Abstract: Computer-implemented methods are included for training an autonomous machine to perform a target operation in a target environment. The methods include receiving a natural language description of the target operation and a natural language description of the target environment. The methods further include generating a prompt such as a reward and/or goal position signature by combining the natural language description of a target task or goal and the natural language description of the target environment. The methods then generate a reward or goal position function by prompting a large language model with the generated prompt. The methods further include computing a state description using a model of the target environment, and training a policy for the autonomous machine to perform the target task or goal using the generated function and state description.

Abstract: A confidence estimation system includes: a neural network including at least one an attention module including N heads configured to: generate attention matrices based on interactions between tokens for words in an input sequence of words, the input sequence of words including a word that is obscured; and determine the word that is obscured in the input sequence; and a confidence module configured to determine a confidence value indicative of a probability of the neural network correctly determining the word that is obscured, the confidence module determining the confidence value of the word that is obscured using a convolutional neural network that projects the attention matrices generated by the attention module over a multi-dimensional space, the attention matrices recording interactions between the tokens in the input sequence of words without information regarding the tokens for the words and the word that is obscured.

Abstract: A confidence estimation system includes: a neural network including at least one an attention module including N heads configured to: generate attention matrices based on interactions between tokens for words in an input sequence of words, the input sequence of words including a word that is obscured; and determine the word that is obscured in the input sequence; and a confidence module configured to determine a confidence value indicative of a probability of the neural network correctly determining the word that is obscured, the confidence module determining the confidence value of the word that is obscured using a convolutional neural network that projects the attention matrices generated by the attention module over a multi-dimensional space, the attention matrices recording interactions between the tokens in the input sequence of words without information regarding the tokens for the words and the word that is obscured.

Abstract: A method for encoding documents includes building or otherwise providing a condensed dictionary including identifiers for block headers identified in text blocks extracted from a collection of training documents. For at least one test document a set of text content blocks is identified. For each of the text content blocks in the set, a block header is identified. Each block header in the training and test documents includes a sequence includes no more than a predetermined maximum number of characters. An encoding of the test document is generated, based on the identifiers of the block headers identified in the test document that are in the condensed dictionary.

Abstract: Methods and systems for electronic crowd-based authentication. An example embodiment includes operations or steps for selecting authentication criteria with corresponding validity criteria through a key ring application, pairing the key ring application with authentication data, and transmitting the authentication data for verification by a security server, thereby allowing a trusted user to endorse an unregistered user to provide the unregistered user with access to a secured resource by the selecting of the authentication criteria, the pairing of the key ring application, and the transmitting of the authentication data for verification by the security server.

Abstract: A method for encoding documents includes building or otherwise providing a condensed dictionary including identifiers for block headers identified in text blocks extracted from a collection of training documents. For at least one test document a set of text content blocks is identified. For each of the text content blocks in the set, a block header is identified. Each block header in the training and test documents includes a sequence includes no more than a predetermined maximum number of characters. An encoding of the test document is generated, based on the identifiers of the block headers identified in the test document that are in the condensed dictionary.

Abstract: Methods and systems for electronic crowd-based authentication. An example embodiment includes operations or steps for selecting authentication criteria with corresponding validity criteria through a key ring application, pairing the key ring application with authentication data, and transmitting the authentication data for verification by a security server, thereby allowing a trusted user to endorse an unregistered user to provide the unregistered user with access to a secured resource by the selecting of the authentication criteria, the pairing of the key ring application, and the transmitting of the authentication data for verification by the security server.

Abstract: A system and method provide for querying a knowledge base. The knowledge base includes a plurality of layers, each layer corresponding to a different type of information, such as user's intent, lack of knowledge, and specific issue. The system includes categorization rules for categorizing at least a part of a user's request according to the types of information and for mapping the request to respective layers of the knowledge base. The user's request is categorized with the categorization rules and mapped to one of the layers of the knowledge base, based on the categorization. A search is conducted for responsive problem descriptions among the problem descriptions in that layer. The system and method can thus provide for differential searching of the knowledge base, depending on the type of information being sought.

Abstract: The present disclosure relates to a computer-implemented method, device, and computer-readable storage medium used for contextual linking information in a financial report. The method can include obtaining portions of the financial report; detecting one or more line items in the portions of the financial report based on one or more properties of the one or more line items; detecting one or more section headers in the portions of the financial report based on one or more properties of the one or more section headers; parsing, by a processor, the one or more line items and the one or more section headers that are detected; and linking the one or more line items to the one or more section headers based on the parsing.

Abstract: A computer implemented system and method are disclosed for updating an electronic calendar. The method includes receiving an electronic message in a natural language in which a change in role is expressed and, with a natural language processor implemented by a computer processor, automatically detecting the change in role within the email message, optionally storing the change in role in a contacts database, and proposing updates for entries in an electronic calendar based on the detected change in role.

Abstract: A method for assisting a driver to locate a parking space includes identifying a user-selected destination with a navigation system of a first vehicle. While the driver of the first vehicle is being guided by the navigation system towards the selected destination, parking availability information is acquired from a first space allocation device (e.g., ticket vending machine), associated with a first parking zone. The parking availability information includes parking availability information for at least one second parking zone associated with a respective second, space allocation device, remote from the first space allocation device. The second parking availability information for the second parking zone(s) is communicated to the first space allocation device by at a respective second navigation system of at least one second vehicle. A parking zone is selected, based on the user-selected destination and acquired parking availability information transmitted from the first space allocation device.

Abstract: A system and method provide for querying a knowledge base. The knowledge base includes a plurality of layers, each layer corresponding to a different type of information, such as user's intent, lack of knowledge, and specific issue. The system includes categorization rules for categorizing at least a part of a user's request according to the types of information and for mapping the request to respective layers of the knowledge base. The user's request is categorized with the categorization rules and mapped to one of the layers of the knowledge base, based on the categorization. A search is conducted for responsive problem descriptions among the problem descriptions in that layer. The system and method can thus provide for differential searching of the knowledge base, depending on the type of information being sought.

Abstract: A method, a system, and a computer program product for processing search queries are disclosed. A first query pertaining to a primary theme is received. One or more secondary queries are then generated. Each of the one or more secondary queries is associated with a corresponding plausible secondary theme. The first query and the one or more secondary queries are then submitted to a search engine.

Abstract: A system and method for proposing candidate solutions for updating a knowledge base are disclosed. In the method, knowledge base solutions in a natural language are each processed to generate a first action sequence of atomic steps, each including a verb and an object including a noun which is in a syntactic dependency with the respective verb. A recorded solution, expressed in a natural language, is received which includes actions performed on a device in the device class. The recorded solution is processed to generate a second action sequence of atomic steps, as for the first action sequence. The second action sequence is compared with the first action sequences to determine whether the recorded solution corresponds to one of the knowledge base solutions. Based on the comparison, provision is made for proposing an update to the knowledge base, based on the recorded solution.

Abstract: A method for speech to text transcription uses a knowledge base containing solution descriptions, each describing, in words, a solution to a respective problem. An audio recording of a dialogue between an agent and a user in which the agent had access to the knowledge base is received. A sequence of phonemes based on the agent's part of the audio recording is identified and from this, a preliminary transcription is made which includes a sequence of words recognized as corresponding to phonemes in the identified sequence of phonemes together with any unrecognized phonemes from the phoneme sequence that are not recognized as corresponding to one of the recognized words. The preliminary transcription is revised by replacing one or more of the unrecognized phonemes with a word or words from a solution description that includes words which match adjacent words of the sequence of recognized words.

Abstract: A method for assisting a driver to locate a parking space includes identifying a user-selected destination with a navigation system of a first vehicle. While the driver of the first vehicle is being guided by the navigation system towards the selected destination, parking availability information is acquired from a first space allocation device (e.g., ticket vending machine), associated with a first parking zone. The parking availability information includes parking availability information for at least one second parking zone associated with a respective second, space allocation device, remote from the first space allocation device. The second parking availability information for the second parking zone(s) is communicated to the first space allocation device by at a respective second navigation system of at least one second vehicle. A parking zone is selected, based on the user-selected destination and acquired parking availability information transmitted from the first space allocation device.

Abstract: A system and method for proposing candidate solutions for updating a knowledge base are disclosed. In the method, knowledge base solutions in a natural language are each processed to generate a first action sequence of atomic steps, each including a verb and an object including a noun which is in a syntactic dependency with the respective verb. A recorded solution, expressed in a natural language, is received which includes actions performed on a device in the device class. The recorded solution is processed to generate a second action sequence of atomic steps, as for the first action sequence. The second action sequence is compared with the first action sequences to determine whether the recorded solution corresponds to one of the knowledge base solutions. Based on the comparison, provision is made for proposing an update to the knowledge base, based on the recorded solution.

Abstract: A system and method for assisting a user to enter data are provided. The method includes detecting a user's selection of a moveable input zone on each of one or more input documents displayed on the user's screen and detecting a user's selection of a moveable target zone on a target document displayed on the user's screen. Data in the input zone(s) and the target zone is captured for one or more locations of the respective zones and a pattern is learned automatically for filling additional data in the target document based on the captured data. A fill arrangement for filling the additional data in the target document based on the learned pattern is generated, which can be presented to the user for validation.

Abstract: A computer implemented method for calculating the probability of an employee leaving an organization is provided. Closely associated groups of employees within the organization are identified on based on their date of joining the organization. The email traffic among different members of the closely associated group is monitored over a predefined time period. The occurrence of an event in the closely associated group is used to compute a risk parameter for the remaining employees of the closely associated group. The risk parameter is used in conjunction with a first parameter and a second parameter to calculate the probability of an employee leaving the organization.

Abstract: A method, a system, and a computer program product for processing search queries are disclosed. A first query pertaining to a primary theme is received. One or more secondary queries are then generated. Each of the one or more secondary queries is associated with a corresponding plausible secondary theme. The first query and the one or more secondary queries are then submitted to a search engine.