanding educator in pattern recognition

NSS Scores

The School had the largest % increase in overall satisfaction across the University, going from 72.7% in 2017 to 84% in 2018
We went from 20^th out of 21 Schools in the University, and 7^th out of 7 Schools in College to being 6^th out of 21 in the University and 2^nd out of 7 in the College
Our scores improved on nearly all measures, with particular improvement on “learning community”, but there is still much room for improvement

Some of the comments from our final year students:

‘Lecturers were often very passionate about what they taught and very willing to answer questions or discuss things beyond the syllabus.’

‘I met some great lecturers who really care about their students and teaching them.’

‘There are very supportive staff particularly, personal tutors and members of student services who have helped me through difficult times at university. It is their ongoing support that has allowed me to persevere through these hard times. Whether it be offering extensions as needed, or providing academic advice, or just having a chat. Some lecturers are terrific and their classes are honestly a pleasure to sit through. Topics that in reality should be boring are made exciting and fun by these staff who think outside the box. They explain things in different ways so that those whose minds work differently can all find a way to wrap their head around difficult concepts. They tell jokes and are open and friendly and allow for questions to be asked and they don't make you feel lesser for asking what may be a stupid question. These lecturers and teaching staff remind me every day of why I wanted to do this degree in the first place. The university also provides opportunities for careers fairs, networking events, societies which allow us to get to know fellow students, graduates and staff on a more personal level. They allow the university to feel more like a community.’

Travel fund

3rd year Computer Science student James O'Donnell was awarded £100 from the Informatics Travel Fund to go to Oxford to take part in the UKSEDS Lunar Rover Competition 2017/18.

Find out more about how the team got on

EUFS won the UK’s first driverless race car competition

Edinburgh students win driverless race car competition

HypED (Edinburgh Hyperloop Society Team)

Informatics students contribute to developing transport of the future

Other information

Recycling

Photo (a photo of the bins)

We all care about the environment and hence, we should all recycle! Confused about what goes into which bin? See links below!

Office bins

How do I dispose of...

Marketing materials

The School now has new marketing materials available. We have postcards for undergraduate, postgraduate taught and postgraduate research as well as the 2019 entry undergraduate prospectus. These are available in the atrium or in the Communication Office (3.18).

Future promotional materials will include a new school brochure, 2019 entry postgraduate prospectus and updated school banners for events. If you require any AdHoc materials or templates, or have any suggestions, please get in touch with Infcomms.

Contact Infcomms

Outreach and Public Engagement

Call for contributions, Curiosity Live, Glasgow Science Centre, 9-11 Nov

Glasgow Science Centre are looking for contributions involving science, arts, history and pretty much anything else! Ideas linking sciences/engineering with other parts of society are especially welcomed. If you have any potential ideas, then contact Sam Langford.

Contact Sam Langford

Internal Call: UoE family events at the Museum at the Edinburgh International Science Festival 2019, deadline - 14th Sept

The University is once again putting together a family events programme based at the National Museum of Scotland 6 -20 April 2019. Janet Paterson is once again coordinating the programme.

If you would like to be part of this programme, then please look at the information documents and complete the application form available via the link below.

Application form

[Password: EISF2018]

General: Edinburgh International Science Festival have a call for ideas just now and up till 14th September

The call is open to anyone – individuals, groups or organisations. In honour of the 50th anniversary of the Moon landings, the 2019 Festival theme is Frontiers, exploring the boundaries of knowledge and the spirit of adventure and enquiry that drives science, technology, engineering and maths. Within this theme, specific areas of focus will include Healthcare Frontiers, Engineering Frontiers, Digital Frontiers, Environmental Frontiers and Planetary Frontiers.

You can respond to the call online via the link below.

Get involved

If you only have a vague idea and would like to discuss it, please contact the Comms Team. You can also contact the EISF team directly, Jennifer Rodger-Casebow is their events developer and can help you develop your idea into an event.

Contact Jennifer Rodger-Casebow

Events

Doors Open Day

The School will be participating in Doors Open Day again this year. See below for details.

Doors Open Day 2018

The Auld Alliance French Movie Club - second movie showing!

Come watch French movies (with English subtitles) and socialise with like-minded folks - the next showing will take place on Thursday 27 September, at 6pm in MF2.

Please contact Ségo if you have any question, or to sign up if you haven't done so already!

Contact Ségo