Presentations

Introduction to AI in the medias

It is often said that the adoption of AI is everyone’s business in an organization. This training aims to democratize AI and enable its use in all organizations, regardless of size, that have completed or are undergoing a digital transformation. We will see an overview of methods, applications and risks allowing participants to better understand the potential and risks of AI in relation to their professional reality.

Mark Heaps, Chief Tech Evangelist & VP of Creative at Groq Inc will showcase how AI technology is not only ready for tomorrow but ready right now. The Society of Motion Picture and Television Engineers attendees will experience a comprehensive overview of the current state of Artificial Intelligence (AI) technology and its applications in the media industry. Through a series of demonstrations, Mark will illustrate the variety of applications built by the Groq AI developer community, showcasing how these innovations enhance the workflow of creators and professionals involved in the production of media, from social media to film. The LPUTM Inference Engine, Groq's cutting-edge technology, enables the world's fastest inference for AI applications, making these advancements possible. Mark will also provide an in-depth examination of the evolution of AI technology and its implications for the media industry. He will discuss the key considerations and opportunities that arise from the integration of AI with creativity, highlighting the potential for Groq to power the next generation of AI-driven innovation in the media sector. Key takeaways: ■ How Groq technology transforms workflows ■ How Groq’s LPUTM will impact media/production environments ■ What augmenting human capacity means for the next generation of media workers

This presentation will provide a comprehensive overview of Intel's latest AI technology. The first topic will cover the new AI accelerators in Intel CPUs -- such as Xeons for data centers or the Core Ultra processors for AI PCs. The presentation will continue by exploring a completely new approach by Intel for deep learning: the new AI processor Gaudi3. Since Intel also stands as a leader in open-source software development, the presentation will explain how oneAPI enables AI workloads across all industries. The presentation will conclude by outlining Intel's future in the foundry business.

To stay competitive in today's media landscape, live media companies are challenged to deliver engaging content that retains audiences with new viewer experiences that are personalized, interactive and immersive. The use of Artificial Intelligence (AI) technologies is one tool to assist studios in achieving this goal. Developing and deploying AI-Enabled products is challenging, requiring an end-to-end developer stack. NVIDIA provides the most advanced platform for the development of AI solutions for media with innovations at every layer of the stack from accelerated computing platforms, to essential AI Software Development Kits (SDKs), pretrained models and AI foundries for the deployment of AI-based media processing on prem, on the edge or in the cloud.

In this session, we will explore the transformative power of AI/ML in the Media and Entertainment industry. We'll delve into compelling use cases such as content moderation, live captioning, automated video highlights, content personalization, and localization. Discover why AWS is the ideal platform for running AI/ML workloads, with its comprehensive AI/ML stack that addresses different user personas. Get ready to dive into real-world architectural patterns, and gain insights into how you can get started by using AWS for your AI/ML needs in the M&E space.

Artificial intelligence (AI) is evolving rapidly, and Microsoft is playing a key role in this acceleration. Microsoft is strategically positioning itself in the field of AI with a focus on generative AI and Copilots. This presentation will discuss this strategy, and the impact of AI in media and entertainment, illustrated by customer case studies. We'll also discuss Microsoft's responsible framework for AI, and conclude with a demonstration of a Copilot to truly accelerate business analytics.

Content creators require more than traditional search methods to meet the demands of speed, accuracy, and relevance. This presentation explores a transformative case study at Avid’s Advanced Research and Development (RAD Lab), where Artificial Intelligence (AI) and Machine Learning (ML) have been employed to overhaul traditional search mechanisms through semantic search capabilities. Leveraging AI-driven technologies, the project has established a more efficient, accurate, and cost-effective method for searching and retrieving media content. The presentation will begin with an overview of the shortcomings of conventional search techniques that rely heavily on manually tagged metadata. It will contrast these with the advanced AI/ML approaches that utilize semantic indexing to understand and interpret media content’s contextual and cultural nuances. The core objectives of this initiative were to enhance the search functionality, improve retrieval accuracy, and reduce operational costs. The project’s planning and implementation phases are outlined, demonstrating the integration of Avid’s established ecosystem components—Nexis, Elasticsearch, and Cloud-UX —culminating in a seamless transition from a proof of concept to a fully functional demo. We will discuss Avid’s engagement with Canada’s Mitacs program, which provides funding to support research internships for graduate students and postdoctoral fellows in collaboration with industry partners. Key outcomes and performance metrics will be shown, highlighting the project’s achievements in accuracy, time efficiency, and cost-effectiveness. A live demonstration will showcase the enhanced semantic search capabilities within the Cloud-UX platform. Finally, the presentation will address the challenges faced during implementation, celebrate the successes achieved, and outline the future roadmap, which includes scaling the solution and extending its capabilities to additional media types.

From Ici Première to Ici Musique Classique, from national broadcasts to regional broadcasts, from podcasts to major series, Radio-Canada produces several dozen hours of content per day. In order to improve our content offering and optimize our operations, transcription is becoming a central element in our work. This allows us on the one hand to make our content accessible to deaf and hard of hearing people and on the other hand to apply different text-based machine learning methods, such as automatic summarization. Our presentation explains the process of implementing a transcription tool at Radio-Canada, from proof of concept to production.

Explore certain tools related to generative artificial intelligence to see if it is possible to simplify and accelerate (or even replace) rotoscoping for a documentary project with animated sequences.

The recent emergence of impressive generative artificial intelligence models is capturing attention as well as raising questions. In order to explore the potential of this technology, it is essential to determine concrete use cases that can be resolved through the use of large-scale language models. I have the privilege of presenting to you our attempt to create business value by exploiting generative artificial intelligence through our corporate ChatBot through the following subjects: - Generative artificial intelligence and how to meet its challenges - The architecture of the ChatBot and the augmented recovery generation (RAG) - The use of generative artificial intelligence beyond ChatBot

The explosion of online multimedia content is putting a strain on distribution and storage infrastructures. Efficient video encoding becomes crucial to reduce costs and ensure a quality user experience. This presentation explores how artificial intelligence (AI) is transforming the video encoding landscape, leveraging three key technologies: - VMAF (Video Multimethod Assessment Fusion): AI for precise and subjective assessment of video quality. - Content-optimized encoding: Adaptation of encoding parameters specific to each segment, for optimal quality to reduce bandwidth costs. - Super-resolution: Stream low-resolution video enriched with metadata to enable AI enhancement at the recipient's end, improving visual quality while reducing bandwidth costs. By highlighting the benefits and practical considerations of AI in these technologies, we guide professionals toward adopting next-generation video encoding solutions for optimized streaming and cost-effective storage.