This Gaming Life
Jason Shankel, Roblox
10/12/2017
In this talk Jason Shankel, a 25-year veteran of the gaming industry, will describe the changes the industry has gone through over the past three decades and what it takes to make it in the world of gaming today
Natural Language Processing For Fake News Detection
William Wang, University of California, Santa Barbara
10/19/2017
In this past election cycle for the 45th President of the United States, the world has witnessed a growing epidemic of fake news. The plague of fake news not only poses serious threats to the integrity of journalism, but has also created turmoil in the political and actual world. However, statistical approaches to combating fake news has been dramatically limited by the lack of labeled benchmark datasets. In this talk, we will describe LIAR: a new, publicly available dataset for fake news detection. We collected a decade-long, 12.8K manually labeled short statements in various contexts from POLITIFACT.COM, which provides detailed analysis report and links to source documents for each case. This dataset can be used for fact-checking research as well. Notably, this new dataset is an order of magnitude larger than previously largest public fake news datasets of similar type. Empirically, we investigate automatic fake news detection based on surface-level linguistic patterns. We have designed a novel, hybrid convolutional neural network to integrate metadata with text. We show that this hybrid approach can improve a text-only deep learning model. We will outline future directions, and conclude this talk by discussing related technologies in natural language processing
Simulating Just Enough Non-Linearity
Theodore Kim, Pixar, Emeryville
10/26/2017
Most of the physical phenomena that are relevant to computer animation are inherently non-linear. These include the equations governing the flow of smoke and water, as well as the dynamics of skin and flesh. Which of these non-linearity’s are visually important, and which just introduce unnecessary trouble? In this talk, I will examine a few case studies.
Framenet And Natural Language Processing
Miriam R. L. Petruck, International Computer Science Institute, UC Berkeley
11/02/2017
This talk presents an overview of FrameNet, a research project in corpus-based computational lexicography, based on the principles of Frame Semantics (e.g. Fillmore 1985, inter alia). FrameNet's initial goals included providing information about the valences, i.e., the semantic and syntactic combinatorial possibilities for the vocabulary of contemporary English, and documented by corpus findings. Initially developed primarily as a lexicographic effort, FrameNet data are used in a range of Natural Language Processing (NLP) tasks, including Semantic Role Labeling (e.g. Das et al. 2014, Roth & Lapata, 2015, Roth 2016), Question Answering (e.g. Sinha 2008, Ofoghi 2009, He et al. 2015) Information Extraction (e.g. Harabagiu & Bejan 2010, Zhang et al. 2010, Søgaard et al. 2015), and Sentiment Analysis (Ruppenhoffer & Rehbien 2012, Ruppenhoffer 2013 ), to name a few. In addition to reporting on recent developments in FrameNet, this presentation will address the implications of these developments for NLP.
Keys, Hollywood, And History: The Truth About Icann, Dnssec, And The Root Key
Richard Lamb, ICANN
11/09/2017
For better or worse, Internet security has gained notoriety recently and with it greater interest in some humble key management operations. Specifically, much hay has been made by Hollywood of the role of those incorrectly referred to as the “7 key holders” of the Internet and ICANN. As the original architect of the DNSSEC root key management system, Rick will describe the truth, unwritten or otherwise, behind this humble, trusted operation and how it came to be. Along the way he hopes to trick participants into learning about DNS, DNS Security Extensions, and how we might all benefit from innovation on this infrastructure.