By Bill Slawski, on February 8th, 2006Google has added another top search scientist to their team.
The Seattle Post-Intelligencer noted yesterday that Udi Manber, chief executive of Amazon’s A9, will be leaving the Seattle-based company to become a Vice President at Google.
Both John Battelle and Gary Price have more on the move.
Udi Manber was a professor of Computer Science at the University of Wisconsin, and the University of Arizona. During his academic career, he co-developed a number of popular search software packages.
He joined Yahoo! as their chief scientist in 1998.
Dr. Manber moved to Amazon.com in 2002, where he was chief algorithms officer. In that role, he worked on such things as Amazon’s “Search Inside the Book” project. There’s a great writeup of a presentation he gave shortly after joining Amazon, as part of the Jon Postel Distinguished Lecture Series at UCLA called Udi Manber: The First 10 Years on the Web
He became the CEO of A9.com in November 2003, working upon their search technologies. It appears that Google had their eyes on him back in 2003, from this ZDnet article: Google, Amazon in a war of search words, where it’s noted that ads for Google jobs appeared on search results when someone would perform a search for “Udi Manber”.
There’s an interesting question and answer session with Udi Manber from the Washington Post this last April: The Future of Web Search. One of the questions asked was “What is the key advantage to using A9 vs. Google?”
In addition to writing more than 50 articles on search related topics, and a book on algorithms, Dr. Manber has been granted a number of patents, and has more patent applications pending. To gain a sense of the scope of the work he has done in the past, I went to the US Patent Office pages and looked those up. A number of them are linked to below, with the abstracts from the documents. I’ve listed patents first, and then patent applications.
Some of the patents granted to Udi Manber as a co-inventor
Inventors: Daniel Alexander Ford, Tobin Jon Lehman, Qi Lu, Udi Manber, and Ron Yair Pinter
Assigned to International Business Machines Corporation
US Patent 6,480,830
Granted November 12, 2002
Filed March 27, 1998
Abstract:
An “active” calendar automatically analyzes a user’s calendar entries and sends machine-readable messages to destinations appropriate to content of the calendar entry. A group of event categories is established, each category specifying one class of anticipated calendar entry.
An action rule database pre-associates each event category with one or more message formats each having a content and a destination. The action rule database also contains data identifying sources containing the content and destination for each message format. These sources include records of the action rule database itself, subparts of calendar entries of the pre-associated event category, one or more other databases, or a combination of the foregoing.
After the calendar receives a user-submitted computer calendar entry describing a planned event, it identifies one of the event categories of the established group that classifies the planned event. For each message format pre-associated with the identified event category, the calendar determines the content and destination for the message as specified by the action rule database, and transmits the message to the destination.
Detecting similarities in Java sources from bytecodes
Inventors: Brenda Sue Baker and Udi Manber
Assigned to Lucent Technologies Inc.
United States Patent 6,282,698
Granted August 28, 2001
Filed December 4, 1998
Abstract:
A technique for detecting similarities in large sets of binary code files, e.g., bytecode files, without requiring access or knowledge of the actual source code itself. In accordance with the technique, bytecode files are disassembled and preprocessed using positional encoding to prepare the disassembled bytecode files for use in conjunction with similarity detection tools.
Systems and methods for matching participants to a conversation
Inventors: Udi Manber, Chi Chao Chang, and Anthony Lamarca
Assigned to Yahoo! Inc.
US Patent 6,651,086
Granted November 18, 2003
Filed February 22, 2000
Abstract:
Systems and methods for connecting two or more individuals to an Internet conversation based on their mutual interests, the current content they may be viewing and what they want to talk about at that time. The techniques of the present invention allow an individual who starts a conversation to maintain full control over who is able to join that conversation as well as how many are able to join at any one time.
A user who desires to start or join a conversation about a particular topic or story selects an indicator, such as an icon, associated with the specific topic or story. The user is presented with an option to start or join a conversation. If the user opts to start a conversation, the user is presented with a comment page, and the user enters a comment, or comments, that preferably is intended to spark an interest in other users. The comment is then presented to other users.
Those users who may desire to join in a conversation with the conversation starter respond with their own comment, which is then sent to the conversation starter. The conversation starter reviews the comment, and if the comment is subjectively acceptable, the conversation starter brings the responding user into a conversation. The responding user and conversation starter are then connected in a messaging session, such as an instant messaging session. Multiple users may be connected in a single messaging session by the conversation starter in this manner.
System and method for personalized information filtering and alert generation
Inventors: Matthias Eichstaedt, Ashvinkumar P. Patel, Qi Lu, Udi Manber, and Kristine Rudkin
Assigned to Yahoo! Inc.
US Patent 6,381,594
Granted April 30, 2002
Filed July 11, 2000
Abstract:
A search engine that forms a compact representation of a plurality of user queries to efficiently find desired information in an information network. The search engine comprises a profile processor having logic to receive the queries from the users and a search module.
The search module is coupled to the profile processor and has logic to receive the information content, to combine the user queries into a master query, and to match the master query with the information content to determine matching content. The search engine also includes logic to analyze the matching content to determine if any of the queries has been satisfied.
Systems and methods for identifying and extracting data from HTML pages
Inventors: Udi Manber and Qi Lu
Assigned to Yahoo! Inc.
US Patent 6,920,609
Granted July 19, 2005
Filed August 24, 2000
Abstract:
Systems and methods for analyzing HTML formatted web pages to automatically identify and extract desired information. A computer algorithm identifies and extracts different pieces of information from different web pages automatically after minimal manual setup.
The algorithm automatically analyzes pages with different content if they have the same, or similar, formats.
Processing of unsolicited bulk electronic mail
Inventors: Brian R. Woods and Udi Manber
Assigned to Yahoo! Inc.
US Patent 6,965,919
Granted November 15, 2005
Filed December 1, 2000
Abstract:
The present invention involves detecting unsolicited electronic mail distributed in bulk. In one embodiment, a method for automatically processing electronic mail loads an electronic mail message. Non non-textual information is removed from the electronic mail message.
A first portion from the electronic mail message is located and a first code smaller than the first portion and indicative of the first portion is generated. A second portion from the electronic mail message is located and a second code smaller than the second portion and indicative of the second portion is generated. The first code and the second code are stored.
Processing of unsolicited bulk electronic communication
Inventors: Geoffrey D. Ralston, David H. Nakayama, Matthew E. Lewin, Ravichandran Menon Jayachandran, Brian R. Woods, Udi Manber
Assigned to Yahoo! Inc.
US Patent 6,931,433
Granted August 16, 2005
Filed January 31, 2001
Abstract:
The invention relates to processing of electronic text communication distributed in bulk. In one embodiment, a method for detecting electronic text communication distributed in bulk is disclosed.
After receiving a first electronic text communication, it is processed with an algorithm to produce a first fingerprint. A time period is begun for the first electronic text communication. After receiving a second electronic text communications, it is also processed with the algorithm to produce a second fingerprint.
The first fingerprint to the second fingerprint are compared to determine if the first electronic text communication is similar to the second electronic text communication. A count for the first electronic text communication is updated based upon the comparison. It is determined if the count during the time period reaches a first threshold.
Processing of textual electronic communication distributed in bulk
Inventors: Geoffrey D. Ralston, David H. Nakayama, Matthew E. Lewin, Ravichandran Menon Jayachandran, Brian R. Woods, and Udi Manber
Assigned to Yahoo ! Inc.
US Patent 6,842,773
Granted January 11, 2005
Filed January 31, 2001
Abstract:
The invention relates to processing of electronic text communication distributed in bulk. In one embodiment, a process for blocking electronic text communication distributed in bulk is disclosed. In the process, a first electronic and a second electronic submission are received.
A first portion is extracted from the first electronic submission and a second portion is extracted from the second electronic submission. A first code is determined for the first portion and a second code is determined for the second portion. The first code is indicative of the first portion and the second code is indicative of the second portion.
After the first code is compared to the second code, the second electronic submission is filtered in response to that comparison.
Some patent applications pending with Udi Manber listed as a co-inventor
Inventors: Barnaby M. Dorfman, Udi Manber, and Jonathan A. Gold
Assigned to A9.COM, INC.
US Patent Application 20050216193
Published September 29, 2005
Filed: March 24, 2004
Abstract:
A method and system are provided for collecting images of objects at geographic locations for publication in an online directory, such as online Yellow Pages. A camera and a GPS receiver, both receiving time information from a synchronized clock, are mounted on a vehicle.
Time-stamped images captured by the camera and time-stamped geographic locations (e.g., longitude and latitude coordinates) determined by the GPS receiver are recorded. Each image is then associated with a geographic location based on corresponding respective time-stamps.
Thereafter, each image is correlated with a street addresses based on each image’s associated geographic location, for example by referring to a lookup table that correlates addresses with geographic coordinates. The address-correlated images may then be used in an online directory, where each listing (e.g., business) can be displayed with an image of an object (e.g., store front) taken at the address in the listing.
System and method for displaying images in an online directory
Inventors: Barnaby M. Dorfman, Udi Manber, Jonathan A. Gold, Huy T. Ha, Timothy Caro-Bruce, and Jason Karls
Assigned to A9.COM, INC.
US Patent Application 20050216186
Published September 29, 2005
Filed: January 26, 2005
Abstract:
Embodiments of the present invention provide a method and system for generating an online listing, such as a yellow pages listing, that includes information about an object, such as a business, and also includes at least one image of the structure containing the object.
For example, the image may be a digital image of the store front of a business taken at a street view. Additionally, a user may select an image from a plurality of images that the user perceives as being an appropriate representative image for the business. Based on votes received a default representative image may be identified that is provided in response to a request for information about the business.
Systems and methods for identifying and extracting data from HTML pages
Inventors: Udi Manber and Qi Lu
Assigned to Yahoo! Inc.
US Patent Application 20050273706
Published December 8, 2005
Filed: May 4, 2005
Abstract:
Systems and methods for analyzing HTML formatted web pages to automatically identify and extract desired information. A computer algorithm identifies and extracts different pieces of information from different web pages automatically after minimal manual setup.
The algorithm automatically analyzes pages with different content if they have the same, or similar, formats. The algorithm is fast and efficient and performs the extraction process quickly in real-time. The systems and methods are useful to build databases from unstructured web information.
The algorithm can be used as an agent that captures information about products, and compares prices or other characteristics. It can also be used to populate structured databases that, given the different pieces of information, can analyze products and their characteristics. And it can also be used for data mining applications looking for patterns useful for marketing analyses, or other uses.
Automated bulk communication responder
Inventors: Geoffrey D. Ralston, David H. Nakayama, Matthew E. Lewin, Ravichandran Menon Jayachandran, Brian R. Woods, and Udi Manber
Assigned to Yahoo! Inc.
US Patent Application 20050172213
Published August 4, 2005
Filed: December 21, 2004
Abstract:
The invention relates to processing similar electronic text communication. In one step, a first electronic and a second electronic submission, that are part of a plurality of electronic submissions, are received. A first portion is extracted from the first electronic submission and a second portion from the second electronic submission.
Content of the first electronic submission influences extraction of the first portion and content of the second electronic submission influences extraction of the second portion. Locations for the first and second portions could vary with content.
A first code is determined for the first portion and a second code for the second portion, where the first code is indicative of the first portion and the second code is indicative of the second portion. The first code to the second code are compared to associate the first and second electronic submission together.
Inventors: Udi Manber and Hilliard Siegel
US Patent Application 20050063612
Published March 24, 2005
Filed: December 31, 2003
Abstract:
Methods and apparatus of the invention enable users to request access to one or more electronic images of pages in a physical text. When the user is identified and user ownership of the physical text is confirmed, the user is given access to the requested electronic images in accordance with the one or more access rules.
Electronic images of pages may be automatically added to a user-personalized library of electronic content for later access. A flag associated with the user and the pages images may be set to indicate confirmed user ownership of the physical text. A user may purchase a physical text itself or purchase an item that the physical text normally accompanies.
Electronic page images may be acquired by scanning printed pages of the text or from a user upload. Access to the electronic images of a physical text is based on user ownership of the physical text.
Systems and methods for matching participants to a conversation
Inventors: Udi Manber, Chi Chao Chang, and Anthony Lamarca
Assigned to Yahoo! Inc.
US Patent Application 20040078432
Published April 22, 2004
Filed: October 14, 2003
Abstract:
Systems and methods for connecting two or more individuals to an Internet conversation based on their mutual interests, the current content they may be viewing and what they want to talk about at that time.
The techniques of the present invention allow an individual who starts a conversation to maintain full control over who is able to join that conversation as well as how many are able to join at any one time. A user who desires to start or join a conversation about a particular topic or story selects an indicator, such as an icon, associated with the specific topic or story. The user is presented with an option to start or join a conversation.
If the user opts to start a conversation, the user is presented with a comment page, and the user enters a comment, or comments, that preferably is intended to spark an interest in other users. The comment is then presented to other users. Those users who may desire to join in a conversation with the conversation starter respond with their own comment, which is then sent to the conversation starter.
The conversation starter reviews the comment, and if the comment is subjectively acceptable, the conversation starter brings the responding user into a conversation. The responding user and conversation starter are then connected in a messaging session, such as an instant messaging session. Multiple users may be connected in a single messaging session by the conversation starter in this manner.
Personalized searchable library with highlighting capabilities
Inventors: Udi Manber, Hilliard Siegel, and Charles Gordon
US Patent Application 20050076012
Published April 7, 2005
Filed: September 23, 2003
Abstract:
A system and method for preparing and searching a user-personalized library of content includes preparing a general library comprised of images of pages of content and electronically-searchable text corresponding to the text in the page images, which is then personalized by receiving from the user a selection of the content to include in the user’s personalized library.
User selection of content for a personalized library may be accomplished manually or automatically as a result of user actions, such as user review or purchase of content. Preferably, full text searching of the content in the user’s personalized library is enabled.
Page images in the personalized library containing text that match the user’s search terms are provided to the user for display. Highlighting of search terms in page images by a user device and access rules that act to limit the viewing of content are further provided.
Server architecture and methods for persistently storing and serving event data
Inventors: Taylor N. Van Vleet, Yu-Shan Fung, Ruben Ortega, and Udi Manber
US Patent Application 20050033803
Published February 10, 2005
Filed: July 2, 2003
Abstract:
A web site system includes an event history server system that persistently stores event data reflective of events that occur during browsing sessions of web site users, and makes such data available to other applications and services in real time.
The server system may, for example, be used to record information about every mouse click of every recognized user, and may also be used to record other types of events such as impressions and mouse-over events. The event data of a particular user may be retrieved from the server system based on event type, event time of occurrence, and various other criteria.
In one embodiment, the server system includes a cache layer that caches event data by session ID, and includes a persistent storage layer the persistently stores the event data by user ID. Also disclosed are various application features that may be implemented using the stored event data.
Method and apparatus for search ranking using human input and automated ranking
Inventors: Udi Manber, and Chi-Chao Chang
Assigned to Yahoo! Inc.
US Patent Application 20040024752
Published February 5, 2004
Filed: August 5, 2002
Abstract:
A search system provides search results to searchers in response to search queries and the search results are ranked.
The ranking is determined by an automated ranking process in combination with human editorial input. A search system might comprise a query server for receiving a current query, a corpus of documents to which the current query is applied, ranking data storage for storing information from an editorial session involving a human editor and a reviewed query at least similar to the current query, and a rank adjuster for generating a ranking of documents returned from the corpus responsive to the current query taking into account at least the information from the editorial session.
Recent Comments