Inspiration
We were inspired to write Age Translator after one of our group members heard one of his father's friends complaining that he couldn't understand the way 'young people' speak, and that there wasn't any way to translate these phrases so he could understand them.
What it does
Age Translator will take any input phrase and translate it to an appropriate, understandable phrase for the selected age, from newborns to centenarians.
How it works
The user inputs a phrase. The phrase is then parsed into words and punctuation. The words are then translated if relevant translation are found in the dictionary. Loops are used to translate the words one step at at a time, until the correct translation has been selected. The phrase is then reformed using the newly translated words and is output to the user.
Challenges we ran into
Significant challenges were encountered while writing the translation system. The first one was - What data structure would be useful to translate in both directions, from arbitrary text into an arbitrary age? A closely related challenge turned out to be the algorithm which could adapt to a wide variety of inputs versus age outputs - We had to devise an algorithm that would be able to 'hop' over a number of intermediate phrases which would be appropriate translations for an audience in between the source and the target audience.
The final notable challenge is one of vocabulary, and we consider it impossible to solve perfectly. Simply put, it's not possible to catalog all phrases which should be translated from one age to another, and it's not possible to establish a single age where any particular phrase should switch over for universal application. Essentially, we do not have enough data to build the dictionary, and our ability to appropriately choose switch-over ages between terms is limited by our own experience with these terms.
Accomplishments that we're proud of
We considered the challenges of potential insensitivity with some of our translations. We made sure that our translations were not ridiculing any particular age group. We set out to build translations which endear popular culture of the past as well as the present. Many of our translations were created from our own personal experiences hearing phrases from our own families.
What we learned
We learned that emulating natural language processing to satisfaction is a difficult project. Our implementation resulted in good behavior, but is limited in its ability to detect certain input phrases, produce grammatically correct output, and avoid removing certain punctuation elements.
What's next for Age Translator
Our team looks forward to adding more phrases and translations, and potentially expanding Age Translator's pattern recognition and natural language production.
Log in or sign up for Devpost to join the conversation.