My background is primarily in artificial intelligence, which I have been working with for over 30 years. Since the late 90s I have worked on a wide range of innovative projects which have been instrumental in driving the transformation of Elsevier from an established scientific print publisher through to a digital information analytics company.
In 2009, I was responsible for developing Elsevier’s first semantic databases for therapeutic drug design, and in 2014 as education director in our health solutions division I introduced new digital business models and launched adaptive learning courses into the NHS in the UK.
Since taking up my current role in 2016, my work has focused on making these information analytics tools truly productive for customers in commercial R&D, and now my work is focusing on the stewardship of data resources, and the validation and auditing of machine learning.
I recently launched Elsevier’s AI and demantic data platform (Entellect.com) and organised a “datathon challenge” to investigate repurposing drugs to treat rare diseases. This project used the new platform in collaboration with the pharma pre-competitive group the Pistoia Alliance, along with not-for-profit organisations CuresWithinReach and Mission-Cure. I am now extending this work to cancer research and sustainable chemistry.
Throughout my career, it has always been important to have a positive impact on society. I recently had responsibility for the launch of Elsevier’s AI and Semantic Data Platform (Entellect.com), which is a culmination of the last couple of decades work at Elsevier in refining science information analytics. The launch of the platform was through a Datathon Challenge, and it is great to see people coming together to work on problems in the environment; and seeing positive life enhancing outcomes as a result. We hope to see benefits brought to patients suffering from chronic pancreatitis as a result of the work.
It is vital to get a comprehensive understanding of the philosophical underpinnings of AI, as well as the engineering principles. I am grateful to Professor Margaret Boden for the broad foundation in AI she gave me as an undergraduate at Sussex University, and I would recommend reading her new book.
My prediction last year was to see AI techniques becoming productive, and this has certainly been the case. At Elsevier, I see this with both the corporate and research customers we support and with the developments our operations teams are driving in analytical technology, from creating deep learning-ready data-sets to developing intuitive user experiences that make gaining insights more accessible. In 2020, Elsevier is publishing a free-report on AI ethics: “Ethics in Artificial Intelligence: Thoughts from the research community”, which gives some great insights into this growing area of focus for the data and AI industry and research area.
In 2020 we will see continued importance placed on data stewardship and the ethical use of AI, which is a key theme for the UK Government Office for AI, and the sector deal. This places great importance on avoiding or managing bias in machine learning models; and emphasizes the importance of data quality for AI. Part of these developments will centre on the use of the FAIR data protocols for data aggregation and integration to support use of AI for R&D uses. Supporting these developments in data stewardship is something we at Elsevier are focused on.
The big opportunity in the areas of science and industrial R&D is that high quality data and predictive analytics technologies (ie AI in the form of machine learning) is to put new tools in the hands of researchers to enhance the work they are doing in generating new knowledge and insights.
These can bring meaningful change in our world, delivering innovation to protect the environment, fight disease, and bring prosperity across our global society. I am pleased to think that data and analytics can support innovation to address these challenges and make the world a better place.
Our customers are looking to bring AI based predictions into their R&D workflows, the challenge is getting the quality AI ready data in place to deliver these predictions reliably. Elsevier plays an important role as a partner who can give trusted outcomes in all analytics workflow steps, including data identification and integration; querying semantic data; generating AI ready features; and running machine learning to make trained artificial neural networks.
This ensures we are putting reliable answers to key scientific questions in the hands of R&D teams, often novel answers not appearing anywhere before in the scientific literature.