Machine learning, data science, & AI are all disciplines that live in the interplay of the applied statistical sciences & computer engineering. When compared to the significant advances in both the science, hardware & computing over the last 75 years the fundamental ways in which we do this work has changed relatively little. To solve a statistical problem we spend most of our time being engineers, typing code into a terminal or similar environment, & dealing with the problems & abstractions of programmers. In this talk, we will evaluate the tools we use today, the utility of their inherent abstractions, & question how we envision the practice of statistical engineering in the years to come.
From the first punch-card machines to distributed cloud computing well explore the evolution of the statistical engineering process, leaving aside specific machine learning implementations or optimizations. Taking Bret Victors Human Representation of Thought as a starting point well focus on the UX of machine learning & imagine media which allows us to maximize creativity & think the unthinkable. Well look at prototypical examples, compare these to environments customized for other disciplines, & discuss how current engineering frameworks & APIs could be designed to create UX optimized media & environments.
Daniel Krasner is the Founder/CEO or Merriam Tech, a company devoted to improving human interaction with large volumes of information, the co-founder of KFit Solutions, a data science consulting firm, & the co-creator of the Pythons Rosetta, a library for high-performance NLP/text processing. He is also the Director of Data Science in eDiscovery at Paul Hastings, where he develops technology for litigation & related legal applications.
Daniel works in the intersection of machine learning, high performance statistical engineering, human-computer interaction & user experience. He is interested in how we interact with machines, how we build software & engineering solutions, & how this will evolve in the years to come.
Previously, Daniel was the technology lead of the Columbia Universitys History Lab project, chief data scientist at Sailthru, a senior researcher at Johnson Research Labs, a lecturer at the London School of Economics & a professor at Columbia University statistics department. Prior to entering the tech sector, he was an assistant professor of mathematics at UCLA. Daniel holds a PhD in mathematics from Columbia University.