Fredric Van Haren was one of the pioneers that developed the world's first speech recognition. While at Nuance, he built a High Performance Computing (HPC) environment even before HPC was a thing. He was one of the first to deploy Graphic Processing Units (GPUs) as opposed to traditional CPUs. He was later involved in many of the voice recognition companies such as Cerence, SIRI, and others.
In this entertaining and informative fireside chat, speech recognition pioneer Fredric Van Haren, CTO & Founder of HighFens, talks with Lynn Orlando, Head of Content Marketing at WekaIO, about how speech recognition training models work, how to architect an environment to support vast amounts of audio data, and what the future holds for GPUs. Here’s what you can expect to learn:
- How speech was “ground zero” for much of the modern AI effort
- Why the data-processing path from voice to text is so complex
- Why the combined role of GPUs and multicore CPUs is important
- What products and services we can look forward to in the future as speech recognition technology continues to evolve