Probabilistic password cracking system

One or more tangible non-transitory computer-readable media having computer-executable instructions for performing a method of running a software program on a computing device, the computing device operating under an operating system, the method including issuing instructions from the software program for a computer processor to generate a probabilistic password cracking system for cracking a targeted password for a secured user account associated with a user, the instructions comprising:

• receiving a plurality of known password strings, said plurality of known password strings formed of at least one category selected from the group consisting of alpha strings, digits, and special characters;
• deriving one or more base structures from said plurality of known password strings, whereby one base structure may include more than one password string from said plurality of known password strings;
• detecting relevant patterns from said plurality of known password strings, wherein said relevant patterns include an A-word, an R-word, an R-pattern, an M-word, and an A-pattern;
• automatically assigning a set of probability values to each relevant pattern of said relevant patterns and to each base structure of said one or more base structures based on a probability value of each alpha string, each digit, or each special character in said each base structure;
• creating a probabilistic context free grammar based on said set of probability values assigned to said each relevant pattern to said each base structure;
• receiving one or more input dictionaries containing a plurality of sequences of alpha characters;
• generating password guess strings in decreasing estimated probability via said probabilistic context-free grammar by utilizing said plurality of sequences of alpha characters;
• accessing a login interface to the secured user account;
• and
• applying said password guess strings from said computer processor sequentially to said login interface, whereby authentication of the user can be achieved.

One or more tangible non-transitory computer-readable media having computer-executable instructions for performing a method of running a software program on a computing device, the computing device operating under an operating system, the method including issuing instructions from the software program for a computer processor to generate a probabilistic password cracking system for cracking a targeted password for a secured user account associated with a user, the instructions comprising:

• receiving a plurality of known password strings, said plurality of known password strings formed of at least one category selected from the group consisting of alpha strings, digits, and special characters;
• deriving one or more base structures from said plurality of known password strings, whereby one base structure may include more than one password string from said plurality of known password strings;
• automatically incorporating a keyboard pattern into said one or more base structures, said keyboard pattern contained within at least one password string of said plurality of known password strings, said keyboard pattern being a sequence of contiguous characters starting from a particular key without regards to actual characters typed but uses a physical sequence shape of the actual characters;
• automatically assigning a set of probability values to each base structure of said one or more base structures based on a probability value of each alpha string, each digit, each special character, or each keyboard pattern in said each base structure;
• creating a probabilistic context free grammar based on said set of probability values assigned to said each base structure, wherein when a known password string of said plurality of known password strings includes only contiguous alpha characters, only contiguous digits, or only contiguous special symbols, said known password string is classified as an alpha string structure, a digit structure, or a special symbol structure and not as a keyboard pattern structure,
• said step of creating said probabilistic context free grammar further including deriving substructures from said alpha strings;
• detecting relevant patterns from said plurality of known password strings, wherein said relevant patterns include an A-word, an R-word, an R-pattern, an M-word, and an A-pattern;
• classifying a relevant pattern as said A-word by checking the presence of said each alpha string in said plurality of known password strings;
• classifying a relevant pattern as said R-word or said R-pattern by checking a repetition of said each alpha string in said plurality of known password strings, followed by checking the presence of the pattern of said each alpha string in said plurality of known password strings;
• classifying a relevant pattern as said M-word only if said relevant pattern is not classified as said A-word, said R-word, or said R-pattern, said step of classifying a relevant pattern as said M-word performed by detecting a lengthy substring that is a word in said plurality of known password strings and identifying said lengthy substring as a first component within said relevant pattern by starting at a rightmost character of said targeted password and recursively calling said M-word classification on the remaining substring of said lengthy substring;
• classifying a relevant pattern as said A-pattern if said relevant pattern is not classified as said M-word;
• automatically deriving word-mangling rules based on said detected relevant patterns, said word-mangling rules derived further based on said targeted password having an association with a targeted group, said targeted group based on at least one criterion selected from the group consisting of language, age, affiliation, and password creation policies;
• incorporating said relevant patterns into said probabilistic context-free grammar;
• utilizing probability smoothing to assign additional probability values to other keyboard patterns for other password strings not found in said plurality of password strings, wherein said step of utilizing probability smoothing is achieved by an equation
• Prob
• ⁡
• (
• p
• )
• =
• Prob
• ⁡
• (
• s
• )
• ⁢
• N
• i
• +
• α
• ∑
• N
• i
• +
• C
• ⁢
• ⁢
• α
• where Prob(s) is the probability of a keyboard shape s given the length of the keyboard pattern, Ni is the number of times an ith keyboard pattern of a shape s was found, α is a smoothing value, ΣNi is a sum of counts of the keyboard patterns found for the shape s, and C is a total number of unique patterns for the shape s;
• receiving one or more input dictionaries containing a plurality of sequences of alpha characters;
• optimizing a primary dictionary of said one or more input dictionaries based on size and content of said primary dictionary;
• assigning an additional probability value to said primary dictionary, wherein an effectiveness of said primary dictionary is measured by coverage and precision of said primary dictionary cracking said targeted password,
• said one or more input dictionaries further including a secondary dictionary for cracking said targeted password;
• generating password guess strings in decreasing estimated probability via said probabilistic context-free grammar by utilizing said plurality of sequences of alpha characters;
• accessing a login interface to the secured user account;
• and
• applying said password guess strings from said computer processor sequentially to said login interface, whereby authentication of the user can be achieved.