International Business Machines Corporation

Benny Kimelfeld

Yunyao Li

Shivakumar Vaithyanathan

Huahai Yang

Zhuowei Bao

Liang Gou

Ching-Tien Ho

Christopher Palmer

Tyler Shore Baldwin

Jerome M Pesenti

Ching-Tien (Howard) Ho

Zhichao Hu

Congle Zhang

Embodiments of the present invention disclose a method, computer program product, and system for searching a database using a user entered search query. A search query for a database is received by the computer and the computer applies condition-action rules based on natural language processing rules to identify one or more phrases within the search query that is associated an entity identifier. The computer further identifies any taxonomy variants that have been established for the identified phrases. The computer creates a search string that includes search query and the entity identifiers. The database search is conducted by the computer and the results are displayed for the user.

Domain centric natural language query answering

An embodiment provides methods and arrangements for visualizing and exploring natural-language text. In an embodiment, natural language text is received, and this is parsed to form a directed graph comprising a plurality of nodes. The directed graph is converted to an outline graph comprising core nodes and layer nodes. The outline graph is simplified, and there is created, for display on a user interface, an interactive visual representation of the simplified outline graph. Other variants and embodiments are broadly contemplated herein.

Visualizing and exploring natural-language text

A method includes receiving one or more natural language dependency parse trees as input. A hardware processor is used for processing the dependency parse trees by creating a mapping from nodes of the one or more dependency parse trees into actions, roles and contextual predicates. The mapping is used for information extraction. The actions include the verbs along with attributes of the verbs. The roles include arguments for the verbs. The contextual predicates include modifiers for the verbs.

Facilitating information extraction via semantic abstraction

Textual data is organized into statement clusters. Sentences are extracted from textual data and parsed. A verb usage pattern is identified and an SVO triplet is determined. The SVO triplet is compared to a taxonomy associated with the domain of the data and a sentiment is derived. A statement cluster is constructed comprising a higher level SVO triplet sensitive to the taxonomy and verb usage pattern, as well as the derived sentiment. Accordingly, the statement clusters may be organized by grouping.

SVO-based taxonomy-driven text analytics

Embodiments of the present invention relate to a customizable text normalization framework providing for domain adaptability through modular replacement generators. In one embodiment, a method of and computer program product for text normalization are provided. An input sequence comprising a plurality of tokens is received. A plurality of generators is applied to the input sequence to generate a set of candidate replacements of the tokens of the sequence. A plurality of subsets of the set of candidate replacements is determined such that the candidate replacements of each subset are syntactically consistent. A probability is determined for each of the subsets. A subset of the plurality of subsets having the highest probability is selected. Each candidate replacement of the selected subset is applied to the input sequence to generate an output sequence. The output sequence is outputted.

Adaptive parser-centric text normalization

Embodiments of the invention relate to automatically suggesting query-rewrite rules. One embodiment includes providing a missing search result for a query. A collection of semantically coherent rewrite rules are generated based on the missing search result. Generating the missing search result includes: selecting candidates including subsequences of the query and subsequences of particular fields of a document, invoking a search engine using the candidates for providing search results, filtering out particular candidates that fail to achieve a desired search result, and classifying remaining candidates based on a learned classifier. Query rewrite rules for document searching are suggested based on the classified remaining candidates.

Automatic suggestion for query-rewrite rules

Using exemplary sentences, usage patterns and thematic roles ascribed in VerbNet to generate 'deep pattern trees' for the exemplary sentences. Then, when an arbitrary natural language subject sentence is input, these deep pattern trees can be matched to the natural language subject sentence in order to assign thematic roles to at least some of the 'grammatical portions' of the natural language subject sentence.

Unsupervised learning of deep patterns for semantic parsing

Methods and arrangements for enhancing search quality. Query search results are displayed, and search query provenance related to the search results is graphically depicted. There is graphically accorded an investigative function to avail investigation of at least one aspect of the search query provenance.

Growing community of inventors

Santa Clara, CA, United States of America

Benny Kimelfeld