Frequently Asked Questions

Background

• Where can I learn more about UCLA SETI?
  You can read about our search by browsing the UCLA SETI website, reading past issues of or subscribing to our newsletter, or asking questions in the "Talk" section. We are also on Facebook, Instagram, and Twitter.

• Where does the funding come from?
  Funding for our research and teaching programs comes primarily from generous donors. We have also secured a three-year NASA grant from the Exoplanet Research Program, which primarily funds a graduate student but does not fund telescope time or equipment. Funding for this collaboration comes from a two-year STEP grant from The Planetary Society and a one-year NASA grant from the Citizen Science Seed Funding Program.

• How can I support the search?
  You can support the search by classifying signals on this platform and by promoting arewealone.earth on your social and professional networks. If you are in a position to help financially, you can make a donation on a secure giving site. Your gift will enable the purchase of equipment and telescope time to collect data, improve the training of the next generation of scientists and engineers, and help bring the excitement of SETI to the classroom and the general public!

• What do you mean by citizen science?
  In this document and other communications, we have used the terms "citizen science" and "citizen scientist". Our use of the term citizen is inspired by the American Astronomical Society ("Every astronomer is a citizen of the community of science") and by the fact that all our collaborators are citizens of Earth, as reflected in http://arewealone.earth.

• Where do the data come from?
  All the data uploaded to the platform were acquired by UCLA SETI as part of a search for narrowband radio signals with the largest fully steerable telescope on Earth, the 100 m Green Bank Telescope in West Virginia. Almost all (>99.5%) candidate signals are automatically classified by our data processing pipeline as anthropogenic radio frequency interference (RFI). This platform hosts some of the remaining (<0.5%) and most interesting candidate signals.

• What are narrowband signals?
  A narrowband signal is a signal whose energy is concentrated in a narrow range of frequencies.

• Why are narrowband signals particularly interesting?
  An extraterrestrial narrowband signal would provide unambiguous evidence for the existence of another civilization. This degree of certainty is unusual in searches for biosignatures or technosignatures, which are usually hampered by the problem of false positives. Narrowband signals are compelling because natural processes produce only broadband signals.

• How many signals have been uploaded to the platform?
  The initial data set consists of 20,000 signals. We plan on increasing that number to 100,000. Data are released in batches, and the first batch contained approximately 7,500 signals.

• Will you be adding new data to the platform?
  Yes, our search is ongoing and we intend to upload fresh data to the platform as soon as the initial data are classified.

• How much search space have you covered?
  Since 2016, we have observed over 55,000 stars and detected over 82 million candidate signals. All of the signals that we have analyzed so far appear to be due to RFI.

• How many stars must be observed to ensure a detection?
  Nobody knows. It is possible that a signal is present in our existing data or will be recorded in upcoming observing runs. It is also possible that a signal will not be detected in our lifetime. SETI pioneer Frank Drake (1930–2022) opined that observations of 10 million stars might be necessary to enable a successful detection. With adequate funding, we could accelerate our search and observe a million stars in the next 5 years.

• Which scientific publications are most relevant to this project?
  A description of the importance of radio technosignatures and an outlook for the next decade is available in a 2019 white paper. A description of our initial methods and search results for planetary systems in the Kepler field is available in our 2018 paper. Improvements to our methods and search results around TRAPPIST-1 and other systems were published in our 2019 paper. A description of our current methods and search results around Sun-like stars near the plane of the Galaxy is available in our 2021 paper. Search results around TESS objects of interest and many other goodies are available in our 2023 paper. A description of a machine learning application that improves our direction-of-origin estimator is available in our 2022 paper.

Using the platform

• How can I access higher resolution images of the available classes?
  You may examine higher resolution images by clicking "Help with this task" or by clicking on the tab labeled "Classification task" in the Field Guide.

• How can I access and navigate the Field Guide?
  The Field Guide is accessible by clicking the words "Field Guide" on the right-hand side of your browser window. You can click on individual tabs in the field guide and return to the main menu by clicking the "<" symbol at upper left. When you are done reading the field guide, you can collapse it by clicking the "x" symbol at upper right.

• The task is about classifying images. Can you show me an example of each class?
  Yes, you can see an example of each class by clicking the tab labeled "Classification task" in the Field Guide.

• When I contribute to the project, am I identifying signals of interest or classifying radio frequency interference (RFI)?
  You are doing both simultaneously. We have uploaded the most promising 0.5% of signals to the Zooniverse platform and your classifications are helping in identifying the most promising signals among those. All single-line narrowband signals and all signals classified as "other" will be inspected by the science team. If the signal survives additional offline testing, we will follow a specific confirmation protocol (see below). Meanwhile, all your classifications are being used to generate a labeled training set for a machine learning application that we are developing. This application will improve the robustness, accuracy, and speed of future SETI searches and will be made available to the radio astronomy community as open source software.

• How are signals of interest ranked?
  All narrowband signals detected by our data-processing pipeline are considered "Level 1" candidates. Most (99.5%) of these signals are automatically classified as RFI. The remaining (0.5%) are considered "Level 2" candidates. The signals that we uploaded to the Zooniverse platform are "Level 2". All single-line narrowband signals and all signals classified as "other" by this collaboration will be considered "Level 3" signals and require additional offline examination by the science team. Signals that survive the science team's initial offline analysis will be considered "Level 4" and require confirmation through additional observations and peer review. A signal that is confirmed with additional observations and peer review will be considered "Level 5" and may constitute the first genuine technosignature ever detected and the first detection of extraterrestrial technology.

• What is involved in the science team's analysis of promising (Level 3) detections?
  We will verify that the signal comes from a single direction on the sky, which can only be done by examining the dynamic spectra of other scans at similar frequencies, a task that cannot be easily implemented on the Zooniverse platform. This analysis is called a direction-of-origin filter, and we have both classic and machine learning implementations at our disposal. If the signal is detected in more than one direction on the sky, it will be labeled as RFI and remain at Level 3. If the signal appears to come from a single direction on the sky, it will be further examined. For instance, we will compare the frequency and rate of change of frequency to known sources of RFI. Signals that cannot be ascribed to a known source of RFI will be considered Level 4.

• How will you handle promising (Level 4) detections?
  Any signal that cannot be ruled out as RFI on the basis of the science team's initial analysis will be treated according to a widely accepted post-detection protocol, which includes principles for confirming the signal. Specifically, we will reprocess the data with independent software, reobserve with the Green Bank Telescope, reobserve with different telescopes, and submit the results for peer evaluation. We will inform the collaboration about the results of these efforts.

• What if I make a mistake?
  Please participate and do not worry about making a mistake. We have several mechanisms for validating the classifications. First, multiple Zooniverse volunteers complete each classification, and we aggregate the responses. Second, the science team reviews the classifications of subjects where responses exhibit unusual variability. Finally, our machine learning application will be trained with thousands of images and will be able to handle occasional misclassifications.

• Why do most signals look vertical? Doesn't the frequency change as a function of time?
  Line-of-sight acceleration between the transmitter and receiver (e.g., Earth's rotation and revolution around the Sun, motion of transmitting platform) cause a shift in the Doppler frequency of the signal that is visible in the dynamic spectra, i.e., signals exhibit a finite slope in time-frequency space. Over the 150 second durations of our GBT scans, these drifts are well approximated by straight lines. Our data-processing pipeline uses computationally efficient algorithms to identify the drift rate associated with each candidate signal (For details, see our recent paper). We removed the overall frequency drift from our images so that you can focus on the time-frequency structure of the signal without being distracted by its slope.

• Which metadata are available?
  We provide the metadata associated with the signal located in the center of each image. The data types and descriptions of the fields are as follows:

  Field	Type	Description
  FILENAME	string	the name of the file as stored on the UCLA SETI servers
  DATASET	string	the name of the dataset as stored on the UCLA SETI servers
  ID	integer	a numerical identifier for the signal
  NAME	string	the name of the source as stored on the UCLA SETI servers (typically includes an astronomical catalog identifier combined with a student or donor name)
  SCAN	integer	the index number of the scan of the source in this dataset
  MJD	float	the Modified Julian Date at the start of the scan
  FREQ	float	the frequency (Hz) at the start of the scan for (X,Y)=(251,1) (counting from 1)
  DFDT	float	the rate of change in frequency or drift rate (Hz/s)
  SNR	float	the signal power integrated over the scan duration (standard deviations of the noise)
  PROM	float	an estimate of the signal prominence above the noise floor (standard deviations of the noise)
  BW	float	an estimate of the bandwidth (FWHM) of the signal (Hz)
  PARTNER	integer	the ID of the signal in a subsequent scan of this source, if known

• How can I access the metadata associated with a signal?
  After you have completed the classification task, press the "Talk" button and click on the information icon.

• What do the RFI codes mean?
  We have named 20 classes of pervasive interferers with three-letter codes. The letters "H" and "V" in the first position stand for horizontal and vertical, respectively. The letters "R", "I", and "S" in the second position stand for regular, irregular, and single, respectively. Other letters distinguish the various classes. The mapping between these codes and the various interferers will be refined as the project proceeds.

• Which sources of radio frequency interference are present in the data?
  We anticipate that a large fraction of signals originate from global positioning systems. We also anticipate signals from Air Route Surveillance Radars and other sources. The classifications obtained on this platform along with the metadata will help elucidate the mapping between RFI classes and specific interferers. Advanced users who wish to elucidate the various sources of RFI may find useful information in the Radio Regulations of the International Telecommunication Union.