In a year shaped by the emergence of ChatGPT and other large language models, with a $25 billion investment in the technology signaling a new gold rush, it’s fitting that the opening keynote at the 15 September ITS: Automation event (being held during IBC at the W Amsterdam) is Konstantin Dranch.
The founder of Custom Machine Translation (Custom.MT), an AI and machine learning operations agency, will walk attendees of the event through recent localisation triumphs and mistakes with models, and see if there is a better way to apply them in media workflows.
His presentation — “Large Language Models in Media – 5 Learnings from the Technology Rush in Localisation” — will dive into how industrial corps teach thousands of employees to write prompts, and even politicians are involved in writing AI acts. Large language models are re-shaping the way humans interact with devices and are changing the future of many professions. Text localisation has become the industry to immediately take advantage of LLMs, and to start grappling with their issues and hallucinations.
“I will speak on the large language models’ impact [and how] they have spread like wildfire in testing scenarios, and got tremendous interest with the managers and investors,” Dranch said. “Now that the hype is dying down, what remains? It turned out that so far, they achieved small incremental improvements across processes, and not as huge life-changing advances people feared and expected. Tiny steps, lots of experiments, reverting often to tried and tested methods.”
Dranch will offer examples of what people have built to date in translation and localization, subtitling, and even in news/broadcasting summaries.
“Based on these cases I will outline the current landscape and present a framework to apply large language models — in processes, products, and in promotions,” he said.
This inaugural ITS: Automation event, produced by MESA and sponsored by Amazon Studios Technology, Fabric, Eluvio, EIDR, and HAND (Human & Digital), will run from 3-6 p.m. and examines the role automation is playing in workflows, metadata, cloud operations, and even how we license, track, and monetize our talent.
To register for the event, click here.
