From Our Mule Sal to AI’s Search and Learn (SaL) - 200 Years of Canals Excellence
From Months to Minutes to Moments: SaL and the New Way to Search Vital, Ancient Canals Documents
The History
The initial construction for the Erie Canal system lasted from 1817 through 1825. But in order to accommodate bigger barges, the canal was enlarged most recently in some sections between 1905 and 1918 to produce the Barge Canal.
In order for this upgrade to be successful in an analog world, ‘human computers,’ i.e., professionals with uncanny mathematical skill, had to produce diagrams, plans, drawings, maps and other paper documents to assure the upgrades were successful. Many of these older documents are beautifully rendered and are works of art on their own.
More than 65,000 pieces of paper documentation were created - known collectively as Final Estimates - and stored in a variety of buildings across New York State for most of the 20th century. This paperwork endures and serves both the New York State Canal Corporation and the NYPA workforce as maintenance and improvements continue into the 21st century.
The Modern Day Issue
While vital to the Canals’ employees who use the documents to maintain locks and other physical assets, these documents were not easily accessible, and a massive project was undertaken with the help of the New York State Archives to digitize all 65,000 of them. This took years, beginning in 2018 and continued through the COVID pandemic. It was ultimately successful and reduced the time for a document search from months to minutes.
The AI Addition
In 2022, after digitization, the next huge leap was taken: building an AI application to crawl over the digitized document database to provide staff with any specific document they need, now in moments rather than minutes.
NYPA’s staff of computer programmers, data scientists, knowledge managers, and AI leaders worked with subject matter experts in Canals to develop SaL, short for Search and Learn. SaL is also named in honor of the Erie Canal’s legendary mule from the song “Low Bridge.”
Future Promise
SaL was developed to help both Canals and NYPA staff find knowledge within the newly digitized documents more efficiently and effectively.
While still a work in progress, SaL's capabilities are impressive, cutting the time needed for a document search dramatically.
And while SaL works for Canals right now, it could potentially be deployed to NYPA and cut the time necessary in searching for Operations-related documents to a fraction of what they are, making some of the work performed by Operations staff far less time consuming and help with the maintenance of transmission and generation assets.
"These Canals documents are not only vital to the ongoing maintenance and operation of the Canals, but they are truly beautiful works of art too," commented Chuck, the SaL project lead.
He continued, “It was truly awesome to watch SaL's development cut the amount of time needed to search for and find these records, down from months to almost no time at all. The AI was even taught to read cursive penmanship and categorize each document. Its future applications on the NYPA Operations side are really promising and amazing."
SaL's popularity is growing, and Chuck will be presenting SaL at the World Canals Conference in Buffalo this September. The presentation is titled ‘Looking Forward and Back: Using AI to Enable the Past to Speak to the Present to Ensure the Future of the NYS Canal System.’
We at NYPA and Canals hope to see more of this particular SaL in a variety of different new roles!