DAWN-IRES Application Challenge Questions
Applicants are strongly encouraged to answer one of two challenge questions. These are intentionally open ended questions, much like most research projects, with the goal of getting a better understanding of how you approach problem solving.
Challenge Question 1:
Living in California long enough means that you are very likely to eventually experience a major earthquake. Although buildings are constructed with this in mind, there are still some locations and activities which would be dangerous in a major earthquake, and even a small amount of advance notice might save lives. In earthquake-prone areas with lower construction standards, this would be even more valuable.
Japan uses a high-quality sensor network in an attempt to solve this problem, because the earthquake (a few km/s) travels much slower than the network (~speed of light). For a large earthquake in which a fault ruptures over hundreds of miles, the network detects earthquakes in approximately 5 seconds, which can result in an early warning of as much as 60-90 seconds in some cases. A similar network does not currently exist in the United States, but on the other hand, you and a very large number of others have accelerometers in your phone, so perhaps a lot of low quality data might be a substitute for a small amount of high-quality data.
Suppose you were given access to the accelerometer data from every phone in the United States. Could you reasonably design a system which would efficiently detect earthquakes within 5 seconds, so that a warning could be sent out? How would you plan on screening for false positives, given that the accelerations you produce when walking around are much larger than in even a major earthquake? Can you distinguish an earthquake from a massive, synchronized movement of many people such as the L.A. marathon? How many cell phones would need to opt in in order for your system to work well?
Challenge Question 2:
Since we're thinking about scientific solutions for disaster scenarios, suppose that we were to discover a large, interstellar object (such as Oumuamua) is on course to collide with the Earth in 10,000 years. If nothing is done about it, the resulting impact would be expected to produce a mass extinction, including human life. The good news is, you have the technology to fix it. It's somewhat complicated, as it involves building a spacecraft, something which we have only been able to do for a little over a half century. But as long as somebody follows your instructions about 10,000 years from now, the Earth will be saved.
The problem is, you need to communicate this information in a way that people 10,000 years from now will be able to understand the danger and follow your instructions. The history of human civilization suggests there may be several dark ages between now and then, during which key technologies and ideas will be forgotten. Many things which were widely known 2,000 years ago are currently unknown or were only recently rediscovered, and some ancient languages are still impossible to decipher.
What would be your plan for communicating information about the upcoming impact and how to avoid it? Do you think this is likely to work? A related question is how one should label nuclear waste, which is toxic for a similar timescale. Would you solve this problem in a different way given that it is necessary to communicate the danger not just specifically 10,000 years from now but continuously until then?