Laney Moy is a rising senior at Thomas Jefferson High School for Science and Technology (TJ) in Alexandria, Virginia, where she took her first computer science classes and began to love coding. In addition to computer science, she is passionate about dance, Latin, and working with younger students. Laney is entering her second season as captain of TJ's Varsity Dance Team and her second year as Head Choreographer of TJ Latin's elaborate International Night performance. She has won several awards for Latin, and she volunteers as a tutor for lower-level TJ Latin students. Laney also enjoys math and is an instructor at Mathnasium, where she gets to do math all the time. In her free time, Laney likes to read books and knit. She is currently working on her fourth scarf.
Many significant works of Latin poetry follow the format of dactylic hexameter, meaning that each line is composed of some combination of six metrical feet, each of which is either two long syllables or a long syllable followed by two short syllables. Scansion is the process of identifying the pattern of syllable lengths. Through this project, I use machine learning to scan lines of Latin poetry in dactylic hexameter.
Summary of Results
I trained a sequence-to-sequence neural network to find the "length" of each character in a line. The input was the plain text line, and the output was a list of probabilities that each character will be a long vowel; short vowel; or consonant, punctuation mark or ignored vowel. The final neural network had an error of 2.9%, and a character is incorrectly scanned around once every few lines.
In the future, markings between metrical feet could be added to the displayed lines. The program could also be extended to other types of meter or poetry in other languages.