SEAS Teams Dominate at NYC Media Lab Summit

 

The grand prize was awarded to CLUE, a team led by PhD student Shyuan Yang (center).

Teams from Columbia Engineering won the grand prize ($1,000), as well as second ($500) and third place prizes ($250) at NYC Media Lab's Annual Summit this fall. Billed as “your crash course in the best thinking, projects, and talent in digital media and communications in NYC and beyond,” the event showcased more than 100 creative coders, product designers, data scientists, and makers from New York City’s universities and more. They presented their research and prototypes to a packed crowd of over 800 participants, including influencers from leading digital media, technology, and communications companies.

The grand prize was awarded to CLUE, a team led by PhD student Shyuan Yang, under the direction of Ioannis (John) Kymissis, associate professor of electrical engineering, for their demo, "Ultra-low cost sensor platform and integration." The demo featured two projects that explore the functionality of low cost electronics in sensing and Internet of Things (IoT) applications. The sensor can be integrated with radio communication for a variety of applications such as plugSTRATE, a low-cost (less than $4) wireless monitoring network system designed to address the needs of the rapidly expanding energy audit market. They also demoed Dollar Sensor, a cheap (less than $1) chemical sensor that uses a direct USB connection and printable sensor materials to eliminate the need for a reader or user interface while delivering digital readouts of chemicals. On this project, the team used a humidity sensor; they are currently developing other sensors, including one that detects glucose, another that senses ammonia vapor, and one that identifies ketones. CLUE team members are Fabio Carta PhD’15, PhD candidate Johannes Bintinger, first-year student Sejal Jain, and Rigers Qeraj from Radiator Labs, a start-up cofounded by Kymissis.

The AMuSe project was motivated by the need to wirelessly deliver video to mobile devices in crowded venues, for instance, to viewers in a football stadium who want to see a replay of a recent touchdown on their phones.

The AMuSe ( Adaptive Multicast Services) Project, a joint SEAS/Bell Labs project, won a second place prize for demonstrating a system for multimedia content (e.g., video) delivery via Wi-Fi to very large groups. The project was motivated by the need to wirelessly deliver video to mobile devices in crowded venues, for instance, to viewers in a football stadium who want to see a replay of a recent touchdown on their phones. Wi-Fi support for multicast, sending the same message to several devices, is limited and so PhD students Varun Gupta and Craig Gutterman, working with their advisor Gil Zussman, associate professor of electrical engineering, and Yigal Bejerano from Alcatel Lucent Bell Labs, have been developing a system that collects accurate feedback (channel quality) from a small number of users and adapts the transmission accordingly. The system was implemented and evaluated in the ORBIT testbed at Rutgers University with over 200 Wi-Fi devices. SEAS undergraduates Raphael Norwitz and Savvas Petridis developed an interactive web-based application to show how the AMuSe system operated, based on traces collected from the real experiments, and to demonstrate the impact of various settings on the video quality at the different nodes. The students used their application to demonstrate the broad capabilities of the AMuSe system at the NYC Media Lab event.

Smaranda Muresan and Debanjan Ghosh won third place for their natural language processing project. (Not pictured: Weiwei Guo)

A third place prize went to a natural language processing project, “Sarcastic or Not: Word Embeddings to Predict the Literal or Sarcastic Meaning of Words.” The ability to detect sarcasm is crucial to understanding what’s really being said. Faced with the statement, “I love going to the dentist,” an automated sentiment-analysis system can have trouble telling whether the speaker likes going to the dentist or hates it. Under the direction of Smaranda Muresan, a research scientist at the Center for Computational Learning Systems at Columbia, Debanjan Ghosh, Muresan’s PhD student at Rutgers University, and Weiwei Guo SEAS PhD’15 are looking at sarcasm-detection as a word-disambiguation problem, where the literal and sarcastic meaning of “love” is determined by its context. Using Twitter data, they investigated several distributional semantic techniques showing a 10 percent improvement over current approaches. 

The winning projects represent the creativity, technical depth, and potential impact of the exciting ideas emerging from faculty and students across NYC universities.

“It’s very difficult to pick winners in a demo pool as broad and deep as this one,” said Justin Hendrix, executive director at NYC Media Lab. “These prizes are intended as a gesture of support for the teams that shared their work at the Summit, and as encouragement for the continued development of their projects, prototypes, and research.”

The full list of winners can be found here.

Original article here

—by Holly Evarts


500 W. 120th St., Mudd 1310, New York, NY 10027    212-854-3105               
©2014 Columbia University