To give AI-focused women academics and others their well-deserved – and overdue – time within the highlight, TechCrunch is launching a series of interviews specializing in notable women who’ve contributed to the AI revolution. As the AI boom continues, we shall be publishing posts all year long highlighting vital work that usually goes unrecognized. You can find more profiles here.
Emilia Gómez is a senior researcher on the European Commission's Joint Research Center and scientific coordinator of AI Watch, the EU initiative to watch the progress, adoption and impact of AI in Europe. Her team contributes scientific and technical knowledge to the EC's AI policy, including the recently proposed AI law.
Gómez's research relies in the sphere of computer music, where she contributes to understanding the way in which people describe music and the methods used to model it digitally. Starting from the sphere of music, Gómez examines the impact of AI on human behavior – particularly the impact on jobs, decisions, and youngsters's cognitive and socio-emotional development.
questions and answers
In short, how did you start with AI? What attracted you to this field?
I started my research in AI, specifically machine learning, as a developer of algorithms to routinely describe music audio signals when it comes to melody, tonality, similarity, style or emotion, utilized in various applications from music platforms to education. I began researching design novel machine learning approaches that address various computational tasks within the music domain and the relevance of the information pipeline, including dataset creation and annotation. What I liked about machine learning back then was its modeling capabilities and the shift from knowledge-based to data-driven algorithm design – for instance, as a substitute of designing descriptors based on our knowledge of acoustics and music, we now used our expertise to design datasets, architectures in addition to training – and evaluation procedures.
Through my experience as a machine learning researcher and observing my algorithms “in motion” in various areas, from music platforms to symphony live shows, I spotted the big impact these algorithms have on people (e.g. listeners, musicians), and I directed my research toward AI evaluation somewhat than development, particularly examining the impact of AI on human behavior and evaluating systems for elements equivalent to fairness, human oversight, or transparency. This is the present research topic of my team on the Joint Research Center.
What work are you most pleased with (within the AI space)?
On the educational and technical side, I’m pleased with my contributions to music-specific machine learning architectures on the Music Technology Group in Barcelona, which have advanced the state-of-the-art in the sphere, as reflected in my citation record. For example, during my PhD thesis, I proposed a data-driven algorithm to extract tonality from audio signals (e.g. if a chunk of music is in C major or D minor), which has change into, and later have, a vital reference in the sphere I co-designed a machine learning methods to routinely describe music signals when it comes to melody (e.g. to look for songs by humming), tempo or to model emotions in music. Most of those algorithms are currently integrated into Essentia, an open source library for audio and music evaluation, description and synthesis, and have been utilized in many suggestion systems.
I’m particularly pleased with Banda Sonora Vital (LifeSoundTrack), a Red Cross Humanitarian Technologies Award-winning project where we developed a personalised music recommender adapted to elderly Alzheimer's patients. There can be PHENICX, a big European Union (EU)-funded project on the usage of music that I coordinated; and AI to create enriching symphonic music experiences.
I like the music computing community and was joyful to change into the primary president of the International Society for Music Information Retrieval, to which I even have contributed throughout my profession, with a selected interest in increasing diversity in the sphere.
Currently, in my role on the Commission, which I joined as a senior scientist in 2018, I provide scientific and technical support to AI policies developed within the EU, particularly the AI Law. Because of this recent work, which is less visible when it comes to publications, I’m pleased with my modest technical contributions to AI law – I say “modest” because you may imagine that there are plenty of people involved here! For example, I even have contributed quite a bit to the harmonization or translation between legal and technical terms (e.g. by proposing definitions based on existing literature) and to assessing the sensible implementation of legal requirements, equivalent to: B. Transparency or technical documentation for civil engineering. Risk AI systems, general-purpose AI models and generative AI.
I'm also quite pleased with my team's work in supporting the EU AI Liability Directive, where we examined, amongst other things, particular characteristics that make AI systems inherently dangerous, equivalent to lack of causality, opacity, unpredictability or their self- and continuous learning skills and assessed the associated difficulties in proving causality.
How do you overcome the challenges of the male-dominated technology industry and due to this fact also the male-dominated AI industry?
It's not nearly technology – I also work in a male-dominated AI research and policy field! I don't have a method or strategy as that is the one environment I do know. I don't know what it might be prefer to work in a various or female-dominated work environment. “Wouldn’t or not it’s nice?”, because the Beach Boys song goes. I truthfully attempt to avoid frustration and revel in this difficult scenario of working in a world dominated by very assertive men and revel in working with excellent women on this field.
What advice would you give to women wanting to enter the AI field?
I might tell them two things:
You are urgently needed – please enter our space as there’s an urgent need for diversity of visions, approaches and concepts. For example, based on the divinAI project – a project I co-founded to watch diversity within the AI field – only 23% of creator names on the International Conference on Machine Learning and 29% on the International Joint Conference on AI in 2023 were female, independent of their gender identity.
You should not alone – there are lots of women, non-binary colleagues and male allies on this field, despite the fact that we will not be as visible or recognized. Look for them and get their mentorship and support! In this context, there are lots of reference groups within the research field. For example, after I became president of the International Society for Music Information Retrieval, I used to be very lively within the Women in Music Information Retrieval initiative, a pioneer in diversity efforts in music computing with a really successful mentoring program.
What are a number of the most pressing issues facing AI because it continues to evolve?
In my opinion, researchers should devote as much effort to AI development as they do to AI evaluation, as there’s currently a scarcity of balance. The research community is so busy advancing the state-of-the-art in AI capabilities and performance, and so desperate to deploy its algorithms in the actual world, that it forgets to conduct proper assessments, impact assessments, and external audits. The more intelligent AI systems are, the more intelligent their evaluations ought to be. The field of AI evaluation is poorly researched and that is the reason behind many incidents that give AI a nasty name, equivalent to: E.g. gender or racial biases in data sets or algorithms.
What issues should AI users concentrate on?
Citizens using AI-powered tools like chatbots should know that AI will not be magic. Artificial intelligence is a product of human intelligence. You should learn the operating principles and limitations of AI algorithms with a purpose to give you the chance to challenge them and use them responsibly. It can be vital that residents are informed concerning the quality of AI products, how they’re evaluated or certified, in order that they know which products they’ll trust.
What is the very best approach to construct AI responsibly?
In my view, the very best approach to develop AI products (with good social and environmental impacts and in a responsible manner) is to devote the crucial resources to assessment, social impact assessment and risk mitigation – for instance to fundamental rights. before you bring an AI system to market. This advantages corporations and trust in products, but in addition society.
Responsible AI or trustworthy AI is a approach to develop algorithms that require consideration of elements equivalent to transparency, fairness, human oversight, or social and environmental well-being from the start of the AI design process. In this sense, the AI Law not only sets the bar for the worldwide regulation of artificial intelligence, but in addition reflects the European give attention to trustworthiness and transparency – enabling innovation while protecting residents' rights. I imagine this can increase residents' confidence within the product and technology.
