Intel co-founder Gordon Moore noted in a 1965 paper titled “Cramming more components onto integrated circuits” that processing power doubled roughly every two years. He predicted conservatively that the trend would last at least another decade.
It is still in evidence and the exponential impact is mind-boggling. A mid-range 2010 laptop, or smartphone, has more computing power than planet Earth, circa 1970. It costs less than a tenth of the first 1980s PCs. Hundreds of millions of such devices are permanently hooked to the Web creating a lot of spare computing power and communication bandwidth.
Researchers have been finding ways to exploit these spare resources. The advent of Web 2.0 and social networking has added a new dimension. The combination of surplus computing resources and easy access to the humans who own them has created a new generation of citizen scientists.
It enables crowd-sourcing and massively collaborative efforts aimed in many different directions. Perhaps the most famous experiment is run by Search for Extra-Terrestrial Intelligence (Seti). Seti analyses radio signals captured from space to find patterns that may indicate communication.
The seti@home programme utilises idle resources. Over 200,000 volunteers donate time on around 300,000 computers. Radio signal analyses are run whenever machines are idle. The results of this distributed computing effort are collated “back home” at the University of California’s Berkeley campus.
Astronomy and space exploration, in general, encourages collaborative efforts and amateurs often spot new phenomena. The National Aeronautics and Space Administration invites amateur inputs in many programmes. In June, a bunch of 12-year-olds found a new Martian cave while steering the camera on the Odyssey via the Web.
Since January 2010, the HiRISE, (High Resolution Imaging Science Experiment) camera on the Mars orbiter has been taking metre-resolution pictures on the basis of emailed suggestions. Each suggestion must be accompanied by a relevant scientific reason.
Another astronomical collaboration involves human pattern-recognition skills rather than idle computing resources. With a day or two of training, humans can recognise supernova explosions with very high accuracy and lower error rates than the best algorithms.
Supernovae are diagnosed when stars suddenly increase in brightness. False alarms can arise from approaching asteroids, variable stars, comets and so on. Hence, humans must compare several images of the same part of the sky to eliminate false positives. Images used to be sorted by a team of eight PhD students at Oxford. In 2010, a team of 2,500 volunteers has already sorted over 14,000 “probables” with 93 per cent accuracy.
Citizen scientists have also been tasked to find meteorites and collate data for various naturalist and climate change programmes. The Zooniverse, which runs a massive collaborative effort called SpaceZoo to identify new astronomical objects, is working with the UK Meteorology Dept to scan ship captains’ logs from the 19th and 20th century.
Logbook data include daily observations of oceanic weather, temperature and air pressure, which can be fed into climate models. While it’s easy to scan and transfer logs, humans are far superior to computers at deciphering faded handwriting. By getting volunteers to man the Old Weather project, data acquisition has been much accelerated.
Modern climate change is also monitored through projects like the Ice Watch programmes in which citizens log in every day with granular, local weather observations. Cornell’s Lab of Ornithology has been highly successful in mobilising bird-watchers to gather data about migration, nesting and breeding patterns. Similar programmes study changes in green cover, prey-predator relationships and plant migration.
Human skills in pattern recognition are also being harnessed in other ways. The gamer community’s skills were first harnessed by Luis Von Ahn, a computer scientist at Carnegie Mellon, who conceptualised the ESP game. In ESP, two gamers (who cannot communicate with each other) are asked to tag the same pictures. When they agree on tags, those images can be labelled. By iteration, with the same image offered to hundreds of gamers, the tags become generic. Google bought the ESP license from Von Ahn and uses a variation of this labelling technique in its image search engine.
In Foldit, a multiplayer game that has around 57,000 active users, players fold and unfold amino acid chains. Determining how amino acid chains fold dense proteins is a cutting edge challenge in molecular biology. Mis-folding is linked to diseases like Alzheimer. The Foldit community has beaten the best computer programs in determining the folding patterns in some new proteins.
Another new multi-player game, Phylo (which is on Facebook) is also focused on genetics. Players move coloured squares representing four DNA nucleotides to find the best way to plug together DNA sequences. These “promoter regions” of DNA determine which parts of the genome control traits like eye colour or a predisposition to diabetes. You don’t need to have a clue what it’s about to play Phylo; it’s designed to appeal to anybody who likes Tetris, or Farmville.
It’s oddly heartening to think that a bunch of people goofing off and playing games when they should be “working”, could actually contribute significantly to scientific advancement. The new era of the citizen scientist could change all our perceptions about idle time.
