Words, links, and patterns: novel representations for Web-scale text mining

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=c_k4sQMhuqU


Duration: 1:11:44
44 views
0

Textual data is everywhere, in email and scientific papers, in online newspapers and e-commerce sites. The Web contains more than 200 terabytes of text not even counting the contents of dynamic textual databases. This enormous source of knowledge is seriously underexploited. Textual documents on the Web are very hard to model computationally: they are mostly unstructured, time-dependent, collectively authored, multilingual, and of uneven importance. Traditional grammar-based techniques don't scale up to address such problems. Novel representations and analytical tools are needed. I will discuss several recent contributions related to text mining from a variety of genres. More specifically these include (a) lexical models of the growth of the Web, (b) graph-based entity classification, (c) evolving news summarization, and (d) mining protein interactions in papers. As it turns out, the right representations, when complemented with traditional NLP techniques, turn all of these into instances of better studied problems in areas such as social networks, statistical mechanics, sequence analysis, and computational phylogenetics.



Other Videos By Microsoft Research


2016-09-05Parameterized Model Checking of Protocols: Two Developments
2016-09-05A Sample of Monte Carlo Methods in Robotics and Vision
2016-09-05Virtual Customer Environments & Customer Involvement in Innovation and Value Creation
2016-09-05Large Margin Generative Models
2016-09-05Strategies for Enhancing Ethnic and Gender Diversity in Engineering and Computer Science
2016-09-05DRM and MSFT: a product no customer wants
2016-09-05A Dynamic Pari-Mutuel Market for Hedging, Wagering, and Information Aggregation
2016-09-05Dynamic Point Samples for Free-Viewpoint Video
2016-09-05Fast Belief Propagation for Early Vision
2016-09-05Culture and Prosperity: The Truth About Markets
2016-09-05Words, links, and patterns: novel representations for Web-scale text mining
2016-09-05Asymptotic Enumeration of Spanning Trees via Traces and Random Walks [1/27]
2016-09-05Traffic Constraints Instead of Traffic Matrices: A New Approach to Traffic Characterization
2016-09-05Unknowable
2016-09-05Neighbourhood Component Analysis
2016-09-05Defying Categorization: DXARTS
2016-09-05Complex Arithmetic for Hardware Implementation: Division and Square Root
2016-09-05Distributed Implementations of Vickrey-Clarke-Groves Mechanisms
2016-09-05Not Even Wrong
2016-09-05Measurement and Monitoring in Wireless Sensor Networks
2016-09-05From TimeSync to EmStar: What's really hard in sensor networks?


Tags:
microsoft research