As usual, the Office of the CTO (OCTO) hosted interns for the summer. We have our interns work on real research projects rather than just create busy work. One of our favorite aspects of the program is that our interns share their results at the end of the summer at our intern showcase. OCTO is spread across locations like London, Boston, Toronto, and San Francisco. Through the miracle of video conferencing, we all got to see the work of all of our interns regardless of physical location.
Here is how our San Francisco interns spent their summer with us.
Machine Learning for Architectural-Scale Robotic Assembly, Rodger Luo, San Francisco
Machine learning can be applied to instruct robots rather than manually programming them to perform desired tasks.
Four Seasons of Work: Future of Work — Worldbuilding Project, Jessica Escobedo, San Francisco
Science fiction has always been a source of inspiration for technological advancement. Four different automation scenarios, depicting how designing and making will change, provide prototypes for how Autodesk customers will work in the future.
MUTU: A new learning paradigm for mutual growth, Humberto Ceballos and Yaoli Mao, San Francisco
A set of data-driven learning and community tools can uncover personalized learning pathways so our users can master our tools. Data can show a user's strengths and weaknesses. Training can then be supplied for the weaknesses.
Boise noise canceling for the brain, Neta Tamir, San Francisco
Just like noise-canceling headphones, we can identify noise in our customers' workflows and develop tools to cancel out the cognitive noise. For example, analyzing command usage can identify customer biases towards more-familiar commands instead of commands that would be more appropriate for a task.
Immersive Design in Augmented Reality and Virtual Reality, Miao Ren, San Francisco
Generative design results can be viewed in context by using augmented and virtual reality.
A Learning Method for Industrial Assembly, Yongxiang Fan, San Francisco
Machine learning can be based on artificial data from simulations instead of having robots perform trial and error operations physically.
Unsupervised Shape Structure Learning, Kaichun Mo, San Francisco
Given a 3D object, it is possible to use computer vision and machine learning to deduce from which shapes it is made and the relationships among those shapes. These shapes can then become the building blocks for future generated designs.
Developing Next Generation of Design Software for Manufacturing, Lin Cheng, San Francisco
Generative design is currently realized with additive manufacturing, but additional research will open it up for use via other fabrication methods.
BIDS — Building Information Design Synthesis, Mohammed Keshavarzi, Spyros Ampanavos, Yi Wang, and Kevin Frans, San Francisco
By analyzing data from millions of past construction projects, it is possible to use machine learning to design appropriate buildings, estimate costs, compute realistic schedules, and predict which contractors and trades will have problems on future projects.
Our CTO once told our department, "I love you guys, but my favorite day of the year is the intern showcase. As you can see, our CTO was not kidding as our interns do some impressive work. This year was no exception. Thanks to all of our interns for the hard work thsi summer.
Autodesk has always been an automation company, and today more than ever that means helping people make more things, better things, with less; more and better in terms of increasing efficiency, performance, quality, and innovation; less in terms of time, resources, and negative impacts (e.g., social, environmental). Working with interns injects new blood and new perspectives in our quest to provide software that helps our customers do more and better with less.
Interning is alive in the lab.