Speech technology is life-changing. It can literally give voice to someone who has lost theirs through illness. Or it can be misused for malign purposes like the creation of ‘deep fakes.’

The ADAPT Centre is at the forefront of designing mathematical algorithms to enhance speech communication between humans and technology. Professor Naomi Harte’s work is underpinned by signal processing and machine learning, but it is her multimodal speech analysis which is driving advances in the technology. Her work to fuse audio and visual cues may soon be applied in settings as varied as the kitchen, the boardroom and the aeroplane cockpit.

From developing systems to help college lecturers engage with remote learners, to facilitating more seamless interactions with our smart speakers, Professor Naomi Harte shares fascinating insights from her work.

TOPICS WE DISCUSSED INCLUDE

• How deep learning can improve our everyday interactions with smart speakers and assistants.

• The life-changing impact of speech synthesis, but how its misuse could be a societal harm.

• Why the analysis of birdsong is being used to improve speech technology.

• How The Adapt Centre will chair the world-renowned INTERSPEECH conference in Dublin in 2023.

GUEST DETAILS

Naomi is an Associate Professor in the School of Engineering in Trinity College. She is Co-PI and a founding member of the ADAPT SFI Centre.

In ADAPT, she has led a major Research Theme centred on Multimodal Interaction involving researchers from Universities across Ireland and was instrumental in developing the future vision for the Centre for 2021-2026. She is also a lead academic of the hugely successful Sigmedia Research Group in the School of Engineering.

Prior to returning to academia in 2008, she worked in high-tech start-ups in the field of DSP Systems Development, including her own company. She holds tech patents and licences through her collaboration with Google in the USA.

MORE INFORMATION

https://www.tcd.ie/research/profiles/?profile=nharte

https://scholar.google.com/citations?user=0GswUfcAAAAJ&hl=en

https://www.linkedin.com/in/prof-naomi-harte/

Email:[email protected]

YOUR HOST

Aideen Finnegan is an IMRO award-winning broadcast journalist with almost 18 years experience, including ten years across two national newsrooms. Find her on Twitter @AideenFinnegan or on Linkedin: https://www.linkedin.com/in/aideenfinnegan/

How do we create global standards for AI so we can mitigate against harms like disinformation? How do we prevent AI from perpetuating human biases?

Meet Professor Dave Lewis, who is at the forefront of developing international standards for trustworthy AI. The EU is the first major regulator anywhere in the world to be drafting legislation governing the use of AI, and The ADAPT Centre at Trinity College Dublin is right at the heart of the process.

Dave explains the challenges faced by society in governing AI, how his research is being used by industry to raise standards and how the use of algorithms can be balanced to meet the needs of users, organisations and society.

TOPICS WE DISCUSSED INCLUDE

● Challenges and complexities of governing AI

● The “mini-UNs” assembled to develop international standards

● Innovations that provide practical mechanisms for data governance

● Human interaction design considerations

● Opportunities for industry to collaborate with ADAPT

GUEST DETAILS

Dave Lewis is an Associate Professor at the School of Computer Science and Statistics at Trinity College Dublin and the head of its Artificial Intelligence Discipline.

He is Deputy Director of the ADAPT SFI Research Centre for Digital Content Technology. His research focuses on the use of open semantic models for Trustworthy AI and Data Governance.

He has led the development of international standards in AI-based linguistic processing of digital content at the W3C and OASIS.

Trinity College: http://people.tcd.ie/Profile?Username=delewis

Linkedin: https://www.linkedin.com/in/dave-lewis-13626a/

YOUR HOST

Aideen Finnegan is an IMRO award-winning broadcast journalist with almost 18 years experience, including ten years across two national newsrooms. Find her on Twitter @AideenFinnegan or on Linkedin: https://www.linkedin.com/in/aideenfinnegan/

Computer technologies are constantly advancing, remember when we thought landlines were as good as it could get? fast forward to today where smartphones are the standard for most. But how much further can we take things? Could we one day have computers who see and act like humans?

Joining us today is Prof. Rozenn Dahyot, an expert in the field of computer technologies. Rozenn explains to us the difficulties in teaching robots human functions such as sight. She speaks about the future of computer technologies, including the metaverse and where she sees this going and also speaks about the ethical limits to how far computer technologies should be taken.

This 34-minute discussion will leave you feeling very excited about the future and show you just how much can be achieved when a group of experts put their minds together.

TOPICS WE DISCUSSED INCLUDE

● Most recent computer technology research funding

● What is Digital Signal Processing

● Visual Data Analytics

● AI and visual detection

● Ethics surrounding computer vision in public

● What the future holds for computer vision

GUEST DETAILS

Prof. Dahyot is a Professor of Computer Science at Maynooth University (Ireland) and a PI in the SFI ADAPT Research Centre.

Her research interest is in pattern analysis and machine intelligence (machine learning, AI) with signal processing applications (e.g. remote sensing, computer vision, computer graphics), applied for instance for automatic asset mapping.

She has been the President of the Irish Pattern Recognition and Classification Society 2014-2020, and she is the general chair organizing the European Signal Processing Conference in 2021.

Maynooth: www.maynoothuniversity.ie/faculty-science-engineering/our-people/rozenn-dahyot

Social Media: https://ie.linkedin.com/in/rozenn-dahyot-a0807116

YOUR HOST

Marie Boran is a PhD candidate and third level lecturer in the area of science, technology & society. Her passion is emerging technologies and as a writer has been covering the technology sector for almost 15 years. She was Irish Science Writer of the Year in 2018 and Technology Reporter of the Year in 2016.

Linkedin - https://www.linkedin.com/in/mboran/

Alexa, Siri and robotic hoovers have become common household appliances. But how would you feel about a robot making your bed in the morning? Could AI robots be the future of healthcare?

TU lead John Kelleher believes it could be. Whether we know it or not, NLP and AI have already become part of our daily lives. From tumour detection to drug discovery research is now delving deeper into healthcare to see what more could be achieved. Have we reached our peak with AI and healthcare or could walking talking robots be in our future? If so, how would medical ethics work?

TOPICS WE DISCUSSED INCLUDE:

● What is NLP or Natural Language Processing

● Where is NLP used

● How does NLP impact our daily lives

● What does the future of NLP look like, will robots soon be a household appliance?

● AI and Healthcare

● How is AI currently being used in the fight against Covid 19

● The ethical concerns surrounding AI in Healthcare

● How can AI be made more sustainable

GUEST DETAILS

John Kelleher is the TU Dublin lead in the ADAPT Centre and the scientific lead for the Digital Content Transformation Strand. Within the ADAPT centre, John leads research projects on language modelling, lexical semantics, machine translation, novelty detection, image captioning, dialogue systems, and making AI more environmentally sustainable.

Adapt Centre - https://www.adaptcentre.ie/experts/john-d-kelleher/

LinkedIn - https://www.linkedin.com/in/johndkelleher

YOUR HOST

Marie Boran is a PhD candidate and third-level lecturer in the area of science, technology & society. Her passion is emerging technologies and as a writer has been covering the technology sector for almost 15 years. She was Irish Science Writer of the Year in 2018 and Technology Reporter of the Year in 2016.

Linkedin - https://www.linkedin.com/in/mboran/

How we represent ourselves online is hugely important. While in the past we used photos and avatars, today we are creating virtual humans.

What do they look like? Are they caricatures or more advanced moving video representations of real selves? Could they develop further with AI and operate independently without us?

Principal Investigator Rachel McDonnell is an expert in this area at the Adapt Centre and has been making amazing progress with virtual human technology. She explains what they are, how they work and the potential to develop in the future. She has a real passion for how our virtual humans live and interact with others online and shares some of the gigantic improvements of recent years.

TOPICS WE DISCUSSED INCLUDE

• What are virtual humans

• Where you have seen them before but never knew

• The pros and cons of virtual humans

• How we are combatting ‘uncanny valley theory’

• The acceptance and believability rates of virtual humans

• When will we start attending VR meetings

• The ethics of how far should we go with virtual humans

• Problems to overcome with Augmented Reality situations

• Why the future is now for the Adapt Team

• Your listener questions

GUEST DETAILS

Rachel McDonnell is an Associate Professor in Creative Technologies at the School of Computer Science and Statistics at Trinity College Dublin.

She is also a Principal Investigator in the ADAPT Research Centre and a member of the Graphics, Vision, and Visualisation Group.

Her research projects and interests include computer graphics, character animation, virtual humans, VR, and perception, with a focus on real-time performance capture and perception of virtual humans.

Trinity College - https://www.scss.tcd.ie/Rachel.McDonnell/index.shtml

Linkedin - https://www.linkedin.com/in/rachel-mcdonnell-10552b155/

YOUR HOST

Marie Boran is a PhD candidate and third level lecturer in the area of science, technology & society. Her passion is emerging technologies and as a writer has been covering the technology sector for almost 15 years. She was Irish Science Writer of the Year in 2018 and Technology Reporter of the Year in 2016.

Linkedin - https://www.linkedin.com/in/mboran/

ADAPT Researcher Mohammed Al-Rawi discussed the role AI plays in the fashion industry. It is used at almost every level, from manufacturing, to transport, and to marketing. You can access the accompanying slideshow presentation here: http://ow.ly/qbdl50BxNTM

ADAPT Researcher Helen McHugh was featured on Raidió na Life on a programme called Blas na Cathrach hosted by Joe Breathnach. She discusses eTranslation, which is a machine translation service provided by the European Commission to all public servants and all start-ups and small to medium enterprises across the EU, completely free of charge.

She also goes into her eTranslation work with ADAPT which focuses on Irish language and sourcing English to Irish translations form bodies such as Government Departments, County and City Councils, and other public bodies such as Conradh na Gaeilge, An Coimisinéir Teanga, Foras na Gaeilge, universities to grow the samples size to improve the eTranslations over time.

Dr Jamie McGann and Prof Johann Issartel have recognised the crisis of sedentary habits among children. A mix of reduced outdoors space, safety concerns, and technology have reduced the amount of movement in children's play and increased screen time. The two are developing MoveAhead which uses kid's bodies as controllers for video games.

Eamonn Donlyn explores the big ideas behind Darwin and Goliath, a micro-recommender system currently used for eCommerce. The Amazons and Netlifxes of the world have the capital for elaborate recommendation systems that are custom-made for their products, but smaller sellers don't have that luxury. The Darwin and Goliath system seeks to democratise the machine learning tech and expand it to uses beyond eCommerce.

Liam Cronin, of the RDI Hub in Killorgin, County Kerry, discusses their aims as a not-for-profit collaborative workspace to foster growth and creativity among budding start-ups. He discusses his own journey that has lead to RDI partnering with the ADAPT Centre and Microsoft to put on the inaugural 'John McCarthy AI Summer School'.

Learn more about the event and register to hear from experts on sustainable AI here: https://www.eventbrite.ie/e/the-international-john-mccarthy-ai-summer-school-tickets-116201069527