Automatic Programming: How Far Can Machines Go • Hila Peleg • YOW! 2018

This presentation was recorded at YOW! 2018. #GOTOcon #YOW Hila Peleg - Doctoral Researcher in the Department of Computer Science at the Technion ABSTRACT Program synthesis is the wild west beyond code generators and onward toward self-programming systems. What can it do for us? The answer, right now, is not a lot. It can harness the wisdom of the crowd to help reproduce scenarios that are repeated often, and in narrow scopes tools can make educated guesses about what the programmer intended, but full automatic program synthesis is likely impossible. Specifically because of this, it becomes most interesting to explore what it can’t and won’t be able to do on its own. Self-programming systems might be past the horizon, but getting a human more involved makes a big difference, and while the computer might never program for us, it might make for a decent pair-programming partner. We’ll look at glimpses of this future, in academia and outside it, and see why they’re more IDE plugins and less the robot apocalypse. Hila Peleg is a doctoral researcher in the Department of Computer Science at the Technion. Hila’s research entails designing synthesis tools to help programmers navigate new programming languages and unfamiliar APIs. Prior to her graduate degree, Hila spent over a decade as a developer on enterprise server projects, and she also holds a degree in literature. [...] TIMECODES 00:00 Intro 03:48 Program synthesis to the rescue 04:47 What if we have automatic programming? 13:29 Query knowledge about some kind of code 15:33 Commit Strip said it best... 17:26 Adjusting our expectations 20:05 What does this all mean? 20:35 Realistic expectations for realistic synthesis 22:37 The synthesis engine 29:22 Trying a different strategy altogether 31:53 The Interaction Model 33:39 Specifying (and re-specifying) intent Task: find the median of a list 35:08 Programming Not only by Example 38:31 When models “out-think“ the programmer 40:19 Keep understanding what’s going on 44:35 Read-Eval-Synth Loops RECOMMENDED BOOKS Christiansen, Foy, Wall & Orwant • Programming Perl • Gamma, Helm, Johnson & Booch • Design Patterns (Gang of Four) • Lee, Wickens, Liu & Boyle • Designing for People • Phil Winder • Reinforcement Learning • Holden Karau, Trevor Grant, Boris Lublinsky, Richard Liu & Ilan Filonenko • Kubeflow for Machine Learning • Kelleher & Tierney • Data Science (The MIT Press Essential Knowledge series) • Lakshmanan, Robinson & Munn • Machine Learning Design Patterns • #AutomaticProgramming #NoCode #LowCode #AI #ML #HilaPeleg #Programming #Synthesis #AITutorial #MLTutorial #AIModels #YOWcon Looking for a unique learning experience? Attend the next GOTO conference near you! Get your ticket at Sign up for updates and specials at SUBSCRIBE TO OUR CHANNEL - new videos posted almost daily.