286 results about "Predictive text" patented technology

The present invention includes systems and methods for retrieving information via a flexible and consistent targeted search model that employs interactive multi-prefix, multi-tier and dynamic menu information retrieval techniques (including predictive text techniques to facilitate the generation of targeted ads) that provide context-specific functionality tailored to particular information channels, as well as to records within or across such channels, and other known state information. Users are presented with a consistent search interface among multiple tiers across and within a large domain of information sources, and need not learn different or special search syntax. A thin-client server-controlled architecture enables users of resource-constrained mobile communications devices to locate targeted information more quickly by entering fewer keystrokes and performing fewer query iterations and web page refreshes, which in turn reduces required network bandwidth. Applications are built by leveraging existing collaborative cloud services that enable the maintenance and sharing of user content.

An electronic messager with a predictive text editor, including a storage unit for storing a data structure associating, for each one of a plurality of a user's contacts, usage data for the user's history of usage of words in communications with the user contact, a data manager coupled with the storage unit for generating the data structure in the storage unit, and for updating the data structure as additional communications with each user contact are performed and additional usage data is obtained therefrom, and a text predictor coupled with the storage unit, for receiving as input a character string and a designated user contact, and for generating as output an ordered list of predicted. A method is also described and claimed.

An adaptation of standard edit distance spell-check algorithms leverages probability-based regional auto-correction algorithms and data structures for ambiguous keypads and other predictive text input systems to provide enhanced typing correction and spell-check features. Strategies for optimization and for ordering results of different types are also provided.

A predictive text personal computer simplified keyboard with word and phrase auto-completion. It has a smaller keypad with each key representing several letters / characters so that only 9 keys are required to represent the entire alphabet of 26 characters. It also has screens and selection buttons which, when combined with predictive text input software such as the T9 dictionary from Tegic Communications, Inc. (www.t9.com), enables the user to type just a few keys to have entire words or phrases automatically typed into their document, thus saving typing time and avoiding the frustration of hunting for characters across a keyboard. It also offers language translation and speech capabilities, both enhanced by predictive text auto-completion.

The present invention includes systems and methods for retrieving information via a flexible and consistent targeted search model that employs interactive multi-prefix, multi-tier and dynamic menu information retrieval techniques (including predictive text techniques to facilitate the generation of targeted ads) that provide context-specific functionality tailored to particular information channels, as well as to records within or across such channels, and other known state information. Users are presented with a consistent search interface among multiple tiers across and within a large domain of information sources, and need not learn different or special search syntax. A thin-client server-controlled architecture enables users of resource-constrained mobile communications devices to locate targeted information more quickly by entering fewer keystrokes and performing fewer query iterations and web page refreshes, which in turn reduces required network bandwidth. Applications are built by leveraging existing collaborative cloud services that enable the maintenance and sharing of user content.

A system and method for supporting predictive text entry in software applications by sharing a common, predictive, software text-entry widget within a consumer device across multiple software applications and input contexts. The method comprises: a software application invoking an instance of a text-entry widget in a particular input context, the application optionally providing the widget a description of allowed symbols and a dictionary of expected symbol strings associated with the current context, the widget modifying a virtual keyboard display and predictive algorithm data according to the allowed symbols and dictionary, the user entering text via the widget, the widget returning the entered text to the application, and the application optionally including information derived from entered text in the associated dictionary to enhance the predictive capability of the widget on future invocations.

The invention relates to an electronic text inputting for an electronic device comprising at least the steps of providing alphanumeric user input, searching words from at least one dictionary on the basis of the provided alphanumeric user input, providing the words resulting from the search to the user, wherein the at least one dictionary is formed based on words from at least one text piece acquired from a memory of the device, the at least one text piece being selected based on at least one person pertaining to the text. The invention relates to methods, to a device and to a computer program.

The invention relates to a hybrid neural network text classification method capable of blending an abstract with main characteristics. The method comprises the following steps that: step A: extractingan abstract from each text in a training set; step B: using a convolutional neural network to learn the key local features of the abstract obtained in the step A; step C: using a long short-term memory network to learn context time sequence characteristics on the main content of each text in the training set; step D: carrying out cascade connection on two types of characteristics obtained in thestep B and the step C to obtain the integral characteristics of the text, inputting the integral characteristics of each text in the training set into a full connection layer, using a classifier to calculate a probability that each text belongs to each category to train a network, and obtaining a deep neural network model; and step E: utilizing the trained deep neural network model to predict thecategory of a text to be predicted, and outputting the category with a highest probability as a prediction category. The method is favorable for improving text classification accuracy based on the deep neural network.

A hand-held communication device provides navigation key-based predictive text entry. The hand-held communication device includes a housing generally sized to be held in a human hand having a display disposed for displaying characters selectable for entry in a character position of a text string being entered and a navigation key assembly for scrolling through and selecting from the characters displayed by the display. The characters displayed by the display during text entry are arranged according to the probability of selection of each character for entry in the character position so that the character with the highest probability of selection is selected with a single input from the navigation key assembly.

The invention relates to mobile terminals. According to the invention, the communication terminal includes a predictive editor application for entering text. The editor is used for editing text for message handling, phonebook editing and searching, etc. The terminal further includes compression and / or decompression software. The invention further relates to a method of compressing data in a mobile terminal on which a language dependent dictionary is stored.

Disclosed is electronic equipment (e.g., a communication device) and method for acquiring remote dictionary information for use in SMS applications. The electronic equipment includes a processor that executes an application program within memory, the application program when executed causing the electronic equipment to: request remote dictionary information related to the user information from an associated remote server; receive remote dictionary information from the associated remote server related to the user information; and output the remote dictionary information in a user-sensible format. The method includes receiving one or more symbols from an input device; requesting remote dictionary information related to the received one or more symbols from a predictive text dictionary application stored on an associated remote server; and receiving and outputting the remote dictionary information in a user-sensible format.

An electronic mail client includes features for preventing addressing errors in electronic messaging such as those caused by predictive text features. A messaging client tracks addressing parameters including the length of time since each previous addressee has been messaged, the quantity of times selected addressees have been co-addressees on messages, if any, whether the addressees are designated as sensitive and whether the messages contain sensitive subject matter. If a high risk of addressing error is determined, the client delays transmission of the message to permit the user to review the message recipient addresses and correct any erroneous addressees.

The present invention includes systems and methods for retrieving information via a flexible and consistent targeted search model that employs interactive multi-prefix, multi-tier and dynamic menu information retrieval techniques (including predictive text techniques to facilitate the generation of targeted ads) that provide context-specific functionality tailored to particular information channels, as well as to records within or across such channels, and other known state information. Users are presented with a consistent search interface among multiple tiers across and within a large domain of information sources, and need not learn different or special search syntax. A thin-client server-controlled architecture enables users of resource-constrained mobile communications devices to locate targeted information more quickly by entering fewer keystrokes and performing fewer query iterations and web page refreshes, which in turn reduces required network bandwidth.

A method and system for populating a predictive text dictionary is provided. A connection between a handheld electronic device and a network is detected. The handheld electronic device is operable to allow a user to enter text. The handheld electronic device has a predictive text dictionary that is operable to receive and employ sets of words. User preferences for the handheld electronic device are retrieved. The predictive text dictionary of the handheld electronic device is populated with a set of words at least partially based on the user preferences.

Systems and methods for interactive, user-driven detection, creation and completion of form fields in a digital document are provided. A document with form fields that require completion by a user is received, after which form fields are detected at the direction of the user. Once the user selects a possible form field, the system creates the appropriate fillable form field based on size, type, location, related text and other parameters of the form field and surrounding document. Additional levels of interaction include predictive text, pattern development and automatic completion of previously completed fields.

Languages based in whole or in part on ideographic characters such as Chinese, Japanese, and Korean, are often are entered in a computerized text-entry system in a two-phase process. In the first phase, symbols from a first pre-conversion set are entered, then in the second phase, these pre-conversion symbols are converted into a second set of post-conversion symbols. This invention teaches a method and apparatus for the automatic conversion of pre-conversion symbols into post-conversion symbols without requiring an explicit conversion signal to be input by the user. It accomplishes this goal though the design of trigger sequences of keystrokes which are substantially functionally equivalent to an explicit conversion signal input by the user. An apparatus constructed according to the trigger sequence method is particularly well adapted for use on reduced keyboards, and in conjunction with predictive text-entry methods. Explicit constructions are shown for Chinese, Japanese, and Korean.

The invention relates to a neural network text classification method based on a multi-knowledge map. The method comprises the following steps: inputting a text in a training set into a long-term and short-term memory network to obtain a context vector of the text; extracting entities from each text in the training set, and performing entity matching in the knowledge map; calculating the attentionweights of matched entities and relationships in the knowledge map under context vector to obtain the overall entity vector, the overall relationship vector and the fact triple vector of the text; computing the fact triple vectors under different knowledge maps, calculating the attention weights of these fact triples, obtaining the text representation vectors and inputting them to the full connection layer of the neural network, and using a classifier to calculate the probability of each text belonging to different categories to train the network; using the trained deep neural network model to predict the category of the text to be predicted. This method improves the model's understanding of text semantics and can classify text content more reliably, accurately and robustly.

The invention discloses a fused attention model-based Chinese text classification method. The method comprises the following steps of: respectively segmenting a text into a corresponding word set anda corresponding character set through word segmentation preprocessing and character segmentation preprocessing, and training a word vector and a character vector corresponding to the text by adoptionof a feature embedding method according to the obtained word set and character set; respectively carrying out semantic encoding on the word vector and the character vector by taking a bidirectional gate circulation unit neural network as an encoder, and obtaining a word attention vector and a character attention vector in the text by adoption of a word vector attention mechanism and a character vector attention mechanism; obtaining a fused attention vector; and predicting a category of the text through a softmax classifier. The method is capable of solving the problem that more redundant features exist in the classification process as existing Chinese text classification methods neglects character feature information of texts, the extracted texts are single in features, all the pieces of semantic information of the texts are difficult to cover and features having obvious contribution to the classification are not focused.

The invention discloses a multi-triad joint extraction method based on knowledge graph embedding, comprising the following steps of: processing an acquired text statement to obtain a text statement matrix; inputting the text statement matrix into a Transformer model to extract semantic information of text statements to obtain semantic feature vectors; applying the semantic feature vectors to an entity recognition sequence labeling task to obtain entity recognition cross entropy loss loss1; applying the semantic feature vector to a relationship classification task, and solving entity recognition cross entropy loss loss2 of relationship classification; constructing an entity word relationship by utilizing an entity labeling prediction matrix and a statement entity word relationship classification matrix, and solving cross entropy loss loss3 of the relationship; calculating a minimized total loss function loss by utilizing an optimization algorithm based on gradient descent of the loss1,the loss2 and the loss3; and obtaining a trained Transformer model according to the text statement to be predicted, inputting the text statement to be predicted into the trained Transformer model to obtain a predicted semantic feature vector of the predicted text statement, and completing a multi-triad joint extraction method.

A device (110) used for composing, compressing and transmitting messages by way of a data network (100) comprises means (116) for employing predictive text entry during composition of a message (MSG) and compressing the composed message for transmission. Increased redundancy and improved compression efficiency result from having the predictive text entry program (201) suggest character strings (207) derived from a corpus of messages (204) that serves also as a basis for a statistical model (206) used for compression. A messaging system comprising the composition device (110) and a device for receiving and decompressing the message may comprise a messaging (MSG) hub (108) for decompressing messages (MSG) from the composing device and reformatting them, for example as e-mail messages (E-MSG), before transmitting them to the addressee (112, 113, 114, 115) and, conversely, compressing messages from the addressee using a similar corpus of messages (204) before transmitting them to the composing device (110). Peer-to-peer messaging (MSG′) may be provided between two users using similar devices capable of composing, compressing and transmitting messages (110′), and received and decompressing messages (110″).

The invention discloses a plain text oriented enterprise entity classification method. The plain text oriented enterprise entity classification method comprises the steps of S1, carrying out type labeling for the enterprise entities in collected plain text data and regarding the enterprise entities being subjected to type labeling as a training set of an enterprise entity identification module; carrying out type labeling for the enterprise entities in the collected plain text data according to business nature and regarding the enterprise entities being subjected to the type labeling as a training sample set of an enterprise entity classification module; and S2, carrying out enterprise entity identification model training through a condition random field model to obtain an enterprise entity identification model; S3, carrying out semantic vectorization construction for the text data of an original training set; S4, training by regarding the data of the training set after being subjected to type labeling and semantic vectorization as training parameters to obtain an enterprise entity classification model; and S5, classifying the enterprise entity in a to-be-predicted text by utilizing the enterprise entity classification model. According to the plain text oriented enterprise entity classification method, as the obtained semantic vector serves as the feature of the entity, dependence on artificial features and external data is reduced, and the universality and robustness are guaranteed.

A method and system for populating a predictive text dictionary is provided. A connection between a handheld electronic device and a network is detected. The handheld electronic device is operable to allow a user to enter text. The handheld electronic device has a predictive text dictionary that is operable to receive and employ sets of words. User preferences for the handheld electronic device are retrieved. The predictive text dictionary of the handheld electronic device is populated with a set of words at least partially based on the user preferences.

A method of providing a predictive text function in a mobile terminal includes detecting entry of one or more alphanumeric characters via a user interface in a currently-executing first mobile terminal application. At least one character string including the one or more characters is identified among words in a default database associated with the first mobile terminal application and among words in an additional database associated with a second mobile terminal application. The additional database includes at least one word that is not found in the default database. Accordingly, the at least one character string including the one or more characters is displayed via the user interface in the first mobile terminal application. Related devices and computer program products are also discussed.

Standard word lists that are often used for such operations as predictive text, spell checking, and word completion are based on general linguistic data that might not accurately reflect actual text usage patterns of particular users. Systems and methods of building and using a custom word list for use in text operations on an electronic device are provided. A collection of text items associated with a user of the electronic device is scanned to identify words in the text items. A weighting is then assigned to each identified word, and the words and corresponding weightings are stored.

An apparatus (600) is provided with a first (635) and a second (645) plurality of input keys. At least some input keys of the first plurality of input keys are associated with both a numeric value and a first set of alphabetic values, and arranged in a first alphabetic keypad layout based on their respective first sets of alphabetic values. The second plurality of input keys are positioned adjacent to the first plurality of input keys. At least some input keys of the second plurality of input keys are associated with at least one alphabetic value, and arranged in a second alphabetic keypad layout. At least some of the input keys of the first plurality of input keys have a second set of alphabetic values that is used as part of the second alphabetic keypad layout and at least some of the input keys of the first and second pluralities of input keys are arranged to provide a reduced group of words generated by a predictive text input routine for an input key sequence entered by a user.

The invention discloses an arbitrary-shape natural scene text detection method based on deep learning. The objective of the invention is to improve the accuracy and recall rate of text detection of natural scenes of any shape. The method comprises the following steps: constructing and training a natural scene text detection model based on a neural network, and detecting a natural scene text in a given image by using the trained natural scene text detection model. According to the method, a general integrity perception loss function is provided to replace a regression branch smooth l1 loss function, and an index related to the IoU is directly used as an optimized object, so that the detection precision and recall rate are directly improved; according to the method, the TextIoU head is addedin the Mask RCNN head to predict the IoU of the text mask, the predicted value is multiplied by the classification confidence of the Box head to obtain the final confidence score, and the final detection result is obtained by screening according to the final confidence score. By adding the TextIoU head, a false positive result can be effectively inhibited, so that the detection effect is improved.