Second Quarter Session
At the Machines + Media Working Group, you’ll meet like-minded, data-driven peers.
Join us for a session—to be held each quarter in 2019—to collaborate, share ideas and tricks of the trade, engage in joint projects, and work to build a repository of open source code that can help advance the media industry.
NYC Media Lab has built a community of industry executives and technologists, faculty, students, and entrepreneurs each pursuing applications of data science to media. Advances in artificial intelligence and machine learning, natural language processing, computer vision, and a variety of other technologies are among the most compelling developments for media companies, offering possibilities for the creation, production, distribution and monetization of media content and interactive experiences.
RSVP to attend
Agenda for june 11, 2019
9:00AM: Breakfast & Networking
Get to know the group members over coffee & breakfast.
Address: Viacom: 1515 Broadway, New York, NY (31st Floor, Room 1)
9:30AM: Introduction and Overview
About the 2019 Machines + Media Working Group and ways to collaborate.
9:45AM: What's State of the Art?
Hear from two leading university researchers, who will share their technical demonstrations and expertise in emerging data science applications.
Michael I Mandel, Associate Professor of Computer and Information Science at Brooklyn College and the CUNY Graduate Center. Mandel works at the intersection of machine learning, signal processing, and psychoacoustics. His work is supported by the National Science Foundation, including via a CAREER award, the Alfred P. Sloan Foundation, and Google, Inc.
About Mandel’s Talk: Very High Quality Speech Enhancement via Speech Synthesis
Noise and reverberation are two of the biggest problems for voice communication technologies like conversational assistants, hearing prostheses, and mobile communication. Traditional approaches mitigate these interferences by modifying the noisy signal, which leads to two problems: under-suppression of noise and over-suppression of speech. As an alternative, we extract from the noisy speech enough information to drive a speech synthesizer, which then generates a high quality, noise-free rendition of the speech from scratch to replace the original noisy speech. In listening tests, the system's output is rated as being higher quality than state of the art speech enhancement methods and as producing more natural pronunciations than traditional text-to-speech systems.
Enrico Bertini, Associate Professor, Computer Science and Engineering, NYU Tandon School of Engineering. His research focuses on the study of effective data visualization methods and techniques to explore and make sense of large and often high-dimensional data. Bertini's work has been applied to several application domains including: biochemistry, cybersecurity, development, healthcare, climate science, human rights.
About Bertini’s Talk: Data Visualization for Interpretable and Interactive Machine Learning
The recent advancement of machine learning has made possible dreaming again of machines helping us do remarkable things with data: detecting diseases with higher accuracy than humans, creating self-driving cars, making smarter business decisions, etc. Increased automation however comes at a cost: as we outsource more and more of traditionally human-driven tasks to automated systems we have the responsibility to understand how they work. What kind of decisions does my model make? Do they make sense? When does it fail? Why does it fail? How can I fix it? These are only some of the questions we should be able to answer. Data visualization, with its power to turn information into insightful pictures can play a major role in this area. In my talk I will introduce the problem of machine learning interpretability and I will describe solutions we are developing in my lab at NYU Tandon, where we develop advanced data visualization systems to help people understand machine learning models
10:30AM: What’s in Market? Moderated Q+A with Michael Mandel and Enrico Bertini
In a moderated Q+A session led by NYC Media Lab, attendees can learn how these new technologies are changing markets.
11:00AM: Roundtable Discussion: What’s Top of Mind?
Engage in a group discussion on top priorities for applications of data science in media.
12:00PM: Event closes
Working group Recap & Related links
The first quarter session was held on Friday, February 22 at The Associated Press. The event included:
A research presentation from: Lydia Chilton, Professor in the Computer Science Department at Columbia University. Her work focuses on crowdsourcing and decomposing hard design problems so that people and machines can collaborate to solve them. She presented VisiBlends, a data system which uses computational techniques and natural language processing to auto-generate visual blends that can pass human visual object recognition.
A startup presentation from: Rungson Samroengraja, President & COO of Satisfi Labs. Satisfi Labs is a conversational AI platform. Its Answer Engine powers chatbots, voice experiences, messaging apps and website forums to provide an answer to all customer questions in real time.
A startup presentation from: Armando Kirwin, Co-Founder of Artie. Artie helps content creators bring virtual characters to life on mobile. Its network shares insights and analytics in real time, allowing creators to understand how consumers interact with each character.