Prof Raymond Bond believes human well being and wellbeing needs to be central in AI growth.
“I’ve loads to be taught,” Prof Raymond Bond says when requested about selling his work. This understanding that there’s all the time extra to be taught and enthusiasm for steady studying has little question contributed to Bond’s success in his subject.
Bond is a professor in Human Pc Methods at Ulster College. His PhD was in computerised electrocardiology. “This concerned processing and visualising ECG information,” he explains. “The ECG is without doubt one of the mostly used diagnostic checks to assist detect heart-related issues comparable to arrythmias and coronary heart assaults.”
In the course of the PhD, Bond developed instruments to visualise ECG information, together with a pc simulator to reveal what can occur if electrodes are positioned incorrectly.
“To be trustworthy, the PhD helped spotlight simply how little I knew and know, and it supplied nice alternatives to satisfy worldwide researchers from many alternative backgrounds, together with analysis leaders within the medical and engineering fields. It was a humbling expertise to say the least,” Bond says.
After his PhD, he labored as a analysis assistant then a educating fellow earlier than taking on a lectureship in pc science and turning into a senior lecturer in information analytics on the Cognitive Analytics Analysis Laboratory (CARL) at Ulster. His curiosity in optimising the interactions between folks and expertise led to the professorship in human pc programs.
“I feel we should be humanity-centered when designing and deploying AI applied sciences,” he says.
“Maybe we have to remind ourselves of what it means to protect human values, wellbeing and happiness, and repeatedly ask ourselves whether or not an AI expertise might negatively have an effect on our wellbeing and happiness in each the brief time period and long run.
“In fact, it is likely to be difficult to foretell the long-term results of AI applied sciences however maybe we will not less than attempt.”
Inform us about your present analysis.
I’m at present engaged on various initiatives about AI ethics, together with learning the position of AI in radiography with reference to the consequences of AI assist in scientific decision-making. For instance, how can we mitigate automation bias? Automation bias is the place an individual could repeatedly over belief and naively settle for AI suggestions. May automation bias outcome within the lack of competency in a talent comparable to studying X-rays or ECGs? We even have a challenge wanting on the high quality of digital well being apps (utilizing information science) in addition to different initiatives in digital psychological well being and AI in cardiology. For instance, one in every of our papers makes use of information analytics to indicate that the consumer scores and the variety of downloads of a well being app doesn’t correlate with the precise high quality of a well being app.
I’m additionally working with others on the usability engineering of automated exterior defibrillators. The usability of this medical machine is vital on condition that customers (members of the general public/lay rescuers) could also be interacting with this consumer interface for the primary time in a probably lifesaving state of affairs.
In one other challenge, we’re learning the position of biodynamic indoor lighting to boost the wellbeing of individuals dwelling with dementia. Mild is essential to our circadian rhythms and for managing our sleep patterns.
We additionally not too long ago acquired awarded funding by the UK Engineering and Bodily Sciences Analysis Council (EPRSC) for a big centre for doctoral coaching in digital well being applied sciences which is able to present nice alternatives for brand new PhD researchers at Ulster College and College School London.
In your opinion, why is your analysis essential?
I feel our digital well being analysis is essential as a result of having and selling good well being and wellbeing is clearly core to society, and we’d like extra analysis to determine the optimistic and unfavourable results that digital applied sciences can have on our well being and wellbeing.
For instance, prime quality digital well being interventions and apps might assist forestall illness in addition to serving to folks handle their circumstances (eg adhering and interesting with therapy plans and so forth).
Well being apps and wearables will also be used to gather well being and wellbeing information ‘in the actual world’, eg symptom information will be collected as repeated measures which may then be shared with healthcare employees to probably improve their scientific decision-making by having larger decision private well being information. In fact, time-series symptom information will also be analysed in actual time by AI algorithms to maybe act as early warning programs and to determine key patterns/tendencies.
Raymond Bond
I’ve a selected curiosity in figuring out strategies for optimising the collaboration between people and AI. And we have to uncover and invent methods to make sure that AI helps and never hindering scientific competence and scientific choices. We have to create methods to calibrate how a lot people ought to belief AI and maybe even how a lot AI ought to belief people (who is aware of). We’ve got created an e-learning course – Introduction to AI for healthcare professionals – which can additionally assist with selling AI literacy amongst healthcare employees.
On the subject of affect, I’m hoping to see research-informed product innovation within the well being and medical sectors together with digital psychological well being improvements by way of collaborations with organisations comparable to Canary Speech, Encourage Wellbeing, Motion Psychological Well being and Pneuma Healthcare, in addition to a brand new biodynamic lighting machine with an organization known as SkyJoy. Hopefully there will likely be new discoveries/insights to tell the design and testing of automated exterior defibrillators at Stryker (HeartSine). I additionally hope to see affect with PulseAI, with reference to the deployment of recent AI algorithms for analysing electrocardiograms.
What impressed you to grow to be a researcher?
Unsure what precisely impressed me, however I feel it was the chance for lifelong studying. I wish to be taught; nonetheless, I’m not essentially nice at it. I’ve realised that the easiest way to be taught (for me not less than) is by educating and doing analysis. I discover that educating helps me construction info in a means that permits me to be taught the core ideas in the appropriate order. Analysis can also be nice since you get to mentor, current work and write. Presenting and writing does appear to assist me ‘assume’ about issues extra deeply. The suitable mixture of phrases can actually assist alter my perspective on a subject or set of concepts. I additionally naturally take pleasure in discussing and producing concepts.
I’m fortunate sufficient to be in a subject that’s broad which permits me to work in lots of utility areas in healthcare. I actually like working throughout disciplines and dealing in groups (once more, you get to be taught).
Science for me appears to be about questions and solutions. And analysing datasets to search out these solutions is inspiring. I additionally really feel impressed after I use information science/machine studying to find new information a few subject. There may be little higher than discovering and inventing.
What are among the largest challenges or misconceptions you face as a researcher in your subject?
Typical challenges embody planning my time to put aside hours to write down and/or lead analysis initiatives, for instance, to hunt additional funding. I might additionally like to organise my diary higher so I might write extra perspective or place papers on key analysis tendencies/challenges in digital healthcare (once more, ‘writing is pondering’, for me not less than). I need to say that after I first got here throughout the Eisenhower matrix, it helped me assume extra concerning the duties which might be actually essential however not pressing. I must keep away from neglecting these ‘essential however not pressing’ duties.
Do you assume public engagement with science and information has modified in recent times?
I’ve not performed any evaluation to see if there was any important adjustments within the public engagement with science. Maybe given the pandemic, extra residents might need engaged extra with subjects comparable to exponential development, statistics, graphs and modelling, which might have probably contributed to enhancing statistical literacy. Readers is likely to be concerned about a paper I co-authored titled Why Pandemics and Local weather Change Are Exhausting to Perceive and Make Resolution-Making Troublesome.
How do you encourage engagement with your individual work?
I attempt to work with stakeholders and corporations when doing analysis. That is primarily to work in the direction of having a real-world affect on merchandise and stakeholders. Collaborating with corporations jogs my memory of the facility of the Ikea impact which means that we will maybe attribute numerous worth to the issues that we construct ourselves utilizing our personal time.
I attempt to encourage press releases about our work; nonetheless, I want to interact extra with requirements and policymakers.
Networking and giving talks at totally different occasions has additionally helped promote our work and I’ve tried to combine my analysis into my educating. For instance, it has been actually nice to see MSc college students go on to publish their work in analysis boards. I used to be additionally concerned in making a research-teaching nexus toolkit to assist promote the mixing of analysis into educating throughout the college.
Learn how rising tech tendencies are reworking tomorrow with our new podcast, Future Human: The Collection. Pay attention now on Spotify, on Apple or wherever you get your podcasts.
