Xerox Corporation

Palo Alto Research Center Inc.

John Turner Maxwell, Iii

Ronald M Kaplan

Matthew Klenk

Kent J Evans

Matthew Adam Shreve

Stuart Kent Card

D Austin Henderson, Jr

Shiwali Mohan

Aaron Ang

Johan De Kleer

Daniel Gureasko Bobrow

Jessica Nicola Staddon

Philippe J P Golle

Prateek Sarkar

Tracy Holloway King

Evgeniy Bart

Hadar Shemtov

Kyle D Dent

A method is provided. The method includes obtaining a state graph that represents a set of objects within an environment and a set of positions of the set of objects within the environment. The state graph includes a set of object nodes and a set of property nodes. The method also includes obtaining user input data. The user input data is generated based on a natural language input. The method further includes updating the state graph based on the user input data to generate an enhanced state graph. The enhanced state graph includes additional nodes generated based on the user input data. The method further includes providing the enhanced state graph to a planning module. The planning modules generates instructions for operating a mechanical system based on the enhanced state graph.

Interacting with machines using natural language input and an enhanced state graph

In one embodiment, a method is provided. The method includes obtaining sensor data indicative of a set of objects detected within an environment. The method also includes determining a set of positions of the set of objects and a set of properties of the set of objects based on the sensor data. The method further includes generating a state graph based on the sensor data. The state graph represents the set of objects and the set of positions of the set of objects. The state graph includes a set of object nodes to represent the set of objects and a set of property nodes to represent the set of properties of the set of objects. The state graph is provided to a graph enhancement module that updates the state graph with additional data to generate an enhanced state graph.

Interacting with machines using natural language input and a state graph

A method is provided. The method includes obtaining sensor data indicative of a set of objects detected within an environment. The method also includes generating a state graph based on the sensor data. The state graph includes a set of object nodes and a set of property nodes. The method further includes obtaining user input data generated based on a natural language input. The method further includes updating the state graph based on the user input data to generate an enhanced state graph. The enhanced state graph includes additional nodes generated based on the user input data. The method further includes generating a set of instructions for a set of mechanical systems based on the enhanced state graph. The method further includes operating the set of mechanical systems to achieve a set of objectives based on the set of instructions.

System for interacting with machines using natural language input

The techniques disclosed herein help designers find interesting designs for small electrical, mechanical, and/or hydraulic mechanisms by exhaustively enumerating the design space given a library of components and a maximum number of components allowed per design. Some embodiments work by creating a design space grammar of designs, solving the equations associated with parts of the grammar, and putting the solutions into equivalence classes. This dramatically reduces the number of designs that have to be evaluated to see if they satisfy the design criteria. The result is often a small number of base designs that show the range of possible solutions to the design problem.

Method for generating designs based on automatically built design space grammars

A method is provided. The method includes obtaining an enhanced state graph. The enhanced state graph represents a set of objects within an environment and a set of positions of the set of objects. The enhanced state graph includes a set of object nodes, a set of property nodes and a set of goal nodes to represent a set of objectives. The method also includes generating a set of instructions for a set of mechanical systems based on the enhanced state graph. The set of mechanical systems is configured to interact with one or more of the set of objects within the environment. The method further includes operating the set of mechanical systems to achieve the set of objectives based on the set of instructions.

Controlling mechanical systems based on natural language input

Described are natural language web browsers configured to provide a natural language interface to an existing web browser or a newly created web browser. That is, users interact with the web browser employing speech (or text) as input and output instead of using a mouse and keyboard as input and an electronic display screen as output. Embodiments of a natural language web browser converts user questions, statements, and/or commands into actions, reads the resulting HTML and converts the HTML into natural language descriptions to provide to the user.

Natural language web browser

Methods and system employing the same for optimizing an objective function are provided. The objective function assesses the quality of a candidate solution. One or more variables of an objective function are selected as pivot variables. Each of the variables include one or more candidate values. An upper bound function is generated from the objective function, where the pivot variables are held fixed. For each combination of the candidate values, one or more candidate solutions are searched using the upper bound function. One or more optimal solutions are selected from among the solutions to the searches.

System and method for efficient interpretation of natural images and document images in terms of objects and their parts

Methods and systems supporting curation of items in a searchable knowledge base are provided. The methods and systems include mining one or more search queries of the searchable knowledge base, where each of the search queries includes a plurality of the items. The method further includes determining one or more pairs of items using a processor, where each of the pairs of items includes a correlation value exceeding a threshold. The correlation values for the pairs of items are based upon the frequency the items of the pairs of items co-occur within the search queries. The method further includes providing the pairs of items to a curator, where the curator reviews the pairs of items.

Detecting duplicates in a shared knowledge base

One embodiment provides a system that detects sensitive passages. During operation, the system receives a document and disassembles the document into a plurality of passages. For a respective passage, the system performs a search through a non-sensitive-passage database to determine whether the passage is a known non-sensitive passage. If so, the system marks the passage as non-sensitive, and if not, the system determines whether the passage triggers a cut-and-paste attack detection. If so, the system forwards the passage to an administrator and allows the administrator to determine whether the passage is non-sensitive and, further, to add the passage to the non-sensitive-passage database responsive to the administrator determining the passage to be non-sensitive.

Non-sensitive-passage database for cut-and-paste attack detection systems

Techniques are provided to generate alternate phrases from a packed meaning representation. The semantics of input items in a packed meaning representation are matched against a lexicon to determine initial entries in a phrasal information structure. Context, syntax and semantic information describing each phrasal entry are added to the phrasal information structure. Linear or other implications are used to determine a set of unordered rewrite rules. The unordered rewrite rules associate semantic information with pseudofacts such that a complete set of choices from the set of disjunctions in the packed meaning representation is selected. Semantic information associated with the unordered rewrite rules identifies entries in the phrasal information structure to be updated with the pseudofact of the unordered rewrite rule.

Systems and methods for the generation of alternate phrases from packed meaning

Techniques are provided to generate alternate phrases from a packed meaning representation. The semantics of input items in a packed meaning representation are matched against a lexicon to determine initial entries in a phrasal information structure. Context, syntax and semantic information describing each phrasal entry are added to the phrasal information structure. Linear or other implications are used to determine a set of unordered rewrite rules. The unordered rewrite rules associate semantic information with pseudofacts such that a complete set of choices from the set of disjunctions in the packed meaning representation is selected. Semantic information associated with the unordered rewrite rules identifies entries in the phrasal information structure to be updated with the pseudofact of the unordered rewrite rule. Entries in the phrasal information structure not associated with false contexts are combined based on ordered rewrite rules. The ordered rewrite rules are based on the language of the packed meaning representation. Source information associated with each new entry in the phrasal information structure reflects the phrasal entry identifiers of the entries used to determine the new entry. For any set of entries in the phrasal information structure produced by an ordered rewrite rule, the context of a missing semantic fact is subtracted from the context of the entry lacking the semantic fact. Entries associated with a false context are discarded and entries sharing the same syntax and semantic information are merged. Any entries in the phrasal information structure associated with the 'ALL' pseudofact are then determined. Alternate phrases are generated based on the source information for the entry in the phrasal information structure associated with the 'ALL' pseudofact.

A process for generating with unification based grammars such as Lexical Functional Grammars which uses construction and analysis of generation guides to determine internal facts and eliminate incomplete edges prior to constructing a generation chart. The generation guide can then be used in the construction of the generation chart to efficiently generate with unification-based grammars such as Lexical Functional Grammars. The generation guide is an instance of a grammar that has been specialized to the input and only contains those parts of the grammar that are relevant to the input. When the generation guide is analyzed to determine internal facts a smaller generation chart is produced.

Generating with lexical functional grammars

Generating with Lexical Functional Grammars

Techniques are provided for determining the efficient conjunction of Boolean variables drawn from an and-or forest of choices. The techniques determine two contexts to conjoin, where the contexts consist of choices drawn from an and-or forest of choices. The choices in the first context are marked with a first mark, along with all of their antecedent choices. The choices in the second context are marked with a second mark, along with all of their antecedent choices. The techniques determine and-nodes in the and-or forest having the first mark from a first dependent or-node and the second mark from a second dependent node. The dependent or-nodes are refined and then partitioned into classes based on the presence of the first mark and the second mark, the first mark only, the second mark only, and neither mark. A cross-product of the partitions for each node is determined and the resultant combinations are grouped based on the presence of both the first and second marks, the presence of the first mark only, the presence of the second mark only and presence of neither mark. A new or-node having choices reflecting the combinations having both marks, the first mark only, the second mark only and neither mark is determined. The new or-node represents a transformed and-or forest of choices that is in free-choice form. The first and second nodes in the forest of choices are replaced with an or-node having choices based on the first and second mark. The conjunction of contexts is then determined based on the choice having both the first and second marks.

Systems and methods for efficient conjunction of Boolean variables

Systems and methods for solving nogood databases involve generating a representation comprising a plurality of contexted disjunctions, conjoining all of the contented disjunctions to form a conjunction of contexted disjunctions, and storing the representation as the conjunction of contexted disjunctions. Nogoods are eliminated by refining the representation until a result of the conjunction of contexted disjunctions is backtrack-free or the result of the conjunction of contexted disjunctions reduces to false. In various embodiments, the refining is carried out without reordering the disjunctions and/or without merging the disjunctions. In various embodiments, the systems and methods are used for various constraint satisfaction problems, such as syntactic processing of natural language sentences, map coloring, understanding line drawings, electronic circuit analysis, and truth maintenance systems.

Growing community of inventors

Cupertino, CA, United States of America

John Turner Maxwell, Iii