Job Description: Research Assistant - Machine Learning College/School: Faculty of Humanities and Social Sciences Department/Subject: Applied Linguistics / Computer Sciences Salary: £30,046 to £33,797 per annum together with USS pension benefits Hours of work: Full-time - 35 hours per week Contract: This is a fixed term position until 31 st October 2022 Location: This position will be based at the Singleton/Bay Campus with extended stay periods at LIS Lab in Toulon University (France) An exciting opportunity has arisen for a research assistant in AI and natural language processing (NLP). We are looking for individuals with skills and experience in the analysis of text using deep learning and a general background in machine learning to work on online chat conversations for detecting online grooming. Our team is multidisciplinary, and we aim to integrate linguistic knowledge into the design of an NLP neural network, to improve its robustness and interpretability. This post is funded by the project DRAGON-S (Developing Resistance Against Grooming Online - Spot and Shield), - see Background Information section. The role-holder will work closely with Dr Paiement (AI) and Professor Lorenzo-Dus (Linguistics), and with the remaining project team, comprising two other researchers in Linguistics, a (web) developer, a project officer and a project risk officer. Introduction The ideal candidate for this position is someone with substantial research experience, at PhD level or beyond. We seek someone who is competent and comfortable with state-of-the-art deep neural networks for NLP and text analysis, including XLNet. Preference will be given to candidates who have experience with chatbots and with other machine learning methods, such as non-linear dimensionality reduction and Markovian models. The successful candidate will join the DRAGON-S team for the duration of the contract, taking part in weekly meetings, and in seminars and conferences at the team's discretion. They will travel between Swansea and Toulon as required by the progress of the project, with a substantial part of the AI work being carried out in Toulon. They will have the opportunity to carry out individual research, network with project stakeholders and end user groups in Wales, France, and internationally and, depending on their career stage, to benefit from mentoring and collaborative research work. Background information Project DRAGON-S (Developing Resistance Against Grooming Online - Spot and Shield) will offer tools based on integrating AI/Linguistics that enable law enforcement to spot online grooming content in real-time. The project will impart specialist knowledge through a learning portal and chatbot to strengthen professionals' abilities to shield children from online grooming. Project DRAGON-S is funded by the UNICEF End to Violence programme - https://www.end- violence.org/grants/swansea-university. Main Duties 1. Using preliminary results (obtained from linguistic analysis), to work with the project team to improve a deep learning model (XLNet) for detection of online grooming. A few methodology candidates are to be compared, including an extension using an HMM. 2. To work with the project team to develop a chatbot using an existing and the newly extended XLNet-based language model. 3. To alpha and beta test all AI based methods in the project, and to bug fix; and to add new functionalities based on results of usability tests with stakeholders. 4. To pro-actively contribute to and conduct research, including gather, prepare and analyse data, generate original ideas and present results. 5. To prepare reports, draft patents and papers describing the results of the research, both confidential and for publication. 6. To be self-motivated, apply and use their initiative, aiming to determine suitable ways to tackle challenges and seeking guidance when needed 7. To interact positively and professionally with other collaborators and partners within the Faculty and elsewhere in the Universities of Swansea and Toulon and beyond as appropriate, especially with project's stakeholders and end-user groups in Wales and internationally. 8. To contribute to the project's organisational matters in order to help it run smoothly and to help raise its external research profile. 9. To keep informed of developments in the field in technical, specific and general terms and their wider implication for the discipline area, commercial applications and the knowledge economy. 10. When requested, to act as a representative or member of committees, using the opportunity to extend their own professional experience. 11. To demonstrate and evidence own professional development, identifying development needs with reference to the Vitae Researcher Development Framework, particularly with regard to probation, PDR and participation in training events. 12. To maintain and enhance links with the professional institutions and other related bodies. 13. To observe best-practice protocols in maintenance and retention of research records as indicated by HEI and Research Councils records management guidance. This includes ensuring project log-book records are deposited with the University/Principal and Co-Investigators on completion of the work. 14. To promote equality and diversity in working practices and maintain positive working relationships 15. To conduct the job role and all activities in accordance with safety, health and sustainability policies and management systems, in order to reduce risks and impacts arising from the work activity 16. To ensure that risk management is an integral part of any decision making process, by ensuring compliance with the University's Risk Management Policy. Essential criteria: 1. A PhD Degree in Machine Learning, AI, Data Science, or equivalent 2. Experience with NLP deep learning models, especially transformers and XLNet 3. Experience with text analysis, preferably chat logs 4. Evidence of the ability to actively engage in, and contribute to, writing and publishing research papers, particularly for refereed journals. 5. A demonstrable ability to conduct research in line with the objectives of the project 6. Evidence of planning skills to contribute to the research project 7. A commitment to continuous professional development Desirable Criteria 1. Experience with designing chatbots 2. Experience with using non-linear dimensionality reduction and HMM algorithms. 3. Experience with analysis of digital texts (e.g. chat logs) Informal enquiries: Professor Nuria Lorenzo-Dus (project related) N.Lorenzo-Dus@Swansea.ac.uk and Dr Adeline Paiement (AI related) Additional Information Shortlisting Date: TBC Interview Date: TBC The University is committed to supporting and promoting equality and diversity in all of its practices and activities. We aim to establish an inclusive environment and welcome diverse applications from the following protected characteristics: age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race (including colour, nationality, ethnic and national origin), religion or belief, sex, sexual orientation. A satisfactory DBS certificate must be provided before a start date can be confirmed