People / Walter Chang
Walter Chang
Principal Scientist
I joined Adobe’s Advanced Technology Labs in 2004. I am currently the principal investigator in two areas of research: advanced data analytics for online recommendation systems, and the semantic understanding of multimodal content. In my role as the architect and project manager of a content intelligence system, our focus has been the development of hybrid symbolic and statistical algorithms to perform deeper semantic analysis of text, audio, and video content for advanced search, content organization, automatic metadata extraction, the synchronization of temporal metadata, and metadata inferencing.
My related research interests include: computational linguistics, automatic ontology acquisition and generation, named entity recognition, and extending relational, entity-relationship, and document object models to represent complex knowledge structures. I was a contributor and Adobe representative to the W3C RDF-Schema Working Group and have published in the areas of Semantic Web knowledge representations, extensible database architectures, and document indexing methods using content signature coding.
Prior to Adobe ATL, I was a managing director and systems architect at Scient (now Razorfish) and previously an engineering manager at Oracle Corporation. From 1984–1989, I was an advisory level programmer in the Extensible Database Group at the IBM Almaden Research Center where I had the privilege of being mentored by Prof. Hans Schek, Dr. Guy Lohman, and Dr. Bruce Lindsay.
I received my BS degree in Electrical Engineering and Computer Science from U.C. Berkeley and also hold a Masters degree in Computer Architecture from Santa Clara University.
Projects |
A Text Analytics-Based Contextual Analysis Engine
An Advanced Data Analytics and Recommendation Engine
A Multimedia Content Intelligence Platform for Video Production Environments
Extraction, Analysis and Time-Synchronization of Multimodal Metadata with Video
Multicore Speech-to-Text Performance and Accuracy Enhancements
A System for the Semantic Analysis of Documents
Patents |
USPTO #8,060,506: A System For Automatically Selecting Document Categories and Keywords For Matching Ads Provided by Ad Aggregators, W. Chang, issued: Nov. 15, 2011.
USPTO #7,873,640 - Method for Semantic Analysis of Documents to Select and Rank Contextual Advertising Keywords, W. Chang, issued: Jan. 18, 2011.
USPTO #7,644,052 - System and method for building ontologies to organize, search and classify documents. W. Chang, N. Ghamrawi, issued: Jan. 5, 2010.
USPTO #7,610,315 - System & method for determining and recommending document retention and security polices. W. Chang, L. Masinter, issued: Oct. 27, 2009.
USPTO# 7,580,918 - Method for efficiently representing and searching directed acyclic graph structures in databases. W. Chang, N. Ghamrawi, A. Swami, issued: August 25, 2009.
USPTO #5,737,736 - Oracle Corporation. Method and apparatus for storing resource objects using c-structure bind descriptors. W. Chang, issued: Apr. 7, 1998.
USPTO #5,319,779 - IBM Almaden Research Center. System for searching information using combinatorial signatures. W. Chang, H. Schek, issued: Jun. 7, 1994.
Patents Pending Issuance:
US Patent App. #13/039,177: Adobe Systems, Automatically Classifying Consumers in Marketing Micro-Segments for Targeted Delivery of Content, G Baum, W. Chang, filed: March 2, 2011.
US Patent App. #13/039,207: Adobe Systems, Method and System for Machine Definition and Specification of Marketing Micro-Segments, W. Chang, G. Baum, filed: March 2, 2011.
US Patent App. #13/039,242: Adobe Systems, Sequential Engine that Computes User and Offer Matching Membership into Marketing Micro-Segments, W. Chang, G. Baum, filed: March 2, 2011.
US Patent App. #13/039,257: Adobe Systems, A Scalable Engine that Computes Marketing Micro-Segments, W. Chang, G. Baum, filed: March 2, 2011.
US Patent App. #13/039,266: Adobe Systems, Persistent Metadata for User Controlled Policy for Advertising, G. Baum, W. Chang, filed: March 2, 2011.
US Patent App. #13/153,328: Adobe Systems, A System and Method to Automatically Expand an Ad Offer Inventory and Provide Automatic Offer Matching Based on Explicit and Implicit Consumer Interests, W. Chang, G. Baum, filed: June 3, 2011.
US Patent App. #12/789,708: Adobe Systems, A System to Time Align Movie Scripts to Video and Audio Content. W. Chang, filed: May 28, 2010.
US Patent App. #12/789,785, Adobe Systems: Movie script to text alignment enhancement by nonlinear interpolation, J. Scoggins, W. Chang. filed: May 28, 2010.
US Patent App. #12/789,791, Adobe Systems: Movie Script Align and Replace (automatic text replacement based on confidence scores) and related GUI, D. Kuspa, J, Scoggins, W. Chang, filed: May 28, 2010.
US Patent App. #12/789,720, Adobe Systems: Movie Script and Replace Umbrella Feature including workflows and enhancements, J.Scoggins, W. Chang, D. Kuspa, P. Ellis, C. Van Winkle, S. Hayhurst, filed: May 28, 2010.
US Patent App. #11/946,680: Adobe Systems, System to automatically generate unique global named entity identifiers to resolve named entities in a federated text mining environment. W. Chang, M. Welch, filed: Oct. 9, 2009.
US Patent App. #12/713,008: Adobe Systems: A System for Automatically Capturing, Converting, and Analyzing Movie Scripts. W. Chang, M Welch. filed: July 2009.
US Patent App. #12/471,814: Adobe Systems. An efficient algorithm to generate layout preserved text from PDF. E. Koh, W. Chang, filed: May 26, 2009.
US Patent App. #12/618,353: Adobe Systems, Multimodal Video Indexing, Search, and Content Disambiguation. W. Chang, M. Welch, filed: Sept. 3, 2009.
US Patent App. #12/332,297, A System for Creating Speech-to-Text Language Models for Generic Audio Content, W. Chang, M. Welch, filed: Dec. 10, 2008.
US Patent App. #12/332,309: Adobe Systems, A Multi-core Processing System For Performing Parallel Speech-to-Text Transcriptions, W. Chang, M. Welch, filed: Dec. 10, 2008.
US Patent App. #12/130,607: Adobe Systems. System for automatically extracting reading-order text and semantic entities from unstructured and semi-structured documents. E. Koh, W. Chang, filed: May 30, 2008.
US Patent App. #12/035,617: Adobe Systems. A system for locating and visualizing people expertise and document resources & their inter-relationships. W. Chang, N. Sakunkoo, filed: Feb. 22, 2008.
US Patent App. #12/035,628: Adobe Systems, A system for creating, visualizing and navigating semantic networks of information. N. Sakunkoo, W. Chang, filed: Feb. 22, 2008.
US Patent App. #11/728,069: Adobe Systems. A method of applying keywords from the analysis of documents to analyzed documents to facilitate ad content metadata insertion. W. Chang, N. Ghamrawi, filed: Mar. 23. 2007.
US Patent App. #11/365,198: Adobe Systems. A method for providing a set of information resources to assist in researching semantic entities in a document. W. Chang, filed: Feb. 28, 2006.
Invited Talks |
W. Chang. 2010. Adobe® Story and Creative Suite 5 – New Semantic Technologies for the Motion Picture and Video Industry . Semantic Technology Conference, June 24, 2010, San Jose, CA.
W. Chang. 2009. An Ontology-Driven System for the Semantic Analysis of Multimedia Content . Semantic Technologies Conference, June 17, 2009, San Jose, CA.
W. Chang, T. Tague. 2008. Emerging Semantic Ad Platforms . SDForum Semantic Web Special Interest Group, Palo Alto, CA.
W. Chang. 2008. The Design of a Semantic Analysis Platform for a Contextual Advertising Application. 2008 Semantic Technologies Conference Session, San Jose, CA.
W. Chang, D. Bedford. 2007. Semantic Content Management Systems. 2007 Semantic Technologies Conference Session, San Jose, CA.
Launched Websites |
Ads for Adobe PDF - an online document analysis and contextualized advertising system for ad enabling PDF content in Adobe Reader using the Yahoo! ad network.
MLB.com - Major League Baseball Advanced Media Gameday - was the chief architect that led the engineering team that designed, built, and deployed a real-time multimedia architecture for MLBAM's live game visualization, player stats, audio and video feeds, live player event capture, event synchronization to video, online video search, and event replay.
Sixdegrees.com - an early dot.com online social networking community, subsequently acquired by Youthstream Media Networks for $125M
Publications |
PixelTone: A Multimodal Interface for Image Editing
Laput, G., Dontcheva, M., Wilensky, G., Chang, W., Agarwala, A., Linder, J., Adar, E., 2013 (Apr 27, 2013)
To appear in Proceedings of CHI : ACM Conference on Human Factors and Computing Systems
RDF Vocabulary Description Language 1.0: RDF Schema, W3C Recommendation
N. Arnett, D. Brickley, W. Chang, et. al. (Jan 31, 2004)
World Wide Web Consortium
A Discussion of the Relationship Between RDF-Schema and UML
Chang, W (Jan 31, 1998)
RDF-Schema Working Group W3C Technical Note. World Wide Web Consortium. 1998
Starburst Database System Mid-flight: as the Dust Clears
Haas, L., Chang, W. et. al. (Jan 31, 1990)
IEEE Transactions on Knowledge and Data Engineering, Vol. 2. No. I , ISSN: 1041-4347
A Signature Access Method for the Starburst Database System
Chang, W. and Schek, H. (Jan 31, 1989)
Proceedings of the 15th International Conference on Very Large Data Bases. Amsterdam, The Netherlands, Pages: 145 – 153, 1989, ISBN:1-55860-101-5, Morgan Kaufmann Publishers Inc. San Francisco, CA, US
