Research Associate

12789BR

Cardiff School of Computer Science and Informatics

Research

Advert

Research Associate
Cardiff University School of Computer Science & Informatics
Applications are invited for a Research Associate post in the Cardiff University School of Computer Science & Informatics, to work on the EPSRC-funded project ELEXIR (Encyclopedic Lexical Representations for Natural Language Processing). The aim of this post is to develop methods for characterising the commonsense relationships that hold between different everyday concepts (e.g. bees produce honey). To this end, we will first develop methods for extracting such relational knowledge from pre-trained language models. Subsequently, we will learn vector encodings of the extracted relationships and develop strategies for reasoning with these encodings in a principled way. This will allow us to improve the quality of the extracted information, and to uncover generic rule-like commonsense knowledge.

You will work closely with Steven Schockaert and Luis Espinosa Anke. You will possess or be near the completion of a PhD in Computer Science or a related area, or have relevant industrial experience.

The position is full-time (35 hours per week), available from 1.4.22 and is fixed term until 31.3.24.

The appointment will be made, dependent on relevant qualifications and experience, within the Grade 6 scale, currently £34,304 - £40,927 per annum.

For an informal discussion of the post please contact Steven Schockaert by email at [email protected].

Job reference: 12789BR

Date advert posted: Thursday, 4 November 2021

Closing date: Thursday, 25 November 2021

This post has previously been advertised to employees of Cardiff University only.  We now invite external applications.

Please be aware that Cardiff University reserves the right to close this vacancy early should sufficient applications be received.

Cardiff University is committed to supporting all its employees in achieving the necessary balance between their work and their personal lives. As part of this commitment, the University has developed a flexible and responsive framework of support for staff in managing their work and personal commitments wherever possible. [Applications for posts on a part-time or job-share arrangement would also be considered].

Cardiff University is committed to supporting and promoting equality and diversity. Our inclusive environment welcomes applications from talented people from diverse backgrounds. Cardiff University positively encourages applications from suitably qualified and eligible candidates regardless of sex, race, disability, age, sexual orientation, gender reassignment, religion or belief, marital status, or pregnancy and maternity.

“Cardiff University is a signatory to the San Francisco Declaration on Research Assessment (DORA), which means that in hiring and promotion decisions we will evaluate applicants on the quality of their research, not publication metrics or the identity of the journal in which the research is published. More information is available at: https://www.cardiff.ac.uk/research/our-research-environment/integrity-and-ethics/responsible-research-assessment”

Job Description

To conduct research within the field of Natural Language Processing and contribute to the overall research performance of the School and University, carrying out research leading to the publishing of high-quality research. To pursue excellence in research and to inspire others to do the same.

Research

• To conduct research within the field of Natural Language Processing and contribute to the overall research performance of the School and University by the production of measurable outputs including bidding for funding, publishing in national academic journals and conferences, and the recruitment and supervision of postgraduate research students. 
• To develop research objectives and proposals for own or joint research including research funding proposals
• To attend and or present at conferences/seminars at a local and national level as required
• To undertake administrative tasks associated with the research project, including the planning and organisation of the project and the implementation of procedures required to ensure accurate and timely reporting
• To prepare research ethics and research governance applications as appropriate
• To review and synthesise existing research literature within the field
• To participate in School research activities.
• To build and create networks both internally and externally to the university, to influence decisions,   explore future research requirements, and share research ideas for the benefit of research projects

• To engage effectively with industrial, commercial and public sector organisations, professional institutions, other academic institutions etc., regionally and nationally  to raise awareness of the School’s profile, to cultivate strategically valuable alliances, and to pursue opportunities for collaboration across a range of activities. These activities are expected to contribute to the School and the enhancement of its regional and national profile.

• To undergo personal and professional development that is appropriate to and which will enhance performance.
• To participate in School administration and activities to promote the School and its work to the wider University and the outside world

• Any other duties not included above, but consistent with the role.

Person Specification

Essential Criteria

Qualifications and Education

The field of Natural Language Processing (NLP) has made unprecedented progress over the last decade, but the extent to which NLP systems “understand” language is still remarkably limited. A key underlying problem is the need for a vast amount of world knowledge. In this project, we focus on conceptual knowledge, and more in particular on: 

(i) capturing what properties are associated with a given concept (e.g. lions are dangerous, boats can float); 

(ii) characterising how different concepts are related (e.g. brooms are used for cleaning, bees produce honey).

Our proposed approach relies on the fact that Wikipedia, and the language models that are trained on it, contain a wealth of such knowledge. Unfortunately, however, important properties and relationships are often not explicitly mentioned in text, especially if they follow straightforwardly, for a human reader, from other information (e.g. if X is an animal that can fly then X probably has wings). Language models similarly struggle with inferring knowledge that requires commonsense understanding. Apart from learning to extract conceptual knowledge, we thus also have to learn how to reason about such knowledge.

A central question is how conceptual knowledge should be represented and incorporated in language model architectures. Current NLP systems heavily rely on vector representations in which each concept is represented by a single vector. This approach has important theoretical limitations in terms of what knowledge can be captured, and it only allows for shallow forms of reasoning. In contrast, in symbolic AI, conceptual knowledge is typically represented using facts and rules. This enables powerful forms of reasoning, but symbolic representations are harder to learn and to use in neural networks.

The solution we propose relies on a novel hybrid representation framework, which combines the main advantages of vector representations with those of symbolic methods. In particular, we will explicitly represent properties and relationships, as in symbolic frameworks, but these properties and relations will be encoded as vectors. Each concept will thus be associated with several property vectors, while pairs of related concepts will be associated with one or more relation vectors. Our vectors will thus intuitively play the same role that facts play in symbolic frameworks, with associated neural network models then playing the role of rules.

Additional Information

Important: evidencing you match the job criteria

It is Cardiff University's policy to use the person specification as a key tool for shortlisting. Candidates should evidence that they meet ALL of the Essential Criteria as well as, where relevant, the Desirable.
As part of the application process you are asked to provide this evidence via a Supporting Statement.  Please ensure that the evidence you are providing corresponds with the numbered criteria outlined below.  Your application will be considered based on the information you provide under each criteria.
When attaching the supporting statement to your application profile, please ensure that you name it with the vacancy reference number:  12789BR_supporting_statement.

Supporting Statement

You need to demonstrate clearly how your experience, knowledge, skills and abilities match those required for the job, as detailed in the Person Specification. Where possible you should give examples of when you have used specific skills and abilities or needed to develop competencies in the relevant areas. You can draw on elements from any aspect of your life, such as education, work, home or community life, as long as you focus on its relevance to the needs of this job. If you do not demonstrate that you meet all of the ESSENTIAL criteria then you will not be invited to interview. You do not need to demonstrate that you meet all of the DESIRABLE criteria in order to be invited to interview.
Meeting all of the Essential Criteria does not guarantee you an interview however, as there may be other applicants who have demonstrated that they meet some or all of the desirable criteria, or that they meet the criteria to a greater depth and/or relevance. Essential criteria may be tested at interview stage.



The School

Cardiff University has a successful School of Computer Science & Informatics with an international reputation for its teaching and research activities. The School delivers a stimulating portfolio of BSc and MSc degrees, and offers research degrees leading to MPhil and PhD qualification. We have a strong research track record recognised for its outstanding impact in terms of reach and significance, with 79% of its outputs deemed world-leading or internationally excellent in the 2014 Research Excellence Framework. Our multi-million pound annual research income is funded by a variety of sources, including the Royal Society, the Leverhulme Trust, the Welsh Assembly (including Higher Education Funding Council for Wales), the UK Research Councils, including Engineering and Physical Sciences Research Council, the European Commission and industrial partners (mainly IBM UK Ltd and Airbus).
For further information visit: http://www.cs.cf.ac.uk/ 
.


We particularly welcome applications from women who are under-represented within Cardiff University in this field.

Job Category

Academic - Research