Category Archives: Activities

An activity refers to a series of events linked together under a specific theme or goal.

Prospects in Data Science – A multidisciplinary symposium

For three days in January some of the key thinkers in the field of data science met together in the elegant and inspiring surroundings of the New Forest just outside Southampton to reflect on the current state of play in Data Science.

Prospects in Data Science – A multidisciplinary symposium

The event took place over 12th – 14th January, 2016 at the Cerys Manor conference centre and hotel in Brockenhurst in the heart of the new Forest.

This was a high-level event on Data Science organised by the University of Southampton which aimed to look at the Prospects for Data Science, understood as the interface between Statistics, Mathematics and Computer Science providing new methods for handling, analysing and extracting knowledge from data, including Big Data. There was a good turnout for the event over all of the days.

IMAG0996

Anne Trefethen, Chief Information Officer, Pro Vice Chancellor, and Professor of Scientific Computing at the University of Oxford presenting her perspective on data science

Outline programme:

Day 1 (Tuesday 12) 

  • David Hand Imperial College London The roles of models in data science
  • Peter Grindrod University of Oxford   Red Herrings and Wild Goose Chases – Creating Analytics for Impact
  • Wendy Hall    University of Southampton   Observatories and data analytics for Web Science
  • Frank Wood University of Oxford     Probabilistic Programming
  • John Aston University of Cambridge
  • Fai Cheng Lloyd’s Register     Big Data and its Transformational Effects

Day 2 (Wednesday 13)

  • Gunnar Carlsson  Stanford University
  • Vitaly Kurlin Durham University  Topological ComputerVision
  • Jane Naylor Office of National Statistics  Data science and its role within official statistics
  • Anne Trefethen  University of Oxford
  • Sofia Olhede
  • Posters and networking

Day 3 (Thursday 14)

  • Arthur Gretton  University College London  Kernel nonparametric tests of homogeneity, independence and multi-variable interaction
  • Dave Coplin Microsoft UK  The Rise of the Humans: How to Outsmart the Digital Deluge
  • Yike Guo  Imperial College

David Hand, Imperial College London, gave an excellent opening talk which set the scene for the rest of the conference. Data Science and Big Data have been around for a while, understood by mathematicians, but only now are they receiving greater interest from the public, media and politicians.

Peter Grindrod, University of Oxford, presented a provocative reflexion on the challenges and opportunities involved in setting up a Data Science programme. In other words, an examination of the big, high-level issues that have to be tackled in order to achieve a successful educational outcome. One of the key aspects of this is the identification that strong leadership skills are needed at all levels in clouding at the funding bodies themselves.

Overall, there were, to my mind at least, a number of key messages that emerged from the talks and debates that look place over these three days. There is something of fight, or at least competition, to own data science. This is problematic as data science is something of a skill set or approach that spans traditional disciplines which still need to thrive. There has not yet emerged an agreed understanding of what data science is and what skills are needed by practitioners. It is also clear that data science is a fast moving and evolving field. Finally, the call to action for senior policy-makers to grasp this topic and understand the need to provide some subtle steer on the trajectory and velocity of change.

Internet of Things and Food: ITaaU/FSA programme outcomes

The Internet of Things offers great potential benefits for health and well-being in many areas, not least in how we manage the production, transportation and storage of the food that we eat. Understanding how we can and should benefit from the mix of technology and process requires that a number of key sectors share thoughts, knowledge and vision. We see this discussion as needing input from policy-makers, agencies, academics and industry. We welcome participants from such areas to learn about the work of our recently funded pilot projects and to contribute to the debate on future direction of this exciting field.

This event brings together researchers from the recent programme of pilot projects run by the Food Standards Agency and the RCUK-funded IT as a Utility Network+. Key outcomes including the benefits and potential for IoT in improving security across the food chain or network will be presented. The role of data, both open and closed, will figure strongly in these conclusions. The event is targeted at researchers and policy makers in the fields of IoT, food, food security, food transportation and storage, and also wider environmental issues.

The event has been structured to present detailed discussions on Monday with an overview and more strategic discussion on Tuesday. So that policy makers might wish to attend only day 2, and academics and other experts both days or even just day 1. (do let us know if only intersted in partial attendance.)

Event details for the 7-8 March, London – further information

Food Standards agency (FSA)/ITaaU IoT projects workshop, 18 January

This event is designed to welcome funded project members to the University of Southampton and introduce them to the ITaaU and FSA teams.

In addition, we will be preparing for the larger event that we are planning in London on the 7-8 March. This meeting, which will be held in Westminster, will act as a showcase for the programme and present the outcomes of the project to a wider audience of specialists, industry figures and policy-makers.

The programme for the day is as follows:

10:30 – 11:00 – registration and coffee

11:00 – 11:30 – welcome and intro to ITaaU and the FSA

11:30 – 12:30 – intro to the 4 projects

12:30 – 13:15 – sandwich lunch in the meeting room

13:15 – 14:30 – project plans, needs and opportunities

14:30 – 15:15 – planning for the London meeting – 7-8 March

15:15 – 15:30 – wrap-up

Are we really multidisciplinary? TRIFoRM bridges some UCD gaps

As the TRIFoRM project draws to a close with results about the levels of trust which motivated users are willing to place in the technology they use, we are particularly mindful of one of the original goals of the ITaaU: to encourage and facilitate multidisciplinary interaction and collaboration. In a nutshell, we need to learn to work with each other, understanding different perspectives and views, and exploiting the final outcomes of truly cross-disciplinary insights. In that vein, TRIFoRM has uncovered another layer of complexity. It is not just about the engineers and researchers from ICT as well as the social sciences who work together to understand the problem and towards its successful resolution, but it’s about what the potential beneficiaries understand and how they react.

TRIF0RM-image

“Estirar de la soga” by Macobru – Own work. Licensed under CC BY-SA 4.0 via Wikimedia Commons – https://commons.wikimedia.org/wiki/File:Estirar_de_la_soga.jpg#/media/File:Estirar_de_la_soga.jpg

Looking at trust and trustworthiness in ICT systems, it’s to be expected that HCI experts, technologists (especially in our case from the IoT and sensor networks) and social scientists would all need to collaborate and contribute to understanding the parameters around technology and service adoption. This cross-disciplinary engagement was needed, of course, to identify just what we need to look for in terms of a potential user’s propensity to trust and their behaviours towards technology: and this is one of the huge benefits of the collaborative networks building up around the ITaaU forum. Far less obvious, though, was the collaborative dimensions around the technical service we were investigating.

In TRIFoRM, we engaged with both clinicians and patients involved in the management of chronic pain. The technology component aims to provide monitoring and self-reporting capabilities on a smart phone or similar which would remotely upload data to a central server where those data would be aggregated such that the clinical team from the consultant to the specialist nurse would access to relevant long-term objective data to support and supplement direct patient contact. We should remember at this point that any such service would involve sensitive personal data. That in itself set off alarm bells: for the technologists, would the clients and servers as well as connections between them be robust enough to protect and maintain data integrity and prevent intrusion? For the social scientists, how could potential users be assured that the technology would appropriately manage such data? And for the end users themselves?

For chronic pain sufferers, there were no such issues. Anything which might help them as well as the medical team manage and relieve their condition would be welcome; data security was not an issue. But are they simply naĂŻve to the risks associated with technology? After all, as our clinician pointed out: systems are only as secure as their users. The point is rather more to do with personal relationships. Overwhelmingly, users would trust the technology in the same way that they might trust the medical team. There is however an important caveat: trust transfer of this kind whereby benevolence is perceived in the technology as a result of confidence in the medical team would occur and persist solely if it added to and did not undermine the person-to-person relationships that the patients valued with the clinical team. In an interesting perspective on ease of use and utility, the other common trustworthiness feature associated with benevolence, our participants highlighted that technology might alleviate or compensate for the physical as well as cognitive limitations which their condition imposed.

At the recent ITaaU event in Southampton, the surgeon David Rew made the point that what he and his colleagues wanted was intuitive presentation of the complex patient data they needed to digest to support their patients; turning away to consult a computer screen was not an option. That is one very important factor. Patients themselves however want to support and enhance that relationship through technology too. What TRIFoRM has shown us and our colleagues is the importance of going beyond user centred design to look at the personal and social context of technology deployment and acceptance.

Taking TRIFoRM out to the community

Last week marked a high point in the TRIFoRM project, as we were able to travel to the ITaaU Community Conference to engage with a wealth of other fascinating projects while showing off our work.

In terms of the “showing off” part, we first gave a lightning talk about what TRIFoRM is doing and why, how it relates to other projects such as OPTET, and – last but not least! – what we have achieved. We followed this up with a poster and demo; three of the project investigators were able to talk with conference attendees in much more detail about TRIFoRM’s aims, objectives and achievements. We were, in fact, so busy talking to people during this session that – alas! – we did not manage to get a photograph of our poster or ourselves in action. Don’t take our word for it that the poster looked good, though:

TRIFoRM

As you can see, we had plenty to talk about. Since our previous blog post, we’ve been busy: led by the social scientists in our team, we have assembled a state of the art on trust (considering trust of humans and also technologies), before using that theoretical grounding to guide semi-structured interviews with stakeholders. What stakeholders, you ask? We chose to focus on healthcare monitoring technologies for people suffering from chronic pain, and spoke to people with such issues as well as someone who provides services for these people.

We got some great insights. The two blue diagrams in the bottom left of the poster are thematic maps showing knowledge we gained about technology acceptance and trust transfer, while on the right are diagrams representing two types of threat to trust that we identified from those interviews.

If you want to know more, don’t worry: our third and final blog post will give more detail about our key findings.

 

 

TRust in IT: Factors, metRics, Models

Guest post by: Dr Brian Pickering of IT Innovation at the University of Southampton

I live in a small, Georgian listed house, and was recently forced into the 21st century when a central heating engineer sucked his teeth (as they do) but instead of selling me a new boiler (as they usually do) he convinced me to install a heating control system which, he assured me, would not only fix the problem he couldn’t seem to, but allow me to communicate with my boiler remotely (which it does). He didn’t just mean I could sit in the living room with a wireless thermostat, but really: from any web browser or tablet or smart phone, I can monitor what’s going on, change schedules, reduce the temperature if my son thinks no one is looking and turns it up. At last a good use of technology…

Triform

Just a minute though. If I can do all of this remotely, then presumably so could anyone else. Forget my central heating for a moment, I need to think about this. I’m used to being concerned about data privacy and protection. If I use social networks and easily available webmail, I should be mindful of what providers expect me to accept[1],[2]; I know about the concerns around care.data[3]. Certainly for medical records, concern has been publically discussed for some time[4]. More recently, though, it appears I should be thinking about other aspects of healthcare: the embedded devices that monitor my condition when training, or keep my heart beating, or release the right levels of insulin at appropriate times, are vulnerable to attack[5].

And all of this is against a background of our increasing use of sensors and embedded devices. Back in 2012, the number of non-PC devices connected to the Internet exceeded the number of PCs; and the number of household devices with embedded ICT capabilities is set to overtake all other IoT connected devices this year (2014) and be almost double that of any other device types by 2020. Sensors and tags in general will experience exponential growth at the same time. I really think it’s time to stop and regroup.

This is not just about technology robustness to misuse and external attack. This is really to do with how we perceive technology and whether we want it to be so pervasive. The big question is whether users really trust the technology, and if so why. The TRIFoRM (TRust in IT: Factors, metRics, Models) project is looking at just that. If we start with a small specific group of users, who either need technology to support them day-to-day or use it as a necessary adjunct to their work, we really want to establish what it is about technology they trust, if at all; why they are willing to trust it; and what makes them trust it at all. Is it because we are simply being forced to take ever greater risks because there are no alternatives? Are we relying on legal regulations? Or is there something inherent in design or implementation that makes it feel OK?

ACKNOWLEDGEMENTS

The Mirror and BBC News logos and headlines are used in accordance with the Section 29 of the Copyright Provisions 1988

The upper image is public domain, available at http://commons.wikimedia.org/wiki/File:ATM_Masalli.jpg?uselang=en-gb; the lower image http://commons.wikimedia.org/wiki/File:Next_Generation_Backscatter_Device.jpg is copyright of Tek84 (www.tek84.com). There is no implied or actual connection between the headline and the product depicted.

 

[1] http://europe-v-facebook.org/EN/en.html

[2] http://www.theguardian.com/technology/2013/aug/14/google-gmail-users-privacy-email-lawsuit

[3] https://medconfidential.org/how-to-opt-out/

[4] Rindfleisch (1997) Privacy, Information Technology and Health Care Communications of the ACM , Vol. 40, No.8

[5] Mansel & Kohno (2010) Improving the Security and Privacy of Implantable Devices New England Journal of Medicine, 362; 13, 1164-1167

BluPoint

BluPoint:

The next 3 billion people connecting to the Internet will be “mobile-only” users and live in least economically developed countries. Access to relevant information about health, education, produce prices, weather, news and entertainment can lead to a better life and a route out of poverty. However many people are unable to harness the power of the internet due to both the availability/cost of data services and the type of handsets that they own being to basic. BluPoint, provides a physical access point to enable people in off-grid low resources communities to both access free digital materials on their basic and smart mobile phones and create/share their own digital content within their communities.BluPoint can be set up as a single unit or a network providing a pop-up, solar powered intranet, providing  digital content and services to locations without electricity or affordable internet.

 

Uplands Rescue Resilience

Uplands Rescue Resilience:

It is very expensive and time consuming for the Police and mountain rescue volunteers when a search is prolonged due to a lack of precise information about where citizens in danger may be located. One of the major problems for rescue teams is to know the correct location of the potential casualty in distress as the area of search can be very large and the weather conditions may be bad. Upland areas contain many “dead” zones that are blind to radio and cellular radio communications.

 

Pat Langdon

Pat Langdon’s experimental doctoral and post-doctoral research has contributed to cognitive science, artificial intelligence, robotics and psychophysical studies of the human visual system. His experimental PhD researched the psychological reality of certain Artificial Intelligence-based theories of vision and he was employed by an Artificial Intelligence Vision Research Unit at the University of Sheffield working on the area of robotics. Subsequently, he moved into Engineering Design Research at the University of Lancaster EDC, working on the application of AI and user centred design principles to complex engineering design systems. Since 1998 Pat Langdon has been employed at the Cambridge EDC, where he has pursued a number of research directions.

These include: (1) the properties and design of Haptic interfaces for use in accessible computer displays for the movement disabled; (2) the representation and formulation of statistical data on disability, for use in Inclusive Design; (3) Integration of software development and empirical methodology for Good Design Research Practice; (4) structure and survey of data for capability assessment; (5) comparative studies of clustering methods for design knowledge exploration; (6) cognitive scales for capability assessment; (7) methods for ethnographic and observational studies of aerospace (7) developing scales for product-capability interaction assessment in design.

 

Trusted Tiny Things

The aims of this pilot project are thus:

 

Lightweight model

To create a lightweight model to represent information about devices in the IoT such as: capabilities, security properties, ownership and provenance of devices and services. This will involve building upon existing work being conducted within the dot.rural Digital Economy Hub on vocabularies to describe sensors and Internet enabled devices using the W3C Semantic sensor network ontology and existing work on provenance.

 

Software framework

To build a “Trusted Things” software framework based on Semantic Web technologies and services to store and query information about devices in the IoT and their associated provenance. (This will involve attaching NFC tags to ‘things’; To develop an initial set of guidelines that could support IoT developers to describe information about devices according to our model. This will include specific guidelines on how provenance of the information generated by such devices can be tracked and how devices can be instrumented to provide real_time information to the “Trusted Things” software platform.

 

Demonstrator

To evaluate this approach using a demonstrator application based on two transport related scenarios: the use of passive NFC tags in bus shelters in Aberdeenshire; the use of in-car black boxes to track the behaviour of drivers for insurance purposes. The latter will be possible by building a black box based on an Arduino programmable board and sensors. This will collect driving data such as speed, acceleration braking forces, route data and timing of journeys using a combination of GPS and accelerometer data and will transmit this to a our framework using an on_board 3G connection.

 

Project leader

Edoardo Pignotti leader of Trusted Tiny Things is a Research Fellow working on trust and provenance issues in Linked Data at the dot.rural Digital Economy Hub at the University of Aberdeen. He has more than seven years experience in Semantic Web technologies, provenance and policy based reasoning gained during his involvement with a number of UK eScience projects.

Libraries of the future

Libraries of the future has been one of the longest running themes that has been explored by the ITaaU Network+. The first related workshop took place at the Bodleian Libraries Delegates Room in Oxford. There have been a number of workshops following on from this around the UK. A self-selected working group has coalesced around this theme with many interesting ideas and concepts emerging.