International Business Machines Corporation

Marcio Ferreira Moreno

Rafael Rossi De Mello Brandao

Renato Fontoura De Gusmao Cerqueira

Renato Fontoura De Gusmão Cerqueira

Guilherme Augusto Ferreira Lima

Daniel Salles Chevitarese

Lucas Correia Villa Real

Rafael R De Mello Brandao

Daniel Salles Civitarese

Ana Fucs

Helon Vicente Hultmann Ayala

Vagner Figueredo De Santana

Paulo Rodrigo Cavalin

Bianca Zadrozny

Emilio Ashton Vital Brazil

Viviane Torres Da Silva

Vinicius Costa Villas Boas Segura

Breno William Santos Rezende De Carvalho

Juliana Jansen Ferreira

Wallas Henrique Sousa Dos Santos

Juliana Soares Jansen Ferreira

Isabela Drabik Chaves Chambers Ramos

Daniela De Mattos Szwarcman

Rodrigo Costa Mesquita Santos

A plurality of datasets may be received. Multiple instances of a machine learning algorithm may be run to create corresponding multiple machine learning models trained for a specific task in a given domain. Each of the multiple instances may use a different subset of the plurality of datasets in training the corresponding machine learning model. The multiple machine learning models may be run with input data. The multiple machine learning models produce corresponding multiple outcomes. A candidate machine learning model may be determined based on comparing each of the multiple outcomes with ground-truth output. A value associated with the different subset of the plurality of datasets may be determined based on comparing of each of the multiple outcomes with ground-truth output. A smart contract may be created which may include the value and the different subset of the plurality of datasets. The smart contract may be recorded in a blockchain.

Data platform to protect security of data

A system and a method of generating and using a multimodal knowledge representation within an artificial intelligence computing system may include prompting a selection of nodes in a graphical visualization; defining an adaptation strategy based; integrating an adaptation strategy representation with the selection of nodes to generate a multimodal knowledge representation; traversing the selection of nodes in response to a user query; performing an adaptation on the selection of nodes; and updating the knowledge representation based on user feedback to a result generated in response to the user query.

Multimodal knowledge consumption adaptation through hybrid knowledge representation

An ore content prediction system is provided. The system receives structured geological data that is derived based on spatial geological information that is associated with an input region. The received structured geological data includes a plurality of multidimensional tensors that are derived from spatial geological information of a plurality of sub-regions of the input region. The spatial geological information includes one or more types of data. The system trains a prediction model to produce a prediction output based on an average grade of an ore of a target mineral type at a target region by using the received structured geological data. The system identifies a relationship of the structured geological data to the prediction output and determines a revised input region based on the identified relationship.

Estimate ore content based on spatial geological data through 3D convolutional neural networks

A modular node includes a semiconductor chip, including modular sensors to detect events or changes in environment, modular actuators for moving or controlling an object, a non-transitory computer readable medium storing a program, a processor executing the program configured to control module networking and setup, and a neuromorphic chip configured to receive information from the modular sensors and modular actuators, autonomously process the information received from the modular sensors and actuators, determine a validity of the information received from the modular sensors and actuators, and autonomously communicate with neighboring modular nodes for network auto-arrangement.

Smart sensing: a system and a method for distributed and fault tolerant hierarchical autonomous cognitive instrumentation

Automatically mapping and combining the application of machine learning models to answer queries according to semantic specification. A query is parsed to extract keywords from the query and to contextualize the query. Based on the keywords, machine learning models are selected that process concepts associated with the keywords. The machine learning models are sorted according to the contextualization of the query. The machine learning models are run on multimodal data according to a sorted order, where data resulting from an output of one of the machine learning models is used as input to another one of the machine learning models. A query result is output based on a result from running the machine learning models.

Mapping machine learning models to answer queries

Entities and temporal information associated with the entities can be extracted from the documents using natural language processing. A graph structure can be created representing the document's temporal semantics, nodes of the graph structure including the entities and edges of the graph structure representing temporal relationships between the nodes. The graph structure can be linked with the document. Multiple documents can be received and a knowledgebase can be created including multiple graph structures representing the multiple documents according to the multiple documents' temporal semantics. An input document for query can be received and transformed into a graph structure for query, the graph structure for query representing the input document's temporal semantics. The knowledgebase can be searched for a matching document having a graph structure similar to the graph structure for query based on a similarity threshold. The matching document can be output.

Retrieving documents according to temporal similarity

A computer-implemented method of performing an incremental specification of a query includes extracting text from each of a plurality of participants in a dialog. A contextual information is determined of the extracted text of one or more of the plurality of participants. A dialog understanding operation is performed by processing at least the contextual information of the extracted text in a knowledge graph to identify in the dialog at least one or more of a structural gap, an information about entities, relationships, and actions. Query information is provided responsive to the dialog for at least one of filling the identified structural gap, or for providing additional information about one or more of the identified entities, relationships or actions in the dialog.

Automatically and incrementally specifying queries through dialog understanding in real time

One or more processor can automatically identify, structure and retrieve spatial and/or temporal sequences of digital media content according to semantic specification. Digital media content can be received and information from digital media content can be extracted. Based on the information, a knowledge graph can be constructed or structured to include at least one of spatial and temporal representation of the digital media content. A search query can be received associated with the digital media content. Based on traversing the knowledge graph structure according to at least one of spatial and temporal criterion mapped from the search query, new digital media content can be composed which meets the search query.

Spatiotemporal sequences of content

A hybrid knowledge representation is searched for a machine learning component corresponding to a search query. The hybrid knowledge representation may be structured as nodes representing machine learning workflow components and edges (e.g., links) connecting the nodes based on relationships between the nodes. Responsive to finding the machine learning component in the hybrid knowledge representation, the machine learning component is returned. Responsive to not finding the machine learning component in the hybrid knowledge representation, the hybrid knowledge representation is searched for machine learning model fragments associated with building the machine learning component, generating a new machine learning component by combining the machine learning model fragments and returning the new machine learning component.

Machine learning engineering through hybrid knowledge representation

Embodiments of the invention are directed to a computer-implemented method of generating a pathway recommendation. The computer-implemented method includes using a processor system to generate an intermediate three-dimensional (3D) virtual reality (VR) environment of a target environment. A machine learning algorithm is used to perform a machine learning task on the intermediate 3D VR environment to generate machine learning task results including predicted features of interest (FOI) and FOI annotations for the intermediate 3D VR environment. The processor system is used to generate, based at least in part on the machine learning task results, the pathway recommendation configured to assist a user with navigating and interpreting a 3D VR environment including the intermediate 3D VR environment having the pathway recommendation.

Route recommendation that assists a user with navigating and interpreting a virtual reality environment

Document retrieval through assertion analysis on entities and document fragments is disclosed. A document is received. Logical structures and entities are extracted from the document by parsing the document. For an entity in the extracted entities, an object representing the entity is created, an assertion made in the document associated with the entity is determined, and the assertion is linked to the object representing the entity. A logical structure from the extracted logical structures and content of the logical structure containing the assertion are identified and linked to the object representing the entity.

Document retrieval through assertion analysis on entities and document fragments

A system, product, and method including automatically performing extraction of slides from multimodal content, performing object extraction from each of the slides, allowing object substitution through semantics and concepts of the objects extracted, processing audio synchronized with the slides enriched with cognitive computing, search engine, and knowledge base, to provide annotations of the slides, processing the audio synchronized with the object being presented in each slide to enhance semantics and understanding, and curating for each step with human-machine interaction to provide a learning process by the system.

System and method to extract and enrich slide presentations from multimodal content through cognitive computing

A document may be received. An ontology associated with a domain may be identified or retrieved. A set of concepts in the document may be extracted, the set represented in the ontology. A set of concepts selected from the ontology may be received. A search may be performed based on the set of selected concepts.

Document analogues through ontology matching

A system and associated processes may generate new modalities by receiving content associated with an original modality. An appropriateness of an alternative modality may be determined based on at least one of a device available to present the content, a user profile, and a contextual factor regarding an environment of a user intended to receive the content. Source content semantics associated with the content may be automatically determined, and an alternative modality based on the determined source content semantics may be generated and used to present the content to the user.

Alternative modalities generation for digital content based on presentation context

Method and apparatus for monitoring collaboration parameters and maintaining user profiles using cognitive computing techniques are provided. The method includes receiving a domain for a first collaborative session, and retrieving a plurality of trait profiles associated with potential participants. The method further includes determining a subset of trait profiles based in part on comparing each respective trait profile with the rest of the trait profiles, and providing the subset of trait profiles. Additionally, the method includes receiving an indication that the subset of trait profiles is acceptable, and a request to initiate the first collaborative session with participants corresponding to the subset of trait profiles. The method includes initializing a plurality of sensor devices, detecting a first contribution from a first participant, classifying the first contribution using cognitive computing techniques, identifying a trait profile corresponding to the first participant, and updating the identified trait profile.

Growing community of inventors

Rio de Janeiro, Brazil

Marcio Ferreira Moreno