Obviousness

Obviousness Analysis of US Patent 9263039 Under 35 U.S.C. § 103

This analysis identifies combinations of prior art elements that would render the claims of US patent 9,263,039 obvious to a person having ordinary skill in the art (POSA) at the time of the invention (priority date: 2005-08-05). Due to the instruction to "Use the results from the Prior Art section of this page," and the provided text primarily containing the "BACKGROUND OF THE INVENTION" section which describes the general state of the art rather than listing specific prior art documents by identifier, this analysis will rely on the understanding of the art as presented within the patent's background and common knowledge in the field around the priority date. For the purpose of citation, as the entire patent text is provided as one authoritative block, all direct references to the patent's content will be cited as ``.

1. Level of Ordinary Skill in the Art (POSA)

A person having ordinary skill in the art (POSA) for US9263039 as of its August 5, 2005 priority date would likely possess a Master's degree in computer science, electrical engineering, or a closely related field, with a specialization in artificial intelligence, natural language processing, speech recognition, human-computer interaction, or software engineering. Alternatively, a Bachelor's degree in these fields combined with several years of practical experience in developing or implementing conversational AI systems, speech interfaces, or complex information retrieval architectures would also be appropriate. Such a POSA would be conversant with contemporary techniques in Automatic Speech Recognition (ASR), Natural Language Processing (NLP), database querying, system modularity (e.g., using agents), and methods for handling uncertainty and ambiguity, such as probabilistic reasoning or fuzzy logic.

2. Scope and Content of the Prior Art

Based on the "BACKGROUND OF THE INVENTION" section of US9263039, the following elements were known or generally understood in the prior art by 2005:

• Speech Recognition and Natural Language Processing: "Speech recognition has steadily improved in accuracy and today is successfully used in a wide range of applications. Natural language processing has been applied to the parsing of speech queries." This indicates that the core technologies for converting speech to text and interpreting text for queries were established.
• Challenges in Natural Interaction: The patent acknowledges that a "machine's ability to communicate with humans in a natural manner remains a difficult problem." It further notes that "verbal communication, such as a person asking a question or giving a command, typically relies heavily on context and domain knowledge of the target person." In contrast, "machine-based queries (e.g., questions, commands, requests, and/or other types of communications) may be highly structured and may not be inherently natural to the human user."
• Need for a "Complete Environment": The patent states that "current systems do not reliably provide a complete environment for users to submit verbal and/or textual communications through natural language queries that are processed to provide natural responses. There remain a number of significant barriers to creation of a complete speech-based and/or non-speech-based natural language query and response environment." This identifies the problem of lacking a fully integrated, natural-feeling system.
• Handling Imperfect Information and Partial Failure: The patent explicitly aims to overcome these issues, stating in the "SUMMARY OF THE INVENTION" that "The robustness to partial failure is achieved through the use of probabilistic and fuzzy reasoning at several stages of the process." This implies that probabilistic and fuzzy reasoning were recognized techniques for enhancing robustness and dealing with incomplete or uncertain information in various computing contexts.
• User Profiles, Context, and Domain Knowledge: The background implicitly highlights the importance of these for natural human communication, and their absence or insufficient use in prior machine systems. The invention claims to make "maximum use of context, prior information, domain knowledge, and user specific profile data". These were known concepts in various fields of computer science (e.g., personalization, dialogue systems, expert systems).

3. Differences Between the Prior Art and Claimed Invention

Claim 1 of US9263039 defines a system comprising a speech unit interface, a multi-pass ASR module, a parser (using probabilistic/fuzzy methods for context), an event manager, a context description grammar module, a user profile module, an agent module (with domain-specific agents), and a response generator module (using probabilistic/fuzzy methods for generating natural language responses, especially with inconsistent/ambiguous results).

The patent's background establishes that basic speech recognition and natural language parsing were known. The primary differences claimed by the patent lie in the integrated architecture that explicitly combines these with user profiles, a context description grammar, and domain-specific agents, and critically, the explicit application of "probabilistic or fuzzy set decision and matching methods" at multiple points:

1. In the parser, for "identifying a context for the command or the question."
2. In the response generator, for "generating a natural language response... including probabilistic or fuzzy set decision and matching methods to deal with inconsistent, ambiguous, conflicting or incomplete information or responses."

4. Motivation to Combine

A POSA, faced with the problems identified in the "BACKGROUND OF THE INVENTION" of US9263039, would have been motivated to combine known elements and techniques to achieve a more "complete" and "natural" speech-based interaction environment.

1. Combination of Core Speech/NLP with User Profiles, Context, and Domain Agents: Given that "Speech recognition has steadily improved" and "Natural language processing has been applied to the parsing of speech queries", a POSA would start with a system incorporating these functionalities (let's call this Prior Art A: a generic speech recognition and natural language processing system for interpreting user utterances). To address the acknowledged reliance of human communication on "context and domain knowledge" and to overcome the "significant barriers to creation of a complete speech-based... natural language query and response environment", a POSA would be motivated to integrate:

   • Prior Art B (User Profiles): The use of user profiles for personalization and recalling past interactions was well-known in various human-computer interfaces. Integrating a "user profile module" to store "user specific data, parameters, and session and history information" would be an obvious design choice to make the system more "natural" and tailored to the individual user.
   • Prior Art C (Context Management): To enable more coherent and natural dialogues, a POSA would incorporate context management techniques. This directly addresses the human reliance on context. Implementing this through a "context description grammar module" would be a conventional approach to define and manage conversational state.
   • Prior Art D (Domain-Specific Agents): To handle queries and commands across "multiple domains" effectively and modularly, a POSA would find it obvious to employ "agents" or modules specialized for different domains, as this is a standard software engineering practice for managing complexity and extending system capabilities. The "agent module comprising a plurality of agents, each agent configured to correspond to a domain" reflects this common architectural pattern.
   • The inclusion of an "event manager" is a standard architectural component for coordinating interactions among modules in a complex, multi-threaded system.

2. Application of Probabilistic or Fuzzy Reasoning: Recognizing that natural language utterances inherently contain "imperfect information such as, incomplete thoughts, incomplete sentences, incomplete phrases, slang terminology, repeated words, word variations, synonyms, or other imperfect information", and that results from various sources might be "inconsistent, ambiguous, conflicting or incomplete", a POSA would be strongly motivated to apply established techniques for handling such uncertainties. Prior Art E (Probabilistic/Fuzzy Logic) represents the widely known methodologies for dealing with uncertainty in artificial intelligence and pattern recognition systems by 2005. The patent itself explicitly states that "The robustness to partial failure is achieved through the use of probabilistic and fuzzy reasoning at several stages of the process." This statement reveals the inventors' own recognition that these methods are suitable for achieving robustness. Therefore, a POSA would have found it obvious to apply these known probabilistic or fuzzy set decision and matching methods:

   • To the "parser" for more robust "identif[ication of] a context for the command or the question" when dealing with imperfect or ambiguous input.
   • To the "response generator module" to intelligently "deal with inconsistent, ambiguous, conflicting or incomplete information or responses" when formulating a natural language output, especially when drawing from heterogeneous sources, thereby making the system's responses more "natural" and reliable.

The combination of these known elements (ASR, NLP, user profiles, context management, domain agents, and probabilistic/fuzzy logic) would be a logical and predictable step for a POSA seeking to create a more robust, natural, and complete speech-based interaction system, directly addressing the deficiencies articulated in the patent's own background. The motivation would stem from the desire to overcome these known problems using well-understood tools and architectural patterns.

5. Reasonable Expectation of Success

By 2005, each of the conceptual elements (speech recognition, natural language processing, user profiles, context management, domain-specific modules, and probabilistic/fuzzy logic for handling uncertainty) was individually established and implemented in various forms. Combining these modular components into an integrated system, as described in the patent, would have been within the capabilities of a POSA. The application of probabilistic or fuzzy methods to enhance robustness in areas like context identification and response generation was a recognized approach in research and development, providing a reasonable expectation that such an integration would yield the desired improvements in handling imperfect information and generating natural responses.

Therefore, the claimed invention, particularly Claim 1, would have been obvious to a person having ordinary skill in the art in light of the general state of the art and common knowledge as described in the patent's own background, and the clear motivations to combine these known elements to solve the identified problems.