In the lead up to awAIre, we’re taking some time to speak to key Executives from the leading companies supporting the show. Today we’re talking to Oliver Brdiczka, Principal AI Scientist at Adobe.
Tell us a little about yourself and your journey in tech?
I grew up in a small town in southern Germany, and went to France for my graduate and PhD studies. During my PhD, I got interested in the AI and IoT space (back then, called Ubiquitous Computing), more precisely I wanted to get a computer system to understand & respond to what humans do (or intend to do) by relying on data from various sensors. This interest or passion continued when I joined Xerox PARC, but the focus shifted to online sensor information, i.e. could we understand human intent from what people say and do online, and provide assistance. Still at PARC, this work led me to participate in various cutting-edge DARPA programs, developing intelligent contextual assistants, as part of the DARPA PAL program (commercialized as meshin and SIRI today), and human personality prediction, as part of the DARPA ADAMS program (commercialized as part of IBM Watson today). I continued on that journey with a stint in cyber-security to identify bad online behaviors, and then my own startup building an AI system that understands & assists users to find jobs. Since I joined Adobe, I came back to the initial vision of understanding human intent from various behavioral and visual cues … can we derive the creative intent of a user and assist them in the creative process?
What attracted you to work in the space of context & AI ?
I like the idea that this space can have a huge impact on people’s lives. Everything people do is highly contextual, humans rely on various contextual cues in their decision making & behaviors. By replicating similar contextual reasoning processes, I believe we come closer to the vision of a human-like AI in the future, and I hope my work will contribute a small puzzle piece to it.
How have you applied context & AI in your current role at Adobe?
In my current role at Adobe, I am tasked with building the architectures for contextual AI systems, with a focus on Digital Imaging and our flagship apps there. The problem to solve is how we can provide contextually aware assistance to users during their creative process, e.g. when they are creating artworks and need specific tools or suggestions. The Discover Panel and Neural Filters in Photoshop are early versions of this shipped in 2020.
How do you see the space of contextual AI developing in the future?
I think one aspect that became more and more important in recent years is the correct handling of users’ data and privacy. In the early days, building contextual AI systems was very complex & challenging, so nobody had the time to even think about privacy implications. However, as some of these systems have now reached maturity, privacy has become one of the most important considerations. How can we build a system that respects the user’s privacy and yet sees & aggregates enough sensor information about the user to be useful? Edge computing and smart privacy-preservation will be key for future success.
Thanks Oliver – Looking forward to hearing more from Adobe at the event !
Register FREE to see Oliver speaking at AwAIre on June 23rd.