Prior art

To identify the most relevant prior art for US patent 7398209, a review of its cited references is necessary. US patent 7398209, titled "Systems and methods for responding to natural language speech utterance," has a priority date of June 3, 2002, and was published on July 8, 2008. The patent describes an integrated environment for natural language speech queries and commands, utilizing context, user profiles, domain-specific agents, and probabilistic/fuzzy reasoning for robust and natural responses. Its claims cover a system including a speech unit, a computer device with speech recognition, a parser, a text-to-speech engine, a network interface, a graphical user interface, an event manager (linked to a dictionary, user profiles, personality, agents, update manager, and databases), and domain-specific agents.

Based on a search of the prior art citations for US7398209, several patents are identified as highly relevant due to their focus on natural language processing, speech interfaces, agent systems, and context management, all of which are central to US7398209. Below is an analysis of some of the most relevant prior art references, providing their full citation, publication/filing dates, a brief description, and potential anticipatory aspects under 35 U.S.C. § 102.

Most Relevant Prior Art for US7398209:

1. US6295521B1: Method and system for providing natural language access to structured information

   • Full Citation: US6295521B1, Invented by W. C. Cheyer, A. Cheyer, A. S. Chaudhri, K. R. Murray, G. L. Myers, D. J. Perdue, K. P. Van Le, R. D. Weintraub, assigned to SRI International.
   • Publication/Filing Date: Publication Date: September 25, 2001. Filing Date: October 2, 1998.
   • Brief Description: This patent describes a system and method for enabling a user to access structured information (e.g., from databases) using natural language input. It involves converting natural language questions into formal queries, executing those queries, and presenting the results. A key aspect is the use of a natural language engine that maps user input to objects, actions, and constraints, and then generates structured queries (e.g., SQL) to retrieve information. It emphasizes the flexibility of natural language input and the dynamic generation of queries.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 5, 12: The concept of receiving natural language input (even if not explicitly speech in the abstract), parsing it, determining meaning, and formulating machine-processable queries to structured data sources is explicitly taught. The system's ability to "map natural language to objects, actions, and constraints" and "generate structured queries" could anticipate elements related to parsing natural language utterances and generating queries to databases.
     • General Concept: The fundamental idea of providing a natural language interface for information retrieval from structured sources is directly addressed.

2. US6076060A: Methods and apparatus for parsing natural language and for representing meaning

   • Full Citation: US6076060A, Invented by P. S. Cohen, G. J. Macias, T. E. Probst, R. Schabes, A. K. Singh, assigned to Lernout & Hauspie Speech Products N.V.
   • Publication/Filing Date: Publication Date: June 13, 2000. Filing Date: December 29, 1997.
   • Brief Description: This patent details methods and apparatus for parsing natural language and representing its meaning using a "head-driven parser." It focuses on converting natural language text into a semantic representation, which can then be used for various applications like machine translation or information extraction. The parsing system efficiently resolves ambiguities and handles complex linguistic structures.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 4, 5, 12: Directly anticipates the "parser configured to parse the natural language speech utterance to determine at least an approximate meaning of the natural language speech utterance" element of Claim 1, as well as the general parsing and interpretation steps in processes described in Claims 5 and 12. While it focuses on text, the underlying parsing mechanisms are broadly applicable to the natural language component once speech is recognized. The emphasis on resolving ambiguities is also relevant.

3. US5819220A: Intelligent agent system for computer network

   • Full Citation: US5819220A, Invented by S. A. G. O'Connell, R. A. S. North.
   • Publication/Filing Date: Publication Date: October 6, 1998. Filing Date: June 17, 1996.
   • Brief Description: This patent describes an intelligent agent system for a computer network where agents are capable of autonomous action, learning, and communication. The system allows users to deploy agents to perform tasks, potentially involving information gathering and processing across the network. It covers aspects of agent management, communication, and interaction within a networked environment.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 4, 31, 35: This patent could anticipate the broad concept of "one or more agents, the one or more agents providing domain specific functionality" as in Claim 1. Furthermore, the notion of agents in a networked environment, providing extended capabilities and potentially allowing users to "disseminate their knowledge" or "extend the system capabilities" (Claims 31, 35) through agents, aligns with aspects of this prior art. While not specific to natural language speech, the foundational agent architecture is present.

4. US6421672B1: Method and system for providing context for a question submitted to a natural language processing system

   • Full Citation: US6421672B1, Invented by W. C. Cheyer, A. Cheyer, S. A. G. O'Connell, D. J. Perdue, assigned to SRI International.
   • Publication/Filing Date: Publication Date: July 16, 2002. Filing Date: October 14, 1999. (Note: Publication date is after US7398209's priority date, but filing date is before, making it prior art under 35 U.S.C. § 102(e) pre-AIA).
   • Brief Description: This patent describes a system and method for providing context to a natural language processing system, particularly for handling ambiguous questions. It involves analyzing previous questions, user profiles, and domain knowledge to resolve ambiguities and improve the accuracy of responses. The system maintains a context stack to keep track of the conversation history.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 4, 12, 17: Directly anticipates the "determine a context for the natural language speech utterance" element of Claim 1, and the extensive use of "context, prior information, domain knowledge, and user specific profile data" as stated in Claim 4. The method of resolving ambiguity using context, user profiles, and conversation history is highly relevant to US7398209's reliance on these factors. The explicit mention of a "user profile" further anticipates Claim 17.

5. US6292770B1: System and method for generating dynamic structured queries from natural language speech input

   • Full Citation: US6292770B1, Invented by D. P. Singh, assigned to IBM.
   • Publication/Filing Date: Publication Date: September 18, 2001. Filing Date: June 30, 1999.
   • Brief Description: This patent describes a system that receives natural language speech input, converts it to text, processes the text to understand its meaning, and then generates dynamic structured queries for retrieving information from a database. It focuses on the transformation from speech to query, including handling of ambiguities and variations in natural language.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 5, 12: This patent directly anticipates many elements of US7398209, specifically: "speech unit configured to receive a natural language speech utterance" (implicitly, by converting speech to text), "speech recognition module" (for speech to text), "parser configured to parse the natural language speech utterance to determine at least an approximate meaning," and "machine processable queries and commands may be formulated after the natural language question and/or command has been parsed and interpreted." The title itself highlights "generating dynamic structured queries from natural language speech input," which is a core function of US7398209.

6. US6009387A: Speech recognition with semantic and contextual processing

   • Full Citation: US6009387A, Invented by K. P. Lang, assigned to General Magic, Inc.
   • Publication/Filing Date: Publication Date: December 28, 1999. Filing Date: June 18, 1998.
   • Brief Description: This patent details a speech recognition system that incorporates semantic and contextual processing to improve recognition accuracy and understanding. It moves beyond simple acoustic matching by integrating higher-level linguistic information and contextual cues during the speech recognition process to better interpret user utterances.
   • Potential Anticipation (35 U.S.C. § 102):
     • Claims 1, 4, 12: This patent addresses "speech recognition" and its integration with "contextual processing," which directly relates to the "speech recognition module" and the "parser configured... to determine a context for the natural language speech utterance" in Claim 1. The emphasis on using context during speech recognition also ties into the broad inventive concept of using context to overcome deficiencies in prior systems as stated in Claim 4 and the capture and interpretation steps of Claim 12.

This analysis highlights that key components and methodologies described in US7398209, such as natural language parsing, query generation, agent-based systems, context determination, and speech recognition integrated with semantic/contextual understanding, were present in the prior art. The novelty of US7398209 likely lies in its specific integration of these elements into a "complete speech-based natural language query and response environment" with features like dynamic personalities, graceful recovery from partial failures using probabilistic/fuzzy reasoning, and extensive user profiling for personalized responses and command control. The extent of anticipation under 35 U.S.C. § 102 would depend on a detailed, word-by-word comparison of US7398209's claims against the entire disclosure of each prior art reference.