Ivan Palomares speaks and is interviewed at Search-Star’19 conference (Bath, UK)

The SearchStar 2019 conference on Analytics of Conversion was held in Bath last 27th June, organised by Search Star , an UK-based company founded in 2005 and focused on advancing digital strategies for marketing and advertisement.

Ivan Palomares from DSRS was one of the speakers at the event, where he delivered the talk Personalisation & Recommender Systems: Perspectives and Challenges.

Image

A summary of the conference, including Ivan’s talk, was published in Search Star blog:

Iván offered his audience an insight into the revolutionary world of recommender systems and the role it will play in consumer choice in the future. Iván started by presenting the audience with a problem – ‘you want to buy a new book.’ Twenty years ago, we would have gone to a neighbourhood bookshop to ask for a recommendation. Nowadays, with the same question, we turn to the internet and are instantly overwhelmed by an extensive amount of results.
What’s the solution? Personalistion through recommender systems. These are evolutionary algorithms designed to provide tailored content to the user, by extracting knowledge about your preferences from previous choices and choices of similar users. If you’ve ever listened to songs on Spotify or bought from Amazon, then these recorded decisions are used to inform these systems about your preferences.
There are many techniques used to tailor these recommendations; you could either be served results based on your demographic, where you are at a certain time and the opening hours of the shops around you (context-aware), or choosing a selection of restaurants based on the reasoning that people with similar interests to you also enjoyed those restaurants (collaborative filtering).
There are challenges to this approach, specifically when they come across a ‘cold user,’ where they lack information on their preferences, or the essential requirement to combine multiple views and sources of user data.
We also had a significant question from the audience regarding data privacy, however Iván responded that all data used to make the recommendations are information that the user has provided either explicitly or implicitly, and it is on the usage of implicit user data where privacy and ethics concerns really come into scene.

Our team member was also interviewed prior to his talk. You can read the interview, entitled ‘What are Recommender Systems?’, here

Zijian Shi – Congratulations and Future Welcome to Bristol and DSRS!

Zijian Shi, our upcoming DSRS researcher

We are pleased to congratulate Mr Zijian Shi, who has obtained a prestigious UK-China PhD scholarship from the China Scholarship Council (CSC) to initiate his PhD studies at the University of Bristol, with the DSRS Lab, after summer 2019.

Zijian has  international experience studying abroad, having received two Masters’ Degrees: one on Management Sciences and Engineering from Tianjin University (China), one of our lab’s international collaborators and a top Chinese university; as well as one on Finance from The University of Texas at Dallas (United States).

Zijian has gained research expertise in decision support systems applied to financial analytics, and he has prior experience in scientific publications, e.g. an authored article resulting from Tianjin University’s collaborations with DSRS, published in the high-impact journal Knowledge-based Systems.

We look forward to welcome you to DSRS soon!

Professor Xueqing Wang (bottom center) and her lab at Tianjin University, with Zijian Shi sitting to the right.

DSRS and Bristol’s Jean Golding Institute organise RESULTS Interdisciplinary workshop – 23rd May 2019

RESULTS: REcommender Systems for engaging Users with healthy Living habiTS

The DSRS research group is organising an exciting interdisciplinary workshop, as part of the 2019 Data Week celebrated next May at Bristol’s Jean Golding Institute. See below for event information and our three confirmed speakers’ profile.

Registrationregister for this event via Eventbrite (PLACES LIMITED)

Event enquiries: Please contact Iván Palomares Carrascosa, workshop chair, for any event-related enquiries: i.palomares (at) bristol.ac.uk

Overview 

This workshop is aimed at researchers, students, academics and practitioners across the Turing network of UK universities and beyond, as well as members of related industry sectors. 

There will be a discussion forum on the challenges and opportunities of personalisation approaches, data-driven Decision Support and Recommender Systems (RecSys) in the areas of fitness, wellbeing and promoting healthy living. We seek to bring together experts across the disciplines of Data Science, Decision Support and AI, Medical Sciences, Data Ethics, Psychology and Behavioural Sciences, Nutrition and Sports/Physical Activity.  

Event focus 

  1. Connecting experts across disciplines (health, nutrition and sport, etc.) with scientists in Data Science/AI and industry, to put forward potential solutions to unaddressed challenges involving personalisation and/or decision support 
  1. Data ethics implications and user data protection regulations in personalized healthcare and wellbeing.  

Participants with ongoing research activity on decision support approaches or RecSys for the health and wellbeing domains are encouraged to bring their laptops to showcase or discuss their results with experts in the field. 

Talks 

Daniele Quercia, King’s College London & Nokia Bell Labs.

Healthy Cities: Tracking Population Health from Grocery Bags and Smart Watches 

Quercia,DanieleBio: Daniele Quercia is with the Department Head of Social Dynamics at Nokia Bell Labs Cambridge (UK) and Professor of Urban Informatics at the Center for Urban Science and Progress (CUSP) at King’s College London. He has been named one of Fortune magazine’s 2014 Data All-Stars, and spoke about “happy maps” at TED. His research has been focusing in the area of urban informatics and received best paper awards from Ubicomp 2014 and from ICWSM 2015, and an honourable mention from ICWSM 2013. He was Research Scientist at Yahoo Labs, a Horizon senior researcher at the University of Cambridge, and Postdoctoral Associate at the department of Urban Studies and Planning at MIT. He received his PhD from UC London. His thesis was sponsored by Microsoft Research and was nominated for BCS Best British PhD dissertation in Computer Science. 

Abstract: We will see how to aggregate both readings from consumer wearable devices and records of food purchases to track people’s well-being at scale. From 11,600 Nokia Health wearables, we collected readings of steps, sleep, and heart rate in the entire cities of London and San Francisco over the course of 1 year. Christmas and New Year’s eve were associated only with short-lived and minor disruptions, while both Brexit and Trump’s election greatly impacted people’s sleep and even heart rates. Then, for another entire year in London, we studied the association between food purchases in grocery stores, as measured by the digital traces of customer loyalty cards, and consumption of medicines. Our results show that analytics of digital records of grocery purchases can be used as a cheap and scalable tool for health surveillance: the distribution of the food nutrients is far more predictive of food-related illnesses (e.g. diabetes) than socio-economic conditions. 

Morgan Harvey, Northumbria University.

Balancing the Healthy with the Tasty: Recommending Nutritious Food that People Will Actually Want to Eat 

Morgan HarveyBio: Morgan Harvey is a Senior Lecturer of information science in the Department of Computer and Information Sciences at Northumbria University, Newcastle. He has been conducting research in the fields of information behaviour and retrieval and recommender systems for over a decade and has published over 50 peer-reviewed conference papers and journal articles, including in ACM SIGIR, ACM CIKM and ACM RecSys. Morgan received his PhD from the University of Strathclyde under the supervision of Prof. Ian Ruthven and has worked as a researcher at the University of Erlangen-Nuremberg in Germany and the University of Lugano in Switzerland. 

Abstract: Poor dietary habits are a major cause of today’s world health problems, especially in the developed world. Evidence shows that such issues can be prevented and sometimes even reversed through good nutrition, however, people are often very poor at judging the healthiness of their own diet and need support to implement positive changes. Computer and information technology, and particularly recommender systems, have been suggested as potential solutions to some of these problems but much of the necessary work is yet to be done. This talk will discuss existing research on the problem of recipe recommendation, the difficulty of finding a balance between healthy and enjoyable choices and early work investigating how people can be “nudged” into making healthier food choices. 

Max Western, University of Bath

Multidimensional physical activity: using the right data from wearable fitness trackers and providing the right feedback

Photo of Max WesternBio: Max is a lecturer in behavioural science within the Department for Health at the University of Bath. He has keen research interests in the factors that influence and facilitate health-enhancing physical activity behaviour, the application of digital health technologies for measurement and intervention and healthy ageing. In recent years Max has worked on several large lifestyle-based intervention trials funded by the MRC and NIHR. Specifically, he has helped develop and evaluate interdisciplinary projects aimed at reducing risk of future chronic disease using wearable monitoring technology to provide feedback on multidimensional physical activity; combatting physical frailty in older age through a community-based physical activity programme; and reducing the decline in cognitive function in older adults with mild-cognitive impairment using a digital web-based behaviour change intervention.

Abstract:  The benefits of regular physical activity behaviour on physical function and health are well documented and it is apparent that there are multiple dimensions to physical activity are important. The accurate measurement of physical activity is key to understanding the relationship between behaviour and health, providing credible feedback to participants and judging the success of behaviour change interventions aimed at increasing levels. Accelerometers are popularised as being the most objective and accurate format of assessment. This talk will draw on research data to discuss the strengths and limitations to their current use, such as their (in)appropriateness for comparisons with public health guidelines and the interpretation of what counts as physical activity, and propose solutions to these issues that should guide future applications of these measurement devices.

Event co-sponsored by: Jean Golding Institute for Data-Intensive Research, University of Bristol, and Bristol’s Intelligent Systems Lab

Bristol’s Jean Golding Institute interviews Iván Palomares

This week, we had the pleasure of being interviewed by the Jean Golding Institute for Data-Intensive Research at our University. The Institute started a new series of interview posts in their blog, where the newly appointed Turing Fellows with the Alan Turing Institute are introduced: the first interview features Iván Palomares Carrascosa. You can read it here

IEEE Smart-PDS Workshop in Leicester, UK – Call for Papers

1st International Workshop on Personalisation and Decision Support for Citizen Service Enhancement at Scale (IEEE Smart-PDS 2019)

https://smart-pds.github.io/

Leicester, UK. 19-23 August 2019

Related image

We are pleased to chair a co-located workshop with the 2019 IEEE Smart World Congress and the 3rd IEEE Conference on Smart City Innovations (IEEE SCI 2019) will be held in Leicester, United Kingdom, between 19th 23rd August, 2019.

The workshop aims at attracting demonstrations of novel research to tackle challenges faced in smart cities, with a particular focus on Big Data-driven user personalisation and large-scale decision-making problems. Research topics include applications, techniques and methods with or without case studies, related to personalisation applications of hybrid and context-aware recommender systems, solving the cold-start problem in smart city contexts, preference elicitation and its integration with Big Data, multi-criteria and group decision support systems at scale.

Researchers and practitioners working in different fields are encouraged to submit their research contributions and ideas in addition to demonstrators and visual materials such as posters.

Topics of Interest

This workshop covers contributions describing recent advances related to personalisation, recommender systems, as well as data-driven and multi-criteria decision making approaches at large scale, to improve citizen and visitor services in highly connected and data pervaded Smart City settings.

Particular topics of interest include (but are not limited to) the following:

Smart city challenges and applications:

  • e-governance
  • tourism
  • leisure
  • socialising
  • health and wellbeing
  • sustainable cities
  • participatory democracy,

… involving:

Large-group decision making

Multi-criteria decision analysis models

User preference modelling and preference aggregation

Context-aware recommender systems

Group recommender systems

AI/Machine Learning approaches for recommendation and decision support

Big Data technologies in personalisation and decision support approaches

Sensor data fusion and uncertainty handling

Role of the IoT in personalisation and decision support services

Ethical and legal aspects of data-driven personalisation and decision support 

Important dates

Paper submission due: April 26, 2019

Notification of acceptance: May 10, 2019

Camera-ready papers due: May 19, 2019

 

For further info and submission instructions, please visit the workshop website

We look forward to see you in Leicester next summer! 🙂

Iván and Ercan, workshop chairs

Ercan Ezin – 1st place Award in 2018 Bristol Data Jam and Hackaton

Congratulations to DSRS’s Ercan Ezin: after a restless and highly devoted weekend at the 2018 Bristol Data Jam and Hackathon, he and his team – led by Frank Kelly – won the 1st place “Databrain” Award to the Best Project Idea developed to promote citizen services in Bristol with its Open Data platform. The event was co-organised by Bristol City Council, ConnectBristol and UoB’s Jean Golding Institute for Data Intensive Research. The project consisted in a GIS-based tool to promote eco-friendly routes in the city.

Since a tweet is more worth than thousand words, here we go:

Hoping the is one out of many successful milestones for you!

Back-to-Basics: Group Decision Making

Supporting collective decisions made by groups of experts or stakeholders with their own different opinions, and supporting recommendations for groups of users that satisfy them all, constitute some of the fundamental research interests in our team. But, what are the principles and basic ideas underlying Group Decision Making?

In this post, we attempt to answer this question by introducing a brief introduction to the topic. The following content extends my originally written overview in the AFRYCA website back in 2014, whose software suite I founded as part of my PhD thesis:

A Group Decision Making (GDM) problem is a decision situation where:

  • There exists a group m of individuals or experts, E = {e1, … ,em}, who each have their own attitudes and knowledge. Consider for example E={Binyamin, Ercan, Hugo, Ivan, James} as a group of five experts. Intuitively, m should be equal or higher than 2, since a group should have at least two members 🙂
  • There is a decision problem consisting of n alternatives or possible solutions to the problem, X = {x1, … ,xn} (again, at least two). Say, for instance, that our five friends in group E want to choose a destination for a research group away day, among four possible options in England: X = {Costwolds, Cambridge, Oxford, Cheddar*}. 
    (*Note: Cheddar is a beautiful place in southwest England, not necessarily the cheese!)
  • The experts try to achieve a common solution, i.e. a final decision on selecting one of the four possible destinations.

Each expert ei , with i=1,…m, expresses his/her opinions or preferences over the alternatives in X, in other words, they provide judgment information indicating to what extent they support – or don’t support – each of the available options.  For this, each participant supplies a preference structure. Some examples of preference structures widely used in scientific literature related to GDM, are:

  • Preference orderings: A ranking established by each individual, establishing their ordering of alternatives from the most to the least preferred one. For instance, Ivan may provide the following preference ordering according to which Oxford is his most preferred place and Cheddar is his least preferred one. (Please note, this is merely and example and Ivan absolutely likes all four destinations! 😉 ):

Oxford > Costwolds > Cambridge > Cheddar

  • Utility vectors: Despite being the most intuitive preference format to be provided by humans, preference ordering are also the least informative ones: they allow to express that Cambridge is preferred over Cheddar, but they do not allow to indicate how strongly Cambridge is preferred over Cheddar (a lot more? just slightly?). A suitable structure to indicate degrees of preference or likeness on each alternative are preference vectors. The following example shows a preference vector over X = {Costwolds, Cambridge, Oxford, Cheddar} with assessments in the [0,1], such that the higher the value of the assessment, the more preferred the alternative is. In this case, the fourth alternative in X (Cheddar) is the most preferred one, whereas the second alternative (Cambridge) is the least preferred one:

[0.6   0.4   0.7    0.8]

  • Fuzzy preference relations: Represented as an nxn square matrix, where each element located at the row and column k  – excluding those in the main diagonal of the matrix – is called assessment and represents the degree to which the lth alternative is preferred against the the kth alternative in X. Assessments therefore describe “pairwise” comparative judgments among alternatives. For instance, assuming we use the [0,1] interval to assess pairs of alternatives, an intermediate value of 0.5 indicates indifference between two alternatives (see 0.5 for indifference between Costwolds and Cheddar), a value higher than 0.5 indicates preference towards the first alternative in the pair (see 0.8 for strong preference on Cambridge with respect to Oxford) and, conversely, a value lower than 0.5 indicates that the first alternative in the pair is less preferred than the second one (see 0.4 for weak preference against Cheddar with respect to Cambridge).
Cost.  Camb. Oxf. Ched.
Cost. 0.4 0.6 0.5
Camb. 0.6 0.8 0.6
Oxf. 0.4 0.2 0.4
Ched. 0.5 0.4 0.6

 

  • Decision matrices: In some GDM problems, each alternative needs to be assessed in terms of multiple criteria, C = {c1, … ,cq}. Consider for instance the following scenario in medical treatment decision-making:
    A group of clinicians with expertise in making treatment decisions for patients with complex health needs, need to prioritise possible treatment options (alternatives) for patients with multiple diagnosis, symptoms and risk factors, by evaluating each treatment in terms of safety, cost and efficiency level (criteria). (Acknowledgements: Dr Rachel Denholm, Population Health Sciences, University of Bristol Medical School)

    In these contexts, we would define the decision framework as a Multi-Criteria Group Decision Making problem, in which case decision matrices would be the best approach to express preferential of judgement information by participating experts. Each element in a decision matrix is an assessment on a specific alternative (row) in terms of a specific criterion (column). For instance, the following decision matrix indicating that Treatment 3 is the safest but least efficient one:

Safety  Cost Efficiency
Treat1 0.7 0.4 0.5
Treat2 0.6 0.5 0.9
Treat3 0.9 0.4 0.2
Treat4 0.5 0.6 0.5

 

GDM problems are usually defined in environments of uncertainty, in which the information regarding the problem is vague and imprecise. These situations are also known as GDM problems under fuzzy contexts. Some information domains for expressing preferences, that have been frequently utilized by experts to deal with uncertainty, are: numerical, interval-valued or linguistic information. This post has shown examples of preference modelling with numerical information between 0 and 1, but other approaches exist to allow participants to assess decision information both quantitatively and qualitatively.

The solution for a GDM problem has been classically determined by applying an alternative selection process, which is composted of two phases:

  1. Aggregation phase: Experts’ preferences are combined at assessment level, by using an aggregation operator. How to aggregate individual opinions into a representative opinion, constitutes a key and very widely investigated question across the decision making and information fusion research communities.
  2. Exploitation phase: Once a group preference has been obtained, an alternative or subset of alternatives is obtained as the solution for the GDM problem, by applying a selection criterion (e.g. dominance or non-dominance degrees).

Interested in delving deeper into Group Decision Making and its extended lines of research (consensus building, large-group decision making)? Then we suggest you taking a look at the newly published book Large Group Decision Making: Creating Decision Support Approaches at Scale, which includes a detailed overview on Group Decision Making and Consensus Decision Making in Chapter 2!

Thank you for reading 🙂

From Tokyo to Bristol: DSRS welcomes James Neve!

Today we were delighted to welcome Mr James Neve on board, who has been awarded with a prestigious DTP scholarship (Doctoral Training Partnership) by the Engineering and Physical Sciences Research Council (EPSRC) to start his PhD studies at the University of Bristol. Under the supervision of Ivan Palomares Carrascosa, James will concentrate his research on Reciprocal Recommender Systemsrecsys approaches to recommend users to each other instead of recommending items to users. Online dating is one of the several related application domains of reciprocal recommendation approaches (along with making friends, job recruitment, etc.), which James investigated and worked on as part of his previous role of Principal R&D Engineer in Eureka Inc. TokyoJapan. The Japanese dating application Pairs, along as an underlying system to detect dubious users in such applications, are part of his pre-PhD work, which he will continue to actively investigate within the frame of a very interesting PhD project!

It is noteworthy that, prior to his official PhD start date, James has already recently participated in the IEEE SMC 2018 conference held in Miyazaki, Japan, where he presented his latest developments and research findings as an accepted full paper.

Here are some pictures of this participation in the IEEE SMC 2018 conference two weeks ago.

And finally, the traditional “growing family picture” with our new DSRS team mate!

On behalf of Ercan, Binyamin and Ivan, welcome on board James 🙂