Videos
See AI2's full collection of videos on our YouTube channel.Viewing 211-220 of 251 videos
Improving Structured Prediction With Locally Normalized Models
March 25, 2016 | Ashish VaswaniLocally normalized approaches for structured prediction, such as left-to-right parsing and sequence labeling, are attractive because of their simplicity, ease of training, and the flexibility in choosing features from observations. Combined with the power of neural networks, they have been widely adopted for NLP…
Beyond the Distributional Hypothesis: Learning Better Word Representations
March 9, 2016 | Manaal FaruquiUnsupervised learning of word representations have proven to provide exceptionally effective features in many NLP tasks. Traditionally, construction of word representations relies on the distributional hypothesis, which posits that the meaning of words is evidenced by the contextual words they occur with (Harris…
Deja Vu: The Story of Vision & AI
March 3, 2016 | Ali FarhadiAli Farhadi discusses the history of computer vision and AI.
Beyond Informational Retrieval
March 2, 2016 | Ashish SabharwalArtificial intelligence and machine learning communities have made tremendous strides in the last decade. Yet, the best systems to date still struggle with routine tests of human intelligence, such as standardized science exams posed as-is in natural language, even at the elementary-school level. Can we…
Strategies and Principles for Distributed Machine Learning
February 16, 2016 | Eric XingThe rise of Big Data has led to new demands for Machine Learning (ML) systems to learn complex models with millions to billions of parameters that promise adequate capacity to digest massive datasets and offer powerful predictive analytics (such as high-dimensional latent features, intermediate representations…
Intelligible Machine Learning Models for HealthCare
February 9, 2016 | Rich CaruanaLocally normalized approaches for structured prediction, such as left-to-right parsing and sequence labeling, are attractive because of their simplicity, ease of training, and the flexibility in choosing features from observations. Combined with the power of neural networks, they have been widely adopted for NLP…
Probabilistic Models for Learning a Semantic Parser Lexicon
January 27, 2016 | Jayant KrishnamurthyLexicon learning is the first step of training a semantic parser for a new application domain, and the quality of the learned lexicon significantly affects both the accuracy and efficiency of the final semantic parser. Existing work on lexicon learning has focused on heuristic methods that lack convergence…
Machine Teaching
January 12, 2016 | Patrice SimardFor many ML problems, labeled data is readily available. The algorithm is the bottleneck. This is the ML researcher’s paradise! Problems that have fairly stable distributions and can accumulate large quantities of human labels over time have this property: Vision, Speech, Autonomous driving. Problems that have…
Adding Structure to Unstructured and Semi-structured Data
December 10, 2015 | Chandra BhagavatulaIn this talk, I will describe two systems designed to extract structured knowledge from unstructured and semi-structured data. First, I'll present an entity linking system for Web tables. Next, I'll talk about a key phrase extraction system that extracts a set of key concepts from a research article. Towards the…
Provable Guarantees for Non-convex and Convex Optimization in High Dimensions
November 3, 2015 | Hanie SedghiLearning with big data is akin to finding a needle in a haystack: useful information is hidden in high dimensional data. Optimization methods, both convex and nonconvex, require new thinking when dealing with high dimensional data, and I present two novel solutions.