Obviousness

The obviousness of US Patent 9092428 under 35 U.S.C. § 103 can be assessed by combining the teachings of the applicant's own prior art references, which are explicitly cross-referenced and incorporated into the patent's disclosure. The claims of US9092428 appear to represent an obvious combination or generalization of features already disclosed in these earlier applications.

Identified Prior Art References (from "CROSS REFERENCES TO RELATED APPLICATIONS"):

The following patent applications, all filed before the priority date of US9092428 (December 9, 2011, based on 61/568,657), are particularly relevant:

1. U.S. Provisional Patent Application 61/568,657 (filed December 9, 2011): Titled "SYSTEM, METHODS, AND USER INTERFACE FOR ORGANIZING DOCUMENT CONTENTS INTO A HIERARCHICAL STRUCTURE AND SELECTIVE HIGHLIGHTING OF TERMS." This document forms the priority basis for US9092428.
2. U.S. patent application Ser. No. 12/972,462 (filed December 18, 2010): Titled "AUTOMATED TOPIC DISCOVERY IN DOCUMENTS."
3. U.S. patent application Ser. No. 12/782,545 (filed May 18, 2010): Titled "SYSTEM AND METHODS FOR AUTOMATED DOCUMENT TOPIC DISCOVERY, BROWSABLE SEARCH AND DOCUMENT CATEGORIZATION."
4. U.S. Provisional Patent Application 61/299,823 (filed January 29, 2010): Titled "SYSTEM AND METHODS FOR AUTOMATED DOCUMENT TOPIC DISCOVERY, BROWSABLE SEARCH, RELEVANCE RANKING, SUMMARY GENERATION AND DISPLAY."

Obviousness Analysis of Independent Claims 1, 13, and 16:

Independent Claims 1 (method), 13 (system), and 16 (computer program product) of US9092428 broadly cover a computer-implemented approach for discovering and presenting information in text content. The core steps involve tokenizing text, performing linguistic analysis to identify grammatical, semantic, or contextual attributes, associating these attributes with tokens, providing a user interface to select attributes and actions, and performing selected actions (extracting, displaying in various formats, showing/hiding, highlighting) on terms associated with the specified attributes.

Combination of Prior Art: U.S. Provisional Patent Application 61/568,657 in combination with U.S. patent application Ser. No. 12/972,462 and U.S. patent application Ser. No. 12/782,545.

Rationale for Combination and Obviousness:

A person having ordinary skill in the art (POSITA) in natural language processing (NLP) and user interface design, confronting the widely recognized problem of "information overload" (as explicitly stated in the Background of US9092428), would have found it obvious to combine the functionalities taught by these prior art documents. The motivation to do so is to create more efficient and effective tools for information discovery and management, building upon the foundations already established by the same inventor.

Let's break down the elements of Claim 1 and their presence in the combined prior art:

• Steps (a)-(d) (Receiving, Tokenizing, Linguistic Analysis, Associating Attributes):

  • US9092428 itself states that the "detailed methods for obtaining such a term prominence or tem importance score are disclosed in U.S. patent application Ser. No. 12/972,462... and U.S. patent application Ser. No. 12/782,545...". These references teach tokenization, linguistic processing to identify grammatical, semantic, and contextual attributes, and associating these attributes (such as term importance) with tokens to analyze text content. The linguistic processing module (120) and dictionary (115) depicted in FIG. 1 of US9092428 perform these functions, and their underlying mechanisms are rooted in these incorporated disclosures.

• Step (e) (User Interface for Attribute/Action Selection):

  • U.S. Provisional Patent Application 61/568,657 explicitly discloses "user interface objects" for "selectively displaying or hiding certain terms in the hierarchical topic tree structure, or selectively highlighting terms representing specific types of information in the original text content." This demonstrates the concept of a user interface allowing selection of a criterion ("specific types of information," which encompasses grammatical, semantic, or contextual attributes) and a desired action.

• Steps (f) & (g) (Performing Selected Actions: Extracting, Displaying, Showing/Hiding, Highlighting):

  • U.S. Provisional Patent Application 61/568,657 teaches methods for "discovering information... and for organizing and presenting the discovered information in a novel hierarchical structure format." It further details "selectively displaying or hiding certain terms" and "selectively highlighting terms."
  • While 61/568,657 specifically mentions hierarchical structures and highlighting, the extraction of terms and their display in other common formats like lists or word clouds (as mentioned in US9092428's abstract and detailed description, e.g., FIG. 5) would be obvious extensions to a POSITA. Displaying extracted data in a list is a fundamental UI practice, and word clouds are a known visualization technique for presenting weighted terms.

Motivation to Combine:

The motivation to combine these prior art references is inherent in the problem addressed by US9092428: the need for "efficient tools for users to gather information from unstructured text contents with less effort."

1. Enhanced User Experience: By integrating the robust linguistic analysis and attribute identification from 12/972,462 and 12/782,545 with the interactive user interface and selective display/highlighting capabilities of 61/568,657, a POSITA would readily understand that a more powerful and user-friendly information discovery system could be achieved.
2. Improved Information Management: The combination directly addresses the problem of quickly and accurately locating specific information by allowing users to act on various identified linguistic attributes (grammatical, semantic, contextual, topical) through intuitive UI controls. The patent itself highlights the difficulty of "digging out information related to such topics from the numerous reviews" by conventional methods.
3. Logical Extension and Modularity: Linguistic processing modules (tokenization, parsing, attribute assignment) are fundamental components of any advanced text analysis system. It would be a natural and obvious step for a POSITA to connect the outputs of such modules with user interface controls that allow dynamic interaction with the identified attributes. For instance, if a system can identify "drug names" (as shown in FIG. 2 of US9092428), it would be obvious to provide a user interface option to "extract" or "highlight" these identified entities.

Context-Sensitive Attribute Interpretation: US9092428 also describes identifying "context information about the terms having a specified attribute... and the effect of the context on the attribute values" (e.g., "not good" vs. "good"). While this offers a refinement, sophisticated linguistic analysis modules (as referenced for topic discovery) are expected to handle nuances like negation and other contextual modifiers to improve the accuracy of attribute identification (e.g., sentiment). A POSITA would recognize the need for and implement such contextual analysis as a standard practice for enhancing the precision of linguistic processing.

Therefore, the combination of the teachings within U.S. Provisional Patent Application 61/568,657 and U.S. patent application Ser. Nos. 12/972,462 and 12/782,545 would render the claims of US9092428 obvious to a person having ordinary skill in the art, motivated to create more effective and efficient text information discovery and presentation tools.