On Wednesday, September 19th, we headed to Washington, D.C. to attend SatSummit 2018. This annual event brings together experts from satellite companies, the mapping community, and academic and humanitarian organizations to talk about the problems and solutions related to satellite-derived data. Although it was our first time attending the conference, we gather that this year was particularly focused on automation, a topic near and dear to our hearts. We attended plenary sessions, smaller workshops, and a stellar afterparty at the Smithsonian National Air and Space Museum. Without further ado, here are our 5 favorite takeaways from SatSummit 2018.
5: AI Might Not Put Us All Out of a Job
A recurring fear voiced in the last few years (and decades) is the possibility of automation making human labor irrelevant. Whether self-driving trucks, automated farm equipment, factory robots, or AI support bots are the subject, humans appear to be in ever-greater risk of obscelesence. Panelists on the aptly named Bot or Not? workshop, including Janine Yoong from Mapillary and Jason Sundram from Facebook, posited that human jobs will shift, not disappear. As we quantify and qualify the world, we continue to want to know more about what it contains, requiring new algorithms and new techniques. And humans will always want to know why an algorithm made the choice that it did: for the near future, AI-based solutions require either human validation or can be used to validate, not supplant, human effort.
4: SAR Is a Powerful Resource SAR, or Synthetic Aperture RADAR is a powerful technique for creating increasingly high resolution 3D scans of large geographic areas. Its power notwithstanding, the vast majority of mapping, vision, and AV companies have focused on laser-based LiDAR scans to produce 3D representations of areas. Panelists including Drew Bollinger from DevelopmentSeed and Dr. May Casterline from NVIDIA, speaking on application of AI to geospatial data, both touched on the information you can derive looking beyond visual imagery. We were intrigued, not least because we have also been working with 3D data of several different types.
3: Humanity is Overwhelmed by Data In a set of flash keynote talks, Andy Revkin from National Geographic spoke about the firehose of data we now have available to us, and the importance of finding new ways to query and search data. On the other hand, several speakers discussed the difficulty of getting enough training data to successfully run machine learning algorithms on huge geospatial datasets. We're all too familiar with both overwhelming quantities of geographic data and training data sets, but we're excited about ongoing advancements toward solving data overload that make it easier to index multimodal data.
2: Customers Want Semantic Information
One solution to the data deluge is indeed parsing, indexing, pulping, and sorting that data. Our particular passion is indexing the physical world, so we love both the growing amount of data available for us to process, and the ways that we can distill that down into queryable semantic information. Jubal Harper from Microsoft spoke on the Artificial Intelligence 101 panel about more people wanting vector maps and information than raw 2D rasters, and even more wanting statistical information about broad areas. On a later panel, several speakers agreed that the utility of AI is in the derived data that is simply too tedious or time-consuming to extract by hand, especially when the source data (the real world!) is constantly changing and evolving.
1: Nothing Can Stop Women in STEM The highlight of the Summit for us was the keynote at the afterhours party, held at the Smithsonian National Air and Space Museum. Amazing women Bronwyn Agrios, Cady Coleman, and Megan Smith, collectively a fascinating Venn diagram of entrepreneurs, activists, CEOs, CTOs, an astronaut, and a rocket enthusiast, held a spirited conversation on their careers and achievements in space and technology. Their stories, their success in the face of an uneven playing field, and their forecast for tomorrow's women in STEM were uplifting.
Comments