Computer aided validation of patent disclosures
AssigneeVeripat, LLC
InventorMichael Robert Kahn
Life of this patent
Estimating the term…
Abstract
A method and system for analyzing a patent disclosure is disclosed. The method and system comprise a disclosure analysis and a separate claims analysis, such that each analysis may be performed independently. Missing and incorrect reference labels are identified within the disclosure. Antecedent basis and specification support are checked for the claim elements. Terms within the specification that do not have a reference number, but may require one, are identified, provided that they fit the profile of one of a set of particular lexical patterns.
Claims9
A method for identifying a term as unnumbered within a patent disclosure, comprising the steps of:
- tokenizing the patent disclosure to form a tokenized patent disclosure comprising a plurality of tokens;
- identifying a pattern of tokens comprising a first token immediately followed by a second token, wherein the second token is a numeric token or an alphanumeric token;
- comparing the first token to a predetermined list of ignorable tokens;
- discarding first tokens that are in the list of ignorable tokens;
- making a term dictionary, wherein the term dictionary contains a list of terms, each term comprising a non-discarded first token, and the term dictionary further contains the associated reference label for each term, the associated reference label comprising the second token immediately following the non-discarded first token;
- identifying a token type associated with each token within the tokenized patent disclosure;
- identifying a token sequence within the tokenized patent disclosure;
- comparing the token sequence to determine if the token types of each token within the token sequence matches a predetermined pattern of token types;
- searching the term dictionary to determine if the term is in the term dictionary;
- and
- indicating the term as an unnumbered term if the term is not in the term dictionary, and
- the token sequence matches a predetermined pattern;
- wherein each of the steps is performed on a computer comprising a microprocessor and non-transitory storage.
A system for computer aided patent inspection, comprising:
- means for identifying unnumbered terms within a patent disclosure;
- means for generating a dictionary of terms from the patent disclosure, and associated reference labels;
- means for editing the dictionary to add, edit, or remove terms;
- and
- means for checking numbering of terms in the patent disclosure against the terms in the dictionary;
- wherein the means for identifying unnumbered terms within a patent disclosure comprises:
- means for tokenizing the patent disclosure to a form a tokenized patent disclosure;
- means for identifying the token type for each token within the tokenized patent disclosure;
- and wherein the means for tokenizing the patent disclosure to form a tokenized patent disclosure comprises:
- means for identifying a pattern of tokens comprising a first token immediately followed by a second token, wherein the second token is a numeric token or an alphanumeric token;
- means for comparing the first token to a predetermined list of ignorable tokens;
- discarding first tokens that are in the list of ignorable tokens;
- means for making a term dictionary, wherein the term dictionary contains a list of terms, each term comprising a non-discarded first token, and the term dictionary further contains the associated reference label for each term, the associated reference label comprising the second token immediately following the non-discarded first token;
- wherein each of the means comprises a computer comprising a microprocessor and non-transitory storage.
Dependencies are inferred from the claim text.