Car manufacturers use computer vision to teach cars to identify and avoid hitting pedestrians. A plot point of a recent episode of the comedy Silicon Valley dealt with a computer vision app for identifying food. This question of humor labor is important. Massive databases of correctly labeled images are crucial to training AIs, and can be hard to come by. Other sites and apps have to find their data elsewhere, often from academic images. Screenshots from the iNaturalist app, which uses "deep learning" to automatically identify what bug—or fish, bird, or mammal—you might be looking at.
But there are some limitations. First, it can be much harder to get an identification of your photograph depending on where you live. In California, where Loarie is based, he can get an identification within an hour.
But someone in, say, rural Thailand may have to wait much longer to receive an ID: The average amount of time it takes to get an identification is 18 days. Another issue: As the site has become more popular, the balance of observers people posting pictures to identifiers people telling you what the pictures are has become skewed, with far more observers than identifiers.
If you're ever bitten by a snake or spider you should hope the species can be correctly identified or the anti-venom administered may be incorrect. That plant may yield a new anti-bacterial drug, but are we sampling one species or two that look alike?
A new mosquito is spreading, but is it a malaria carrying species or not? The real question of course though is how can species be recognised and identified? This is where things get complex and disagreements can arise between biologists, since species are more fluid than elements or atoms. By definition, species evolve and over time populations change, diverge and lineages split into new species. Humans now might be considered the same species as humans ten thousand years ago, but it's also undeniable that we have changed in that time.
You might well be familiar with the definition of a species that runs roughly as "a group of animals that can reproduce and have fertile offspring", and that's all well and good, but it's also profoundly limited. Plenty of species don't have sex bacteria, some lizards and sharks, many plants so this definition is irrelevant for these cases and there's tons of them. We can't separate out fossil species by this definition either, and some things can produce fertile offspring despite being very different in appearance, or being separated by another non-genetic barrier behaviour, geography etc.
Users take part in a worldwide citizen science project—called GLOBE at Night — by mapping light pollution and star visibility. The results are added to a database that scientists use to investigate the reasons for the increasing illumination of the night, its ecological, cultural and socioeconomic effects as well as the impact on human health. For example, excessive light can cause sleep disorders in people. The research is expected to aid in the development of improved lighting concepts and sustainable technologies.
How it works: Loss of the Night allows for the measurement of light pollution in three steps. The first is an arrow that guides users to a star, similar to a compass. The app then asks users to select visible stars in various constellations and submit their data once observations have been completed. The observations contribute to helping communities learn what works in terms of mitigating light pollution and how improved street lamps can save energy and money.
Measurements from the app are sent anonymously to the GLOBE at Night database, a citizen science project launched in Similar apps: For citizen scientists with iPhones, a similar app is available for download in the iTunes App Store. The greater the difference between your dark shot and sky shot, the more reliable the data.
Image: Kinsey Reporter. Kinsey Reporter: Kinsey Reporter , as the name suggests, is a global mobile survey platform for sharing anonymous data about sexual behaviors. The data collected with the app is aggregated and shared openly at KinseyReporter. How it works: Using Kinsey Reporter, participants can submit an anonymous report of sexual or intimate activity within 24 hours of an event.
The resulting surveys—or reports of information shared by many individuals—cover sexual behaviors and events, sexual health issues, violence reports, public displays of affection and other experiences. To protect anonymity, the reports only use approximate locations selected by survey participants. Reports are aggregated over time and are not published until a sufficient number of entries have been received from the same location. Image: Marine Debris Tracker.
Marine Debris Tracker: Marine Debris Tracker can be used to find and log marine debris items on beaches or in the water. Platforms supported: Android, iPhone. Note: An updated version of the app is available for Android devices. The last update for iPhones was released in , but a newer version is planned for next summer.
How it works: The app records the location of debris via GPS, and allows the user to review the data and submit it to the Marine Debris Tracker Web site. Users are initially required to create a user name and password to view and download data from the site. All downloadable and mapped data with GPS coordinates is visible to the public anonymously. Marine Debris Tracker comes with a preloaded list of commonly found marine debris items, split into several categories. Examples of debris that citizen scientists can report include beverage bottles, clothing and shoes, car tires, building materials and fishing gear.
Similar apps: iPhone users looking for similar citizen science apps should check out Creek Watch. The app allows users to snap a picture and report on any waterway they come across. By monitoring the health of local waterways, including water amount and trash, participants can help watershed groups, agencies and scientists track pollution and manage resources. Image: NoiseTube. NoiseTube: NoiseTube was created with the purpose of turning smartphones into mobile noise level meters.
How it works: The app allows citizen scientists to participate in the collective noise mapping of their city or neighborhood. NoiseTube has three features: measure noise, localize it and tag it. Tags include the level of annoyance and the source of sound, such as an airplane. The collected data is wirelessly sent to the NoiseTube server in real time. Additional features: Citizens concerned with noise can visualize their measurements on a map and contribute to creating collective, city-wide noise maps.
0コメント