Quantum Artificial Intelligence Lab Comes to Ames
NASA Collaborates with Google and USRA for Quantum Artificial Intelligence Lab, SpaceRef
“Quantum computing may be the key to solving some of the most challenging computer science problems. This is why Google in collaboration with NASA and the Universities Space Research Association today announced that they will launch the Quantum Artificial Intelligence Lab.”
Watch out for those Quantum tin whiskers. Curiosity kills the electron
Considering how vulnerable NASA is to budget cuts, it might have been smarter to try and cover this stuff under “defense” spending. In any case, I hope it works out – this is cool stuff.
In a quote from a BBC article (http://www.bbc.co.uk/news/s… about this, “Nasa will likely use the commercially available machine for scheduling problems and planning.” Does NASA have serious scheduling and planning problems?
Yes, NASA does have scheduling and planning problems, but not of the sort you might expect. This is a significant issue for flight missions: What observations a spacecraft should make (if two or more sorts are desirable and only one can be made at a time), what data to downlink from a spacecraft (if sending it all down isn’t feasible), which spacecraft may use the DSN antennas (or, more correctly, how the time on the antennas should be schedules), etc. The current solution is time and workload intensive, and effective ways to automate the process would be an improvement. Having been involved in this sort of work, I think it would be very difficult to automate, and I’m especially skeptical when people start using buzz words like artificial intelligence and quantum computing. But I will say there are scheduling and planning problems that could benefit from a better solution.
D-wave’s quantum computer is a fascinating development. It’s like something from the future. It can learn how to generate the optimal solution to a problem based on selected inputs. I’m confident its usage will be 24/7 and there will be a lot of valuable results. I hope Google buys a few more quantum computers and provides open access. I recommend a tour of D-Waves web site. The architecture and software stack are very well presented. Say goodbye to sequential computers.
One aspect of this, I’m assuming, is the calculation of launch times for planetary and other BEO missions. Since launch time depends on trajectories, and trajectory parameters are basically approximations (since no one has yet solved the n-body problem), the math can be complex and time-consuming. Anything that speeds up or automates this process is a plus, especially in the event of unexpected changes. The ability to recalculate a burn in minutes instead of hours could save lives.
It’ll be interesting to see how forthcoming any progress from this group is. With NASA and the Universities Space Research Association involved there’s at least a chance that they will share non-patentable stuff that they learn or develop with the rest of the industry and the world at large.
However, if Google happens to be the major financial contributor, which seems likely to me, then I wouldn’t expect them to give anything away. As the capabilities of their search indexes and algorithms continue to drop over time, Google seems to get more and more cut-throat, diversifying into many areas and holding everything they learn close to the vest. This may be good policy from the perspective of the shareholders and Board of Directors, but the rest of us lose out as Google continues to try to controi the whole world’s digital data for their our profit.
Google’s social habits are very different today from what they were in the beginning. It makes groups like Wikipedia look that much better.
I suspect we’ll hear very little from this Quantum AI lab after this initial announcement. I hope I’m wrong.
Very hard to justify that this sort of research is consistent with NASA’s mission. Sounds a lot like Ames looking for ways to justify it’s existence through diversification.
Although this particular work appears to be oriented toward earth-based applications, AI has major applications in the control of robotic spacecraft and rovers at distances which make teleoperation impractical, indeed NASA is (SFAIK) the only government agency that conducts robotic missions at distances that require autonomous control.