SI MEDIA – INTERVIEW ON YES!AI
Davide Sarraggiotto talks about SI Media’s progress with AI
The world of broadcasting is on the verge of a transformation fueled by artificial intelligence. Machine learning is poised to transform the way broadcasters not only create content, but also manage and deliver it to audiences. Traditionally, broadcasters have relied on manual processes and human expertise to manage vast libraries of video, audio, and text. However, machine learning offers a glance into a new reality where intelligent automation takes a meaningful role.
Today we talk to Davide Sarraggiotto, SI Media Product Manager.
Hello Davide! Tell us about yourself and how you started your journey at SI Media.
“My name is Davide Sarraggiotto and I am currently working in SI Media on AI/ML applications. My journey here began in January 2024, right after graduating from the University of Padua with a Master Scientist in Computer Engineering. I had a deep-seated desire to put into practice all the knowledge I had accumulated in the fields of artificial intelligence, machine learning, and deep learning. It was important to me not only to apply what I had learned, but also to keep expanding my understanding and to stay updated with the latest developments in these rapidly evolving domains. Fortunately, SI Media has offered me exactly that, an opportunity to immerse myself in a stimulating and dynamic environment where I can work on the development of concrete and innovative AI solutions.”
How did SI Media get started with AI?
“SI Media has taken its first steps in AI by focusing on the development of much-needed intelligent algorithms. These include automatic content moderation, which is essential to safeguard broadcast channels and audiences. In addition, we have prioritised the integration of a facial recognition algorithm, which enables the automatic identification of people appearing in video content (e.g. the lead actors in a film). Finally, automatic speech recognition was another key area of focus, allowing spoken dialogue to be converted into text for purposes such as captions and searchability. This emphasis on core functionalities laid a solid foundation for further exploration of AI applications in the broadcasting landscape.”
In the initial updates with AI, how has the integration with YES! proved crucial?
“Building upon the initial functionalities, SI Media’s latest developments aim to further empower our AI suite with groundbreaking features. In the realm of face recognition, we’ve made significant strides by enabling the algorithm to learn dynamically. This means it can now recognize new faces across different assets, allowing users to assign names later at their convenience. Similarly, we’ve taken speech-to-text (or automatic speech recognition) to the next level by integrating a translation algorithm. This creates a comprehensive set of tools, complete with a dedicated interface, for managing and generating captions and subtitles in multiple languages. Expanding the capabilities even further, we’ve introduced a brand-new object recognition algorithm. This enables broadcasters with the flexibility to adapt the set of recognizable objects to their specific needs.”
Moving into the future, what are the next goals SI Media wants to achieve with AI?
“Looking ahead, SI Media is aiming to incorporate semantic search into our AI suite. This innovative technology overcomes the limitations of keyword-based search by going to the heart of content: its meaning and context. Imagine searching for a specific event not only by its title, but also by the central themes, the emotions evoked or even the personalities involved. Semantic search unlocks this potential by analysing the content itself, including audio transcripts, closed captions and even visual elements such as scene recognition.”
We would like to thank Davide for this interview on YES!AI and SI Media’s future goals to remain competitive in a market that, with the advancement of artificial intelligence, is constantly seeking innovations that meet the needs of media management.
