This opportunity is now closed.

Funded PhD Opportunity

Annovate – enabling meaningful dataset insight through novel annotation mechanisms and approaches to dataset development

Subject: Computer Science and Informatics


Summary

Recently there has been an exponential increase in the quantity of data that is being produced globally.  Recent estimates indicate that there are 2.5 quintillion bytes of data created every day with this number set to continue its exponential increase (Taleb, Serhani and Dssouli, 2018). This data is generated from a variety of sources including human generated content and data produced by computational systems (Jones and Shao, 2013; Oussous et al., 2017; Ramakrishnan et al., 2017).

Human generated content includes microblogs, videos and photos. Data produced by computational systems includes (i) sensor data, such as temperature sensors and (ii) output from processes, such as automated analysis of video. In conjunction with the growth of datasets there has been a significant increase in the capability and adoption of machine learning to produce solutions which augment everyday life (Holzinger, 2016; Hazelwood et al., 2018), such as voice control through assistive agents.

To produce these solutions, machine learning techniques are generally applied to large datasets to produce a model which can subsequently be used to identify and predict patterns/trends of interest. In order to effectively learn from large datasets and produce useful models, machine learning algorithms require annotations within datasets to provide context and aid classification of outcomes (Yordanova et al., 2018).  An example of an annotation could a label in a dataset indicating the location and time of a food making activity within a sensorised environment (Rafferty et al., 2017).

There is a large demand within diverse research and industrial communities for well annotated datasets (Cordts et al., 2016; McChesney et al., 2017; Yordanova et al., 2018). An obstacle to the production of such datasets are the techniques currently employed to produce meaningful annotations.

Current approaches to dataset annotations exist but have a range of deficiencies which reduce their efficacy, including:

-Reliance on onerous annotation mechanisms, such as manually recording annotations and transferring these to a database

-Computational approaches which have an inability to flexibly model a variety of annotations dynamically

-Inability to model and reconcile time synchronisation issues

-Absence of a mechanisms to resolve annotation uncertainty

-Lack of a common dataset and annotation interchange format

This project aims to investigate approaches to annotation of datasets through pervasive computing technologies. Specifically, this would produce solutions to aid individuals in production of accurately annotated datasets with minimal friction. Primarily, this would produce an annotation solution which extends the sensing technologies, computing infrastructure and physical laboratory environments that are present within the School of Computing. Notably, the new living lab located within the school will augment research activities by providing a ground truth.

Research avenues to be explored may include:

i.Approaches to technology-assisted annotation, such as through smart object interaction, video recognition and voice assistants

ii.Exploration of augmented reality annotation and sensing solutions

iii.Investigation into mechanisms to resolve uncertainty within annotations

iv.Mechanisms to manage time synchronisation issues

v.Development of a common interchange format for annotated datasets

vi.Production of exemplar gold-standard datasets


Essential criteria

  • Upper Second Class Honours (2:1) Degree or equivalent from a UK institution (or overseas award deemed to be equivalent via UK NARIC)


Funding

    Vice Chancellors Research Scholarships (VCRS)

    The scholarships will cover tuition fees and a maintenance award of £14,777 per annum for three years (subject to satisfactory academic performance). Applications are invited from UK, European Union and overseas students.

    DFE

    The scholarship will cover tuition fees at the Home rate and a maintenance allowance of £ 14,777 per annum for three years. EU applicants will only be eligible for the fees component of the studentship (no maintenance award is provided).  For Non EU nationals the candidate must be "settled" in the UK.


Other information


The Doctoral College at Ulster University


Reviews

As Senior Engineering Manager of Analytics at Seagate Technology I utilise the learning from my PhD ever day

Adrian Johnston - PhD in Informatics

Watch Video  

Key dates

Submission deadline
Monday 18 February 2019

Interview Date
25 to 29 March 2019


Campus

Jordanstown campus

Jordanstown campus
The largest of Ulster's campuses


Contact supervisor

Dr Joseph Rafferty


Other supervisors


Applying

Apply Online

Related Funded Opportunities in: Computer Science and Informatics ,

Computing

Subject:
Computer Science and Informatics

 View Details

Intelligent Systems

Subject:
Computer Science and Informatics

 View Details