By Brett Daniel Shehadey
Special Contributor for In Homeland Security
Rice University is one of a small group of institutions receiving funding of $11 million by the Defense Advanced Research Projects Agency (DARPA) in order to develop an auto-correct capability feature for programmers within four years. The program is called PLINY, named after Pliny the Elder, who was the ancient Roman naturalist credited for writing the first encyclopedia. PINY will act as a lexicon of hundreds of billions of lines of open-source computer code and use Bayesian statistics to predict the expressed intent of the programmer and autocomplete the coding process assisting in software development.
PLINY is a bold attempt of predictive software analytics that will, in a crude analogy, function and progress similar to the advent of predictive typing, when texting. One might expect many of the same hurdles as the development of predictive typing for predictive coding, but just having surpassed those hurdles to a degree of natural efficiency on mobile device platforms is encouraging for other pursuits such as PLINY.
PLINY is a related to DARPA’s larger effort called Mining and Understanding Software Enclaves (MUSE). MUSE will analyze searchable database of properties, behaviors and vulnerabilities. From a defense perspective, this could be seen as a cybersecurity tool, in which these programs or similar programs will aid in in checking for bugs or security weak points within software coding. The reason for this is to automate more of the grunt work and make the overall coding more precise.
But the key benefits of PLINY will likely have both civilian and military cross applications. Think of DARPA’s Speech Understanding Research (SUR), for example, in the 1970s. It took a while but eventually, America’s lead in speech recognition technology is at the point that it is becoming the standard model, replacing human operators, aiding hands-free human-computer interaction, etc.
“Software today is far more complex than it was 20 years ago, yet it is still largely created by hand, one line of code at a time,” said co-PI Swarat Chaudhuri, assistant professor of computer science at Rice. “We envision a system where the programmer writes a few of lines of code, hits a button and the rest of the code appears. And not only that, the rest of the code should work seamlessly with the code that’s already been written.”
Over two dozen computer scientists from Rice, the University of Texas-Austin, the University of Wisconsin-Madison and the company GrammaTech will participate in programing the auto-correct programing engine PLINY. Eventually, the inevitable goal will be to get programs to program other programs. PLINY is the forefront of more intuitive software development tools, perhaps using more general commands in the future. Looking decades ahead, humans may eventually develop and program software primarily through expressed simple verbal or neural commands alone. PLINY is a natural pathway to that reality.
Comments are closed.