During the final presentations, student teams also described their software contributions to our data processing pipeline. The student contributions were exceptionally strong this year. Here is a description of the students' various projects:
- Antoine and Robert wrote code to perform GPU-accelerated calculations of power spectra and quantified the reduction in execution time.
- Calvin calibrated power spectra by estimating the mean and standard deviation of the noise from a fit to the relevant probability distribution.
- Yugantar and Sai trained a denoising autoencoder and quantified the performance with respect to signal detection.
- Anthony wrote a program that computes the frequency drift rate for wideband signals.
- Atharva wrote a program to verify whether high-drift rate signals are genuine or incorrect identifications.
- Divij and Joshua wrote a program that verifies whether a detected signal intersects another detected signal in time-frequency space.
- Zefyr and Chester implemented code to improve the pairing of signals from separate scans of the same source.
- Megan, Jian, and Travis improved our data processing pipeline documentation with standard docstrings.
- Laura and Vedant wrote a program to compute the Doppler-shifted frequencies of GPS satellites.
- Maria and Janice wrote a program to compare the strength of signals according to the GBT antenna beam pattern.
- Michelle implemented a moving average calculation to estimate the number of detected signals per unit frequency.
- Anchal and Jordan created a graphical user interface for our data processing pipeline.
- Nadine wrote a signal injection and recovery program to test the efficiency of our data processing pipeline.
- Sarah created a testing framework for our data processing pipeline.
Despite the COVID-19 handicap, the Spring 2020 class enabled progress at a level comparable to the historical Spring 2016 class, which developed a data-processing pipeline from scratch. The results and contributions truly lifted my spirits during a challenging period in all our lives.
The students seemed to have enjoyed the course as well. In anonymous student evaluations only visible to instructors after grades were submitted, the median score was 9 on a 1-9 scale for overall rating of the course and overall rating of the instructor.
On a related note, UCLA graduate student Paul Pinchuk and I have been working on a manuscript that describes our 2018–2019 search results as well as recent improvements to our data processing pipeline. We have already sent the manuscript to co-authors for review and approval, after which we will submit it for publication. We are excited about the work and hope that our co-authors share our enthusiasm!
Warm regards,
Jean-Luc Margot
|
|
