AI Tools for Selection of Government Records Researcher

Tender description

Summary of the work A review of the whole-of-market for AI tools to aid selection of born-digital records and provide a written report on the findings including a recommendation to test between two and four tools. TNA will also use them to understand the content of records to provide more context for researchers. Specialist role Data scientist Expected Contract Length 1 month Latest start date Tuesday 1 October 2019 Maximum Day Rate £750 per day + VAT Early Market Engagement None Who Speclialist Work With You will need to work collaboratively with the project team at The National Archives. The project will be led by the Service Owner for Digital Preservation, working with data scientists, digital archivists, information managers and a project manager. What Specialists Work On The National Archives is responsible for providing advice to government departments on their selection of records for transfer to the archive (http://www.nationalarchives.gov.uk/information-management/legislation/public-records-act/public-records-system/). One of the biggest challenges for departments is the selection of born-digital records. The huge volumes of these records on shared drives and in EDRMS means that the manual methods used for the selection of paper records are not feasible. The researcher will survey the whole of the AI landscape to identify the AI tools and approaches that could be used to identify records suitable for the selection of born-digital records for transfer to The National Archives. Skills & Experience • Have an understanding of AI and specifically machine learning tools • Have an understanding and experience of evaluation of technical products • Have experience of report writing on technical products or services • Be able to work to a fixed timeline and budget Nice to Haves Have knowledge and experience of digital information management Work Location The AI for Selection Project team is based at The National Archives, Bessant Drive, Richmond, Surrey TW9 4DU. Working Arrangments Ideally co-located (for at least 3 days a week) so you can refine requirements with the team as you carry out the research. Normal working hours are 9am to 5pm, there is some flexibility available. Security Clearance None required, unless you co-locate at The National Archives. In this instance, baseline security clearance will be required. Additional T&Cs None No. of Specialists to Evaluate 5 Cultural Fit Criteria • Be willing to work at The National Archives • Be able to work alone to deliver work to a deadline • Be open to sharing findings and checking if on track • Be able to provide evidence when recommending (or otherwise) tools and approaches • Provide evidence of working with non-technical team members Evaluation Weighting Technical competence 70% Cultural fit 5% Price 25% Questions from Suppliers 1. Is the contract likely to extend beyond the stated 1 month? It could be extended to 6 weeks but we would need at least one tool recommended by the third week. 2. Can the Authority confirm if they are working with an incumbent supplier at present ? We are not working with anyone currently. 3. Could you please clarify what is meant by "co-located"? E.g. will the specialist be required in multiple locations as there is only one location currently stated? Bu "co-located" we mean working with us at The National Archives' (sole) premises in Kew. Please see the section 'Working arrangements " for additional information. 4. What tools and technologies will be employed in this role? i.e. R, Python, SQL etc… None. This is a research task, identifying tools and approaches and assessing them for their relevance to selecting digital records against various criteria. We are not expecting a new tool to be built as a result of this. 5. Is the data mostly text based? Will experience in Natural Language Processing (NLP) be necessary or beneficial? Yes, mostly (if not all) text based so experience in NLP will be an advantage. 6. Is travelling and sustenance included in stated maximum day rate of £750? Yes, travel and food is included in the £750 (there is no extra allowance for those). 7. Will the work be carried out on TNA’s own IT infrastructure or will the data scientist need to provide their own IT equipment, i.e laptops? The appointed person will need their own laptop as they will not have access to the corporate network. We will give them some open files if required. 8. Considering this is a research only project, is the academic knowledge more important than the hands-on technical expertise? The researcher does not need to be a machine learning engineer but knowledge of machine learning would be beneficial in order to assess the viability of the tools. 9. It is mentioned that “knowledge and experience of digital information management” would be beneficial. Could you state some of the tools and techniques that TNS uses at a high level? Would familiarity of these tools and techniques be beneficial? Familiarity with records management practices in government would be beneficial in order to understand how records managers might approach their use of AI tools. There is guidance on TNA's website (see link on the Digital Marketplace) and also: http://www.nationalarchives.gov.uk/information-management/manage-information/ in particular the section on Digital Transfer. 10. Is each data scientist expected to produce two to four tools individually, or as part of a team? What will be the level of teamwork? We do not require any tools to be built. Instead we want a review of the whole of the market and recommendations for between 2 and 4 tools to test. The appointed person will work with members of our team during this project, though resource is limited and the appointed person should be able to work with little or no supervision. 11. Will the research cover open source as well as off-the-shelf proprietary solutions? Yes, we prefer open source tools but we want to look at as wide a range of tools as possible.