We held our next tinyML Talks webcast. Dominic Binks from Audio Analytic has presented Making ML work in the real world on February 23, 2021.
While ML researchers continue to deliver new techniques and tools to improve the performance of ML-based systems, there’s often a gap between those developments and ML systems working in the real world: why is it often the case that when ML models are deployed in the real they do not work anywhere near as well as lab experiments suggest? What is it that leads to this disconnect?
In this talk, Dominic will share elements of our product design process, share some amusing stories from deploying ML models to real-world environments, see how the real world can throw up surprising, unexpected and plain strange scenarios and propose tools and techniques to help build ML models that really work.
Dr. Dominic Binks was previously a Staff Engineer at Qualcomm working in a variety of different software roles prior to joining Audio Analytic. At Qualcomm in Cambridge, he worked on mShop, a BREW-based shopping application and Vuforia, Qualcomm’s cross-platform augmented reality SDK. In addition Dominic spent time in San Diego working on Qualcomm’s core Android porting team with responsibility for the build and release team. Prior to Qualcomm, Dominic worked in technical presales at SavaJe, Android’s forerunner. Before joining SavaJe Dominic worked as a technical consultant at Scientific Generics (now Sagentia) and prior to this he worked on pre-paid calling platforms deployed to a number of mobile operators worldwide. Dominic’s PhD investigated techniques for automating fault finding (debugging) in pieces of software.
Watch on YouTube:
Download presentation slide:
Feel free to ask your questions on this thread and keep the conversation going!