Dr. Amelia Kelly
Nominated Award: AI Person of the Year
Linkedin profile of Person: https://www.linkedin.com/in/amelia-kelly/
Dr. Amelia Kelly is an artificial intelligence engineer and scientist specializing in speech recognition of children’s voices. She is currently the Vice President of Speech Technology at Dublin-based SoapBox Labs and has more than a decade of experience in speech signal processing, natural language processing, machine learning, and AI.
After finishing her PhD in Linguistics and Speech Technology at Trinity College Dublin, Amelia went to Silicon Valley to work for Fluential, a startup providing voice UI for mobile apps. It was there that she dug into the practical aspects of speech recognition, NLP, and intent classification and found that she enjoyed working in a startup environment.
Amelia then worked for IBM Watson as a Cognitive Computing Solution Engineer. At IBM, she contributed to the very first iterations of what is now the Watson AI assistant. She also gained invaluable customer experience in the US and insight into how a product goes from conception to being incorporated into some of the most widely used banking, insurance, and health systems in the world.
In 2015, Amelia joined SoapBox Labs as one of our first three employees to help build our proprietary speech recognition technology for kids from the bottom up. Fast forward seven years and she’s now the VP of Speech Technology, overseeing a team of world-class computational linguists, speech engineers, and scientists from eight different nationalities, all of whom have chosen to work for SoapBox for our groundbreaking work in speech recognition for kids and to benefit and learn from Amelia’s leadership, knowledge, and expertise in the industry.
When Amelia began working at SoapBox, her primary role as the speech engineer was to write the code to build — from scratch — our models and speech systems. Over time, as the company grew, she became more involved in the big-picture planning, creating “grand visions” of what we wanted to build, and managing the growing number of speech engineers who were implementing it. She also became more customer focused, and today, she speaks to companies all over the world about SoapBox’s voice engine, why it’s so accurate and trustworthy, and how it can help clients maximize the potential of voice technology to deliver immersive and fun voice-enabled learning and play experiences for children.
Reason for Nomination
The primary application of SoapBox’s speech technology is voice-enabling PreK-12 learning and assessment tools. In an educational setting, speech recognition systems must be highly accurate and understand all kids’ voices, regardless of their age, accent, dialect, or socioeconomic background to ensure equitable learning experiences. Anything less, and kids experience false negatives, which are frustrating, or false positives, which can negatively impact their learning journey.
For example, if a biased speech system fails to understand a child’s accent or dialect, it can consistently tell that child he or she is a poor reader when, in fact, they are reading correctly. An unbiased system, on the other hand, can offer fair and uncompromised information to facilitate edtech platforms and services.
Amelia has been instrumental in driving our culture and approach to equity in how we build, model, and test our speech technology. Under her leadership, our voice engine has been built with equity at the heart of its design and development to mitigate bias across accents and dialects. We’ve received independent validation from researchers and clients that our voice engine shows no bias across race, accents, and socio-economic backgrounds.
This summer, Amelia spearheaded our submission for Digital Promise’s Prioritizing Racial Equity in AI Design certification, which resulted in us becoming the first and only AI company to date to earn this certification. It recognizes that our voice engine prioritizes racial equity by proactively identifying and minimizing racial bias in our algorithms and design and that we’re transparent about our design processes and the actions we take to identify and mitigate racial bias. (The attached PDF is a redacted version of our submission to Digital Promise, highlighting the strategies and practices Amelia and our Speech Technology and Engineering teams employ to mitigate racial bias in our datasets and models.)
Amelia shares her passion for and thought leadership on equitable speech technology with our clients and the tech and AI community by writing articles, presenting at company webinars, and speaking at industry conferences, including the following:
• Blog: Debiasing (https://buff.ly/2So52v1)
• Glossary: Speech recognition for education (https://buff.ly/3AN6qLD)
• Webinar: Why kids will change how with think about voice AI (https://buff.ly/3m4wDgU)
• SXSW EDU 2022: What my kid taught me about edtech during COVID (https://buff.ly/3eq3onR)
Amelia was also instrumental in driving our successful 2019 application for €12M in funding under the Disruptive Technologies Innovation Fund (DTIF). In conjunction with consortium partners XPERI and the National University of Ireland, Galway (NUIG), this funding has allowed SoapBox Labs to embark on a new and truly disruptive AI project: Data-center Audio/Visual Intelligence on-Device (DAVID).
DAVID is an on-device multi-modal AI platform, encompassing both voice and vision, that offers the user complete privacy while driving deeper and more natural interactions with smart toys, games, AR/VR, and smart home devices.
Under Amelia’s leadership, the DTIF project has:
• Enhanced SoapBox’s credibility as a leader in speech technology and AI in Ireland and globally.
• Unlocked a new market for SoapBox in the toy space based on the disruptive nature of the project research.
Without a doubt, Amelia is a formative and essential member of SoapBox, and we’re all the better for her passion, skillset, and expertise in speech technology and AI that she brings to the team every day.
Additional Information:
Below, we are delighted to list some of Amelia’s achievements and contributions to AI outside of SoapBox (with links for more information where applicable):
• Amelia won the 2022 Women in Voice “Technical Excellence” Award: https://buff.ly/3BgI25p
• In 2021, she was recognized by the Irish Independent as one of Ireland’s top 30 women in Tech: https://buff.ly/2QLOIDH
• Amelia is on the advisory panel for the University of Limerick’s Masters in AI program.
• She was awarded a Fulbright Tech Impact scholarship in March 2021 and spent a few months in Boulder, Colorado collaborating with a university group at the Institute of Cognitive Studies on how they could use child speech recognition to improve learning outcomes for children in math and science classes.
• She is on the National Standards Authority of Ireland’s (NSAI’s) team to establish standards in AI.
• Amelia was the keynote speaker at the 2021 Irish Systems and Signals Conference (ISSC).