170 results about "Logical reasoning" patented technology

The invention discloses a natural language semantic analysis system and method based on the depth neural network. The method comprises the steps that a knowledge map is built, a training set is inputted, and an N-Gram probability model is obtained, a matrix is obtained as an input by representing words as vectors using the word2vec, a deep belief network model is used for the entity identification and the input validation set, the classifier parameters and the input test set are adjusted, the group abilities of the models are tested, the knowledge graph method is adopted to apply reasoning to the entities in the descriptions of the language, and corresponding conclusions are obtained. Compared with the prior art, the natural language semantic analysis system and method based on the depth neural network uses the knowledge graph method to apply reasoning to the entities in the descriptions of the language and to obtain the corresponding conclusions, so that our natural language understanding abilities are provided not only with the capacity to understand the literal meaning, but also with logical reasoning and the understand of the meaning on a deep level, and the method has promotable and practical value.

A method, system, and computer program product provides simultaneous debugging of multiple OS image and/or system dump pairs in a distributed storage repository. A management console receives a terminal debugging session request and a cluster selection from an interface and starts a debugger instance. The debugger instance autonomously identifies client LPARs and loads the system dump images assigned to the client LPARs. In response to receiving a selection of a first and second client LPARs, the debugger analyzes the first and second system dump images, respectively, and calculates relational information between the first analysis and the second analysis via one or more logical reasoning utilities of the management console. The debugger then loads the relational information to the management console interface with an analysis of one or more similarities between the first and second system dumps.

The invention belongs to the technical field of computer word representation, and discloses a deep dynamic context word representation model and method. The deep dynamic context word representation model is a masking language model formed by stacking multi-layer bidirectional Transformer encoders with layers of attention mechanisms. The network is a multi-layer neural network, and each layer of the network captures context information of each word in the input statement from different angles. Then different weights are given to each layer of the network through a layer attention mechanism. Finally, different word representations are synthesized according to the weights to form context representations of the words. Words generated by using the model represent three tasks of logic reasoning(MultiNLI), named entity recognition (CoNLL2003) and reading understanding (SQuAD) on a public data set, and the tasks are improved by 2.0%, 0.47% and 2.96% respectively compared with an existing model.

The invention relates to the technical field of artificial intelligence, and provides a ten-thousand-level intention classification method and device, a storage medium and electronic equipment. The ten-thousand-level intention classification method comprises the steps of obtaining a dialogue statement of at least one round of dialogue with a user; performing context analysis on the dialogue statement to complement the context information of the dialogue; performing semantic analysis on the dialogue statement to obtain a plurality of candidate intentions of the user; and based on the pluralityof candidate intentions and the complemented dialogue context information, determining a real intention of the user by utilizing an intention decision model constructed based on a reinforcement learning algorithm. The method is realized based on a brand-new three-layer man-machine conversation technology framework. According to the method, the plurality of candidate intentions are obtained throughsemantic analysis on a semantic understanding layer, complemented context information is obtained through context analysis on a logical reasoning layer, dynamic decision making is conducted on a decision judgment layer through an intention decision making model according to the candidate intentions and the complemented context information, and the accuracy of intention classification through themethod is high.

The invention relates to a mechanical arm kinematics formal analysis method. The method comprises the steps of 1, determining structural parameters of a mechanical arm mechanism; 2, using high-order logical language for building a mechanical arm kinematics logical model; 3, using a logical formula for describing kinematics constraints or attributes needing to be verified; 4, enabling the mechanical arm mechanism kinematics logical model and the kinematics constraints or attributes needing to be verified to form a logical proposition; 5, utilizing a logical reasoning engine for proving whether the logical proposition is true or not; if the logical proposition is true, it represents that the model meets the kinematics constraints or have the attributes; if not, it represents that the model does not meet the kinematics constraints or does not have the attributes. According to formal verification, on the basis of system formal specifications or attributes, the correctness of the system is verified by using the mathematical method, and the verified attributes are precise and complete.

The invention discloses a product allocator and constructing method, which comprises the following steps: building product knowledge base based on logical language; realizing intelligent option and highly effective allocation calculation; designing logic rational regulation and real-time detecting function to realize effectively rapid definition of logic relationship among characteristic parameters; affirming the logic relationship of the parameter not mutually contradiction; making the product allocator have certain rational intelligence; realizing to extract the BOM making information in the allocated BOM. The main content of knowledge base of characteristic parameter contains product element file, allocated BOM, characteristic parameter information and exclusive or coexistent logic relationship information among characteristic parameters.

The invention discloses a multi-source data aggregation sampling strategy based on a big data environment. The method comprises the following steps that on the basis of a multi-source data set encoded by GBK, fusion of multiple attributes in one data set and fusion of different source data sets are achieved, and multi-source data fusion operation is completed; fused files are segmented to form a two-dimensional word frequency matrix of file words; high-frequency words are displayed to a user for reference, and random migratory selection of seed root node words required by snowball sampling is conducted; business-goal-oriented seed root node keywords are selected, and the depth of snowball sampling is input; on the basis of seed root node data, a balance check value is set, and the words are circularly matched for snowball sampling; a directed acyclic graph and an adjacent matrix are constructed; a root node clustering network diagram and a logical reasoning diagram which are related to business goals can be obtained. The interference of sample noise on subsequent reasoning can be reduced through the strategy.

The invention discloses a wireless network switching management method and system; the method uses software to define network inherent centralized control and programmability advantages, and fully considers influences on the network switching management strategy performance by multi factors such as the network, users and applications; the method selects a plurality of influence factors as the input parameters, such as user mobility, network states and user preferences, senses network demands in many aspects, and uses fuzzy logical reasoning to obtain the final network scoring result so as to provide a network switching final decision. The method considers objective factors on the network side and the user side, and considers user subjective customized demand preferences, thus providing differentiated network switching management services for different users.

The invention relates to an electric power equipment fault early-warning method which comprises the following steps of acquiring acting quantity information of a power grid in a fault occurrence time, namely switch information and protection action information; acquiring action information from a data acquiring and monitoring system and a circuit breaker, and obtaining fault occurrence time of each action node of a suspected faulted component; discriminating faulted equipment in the power grid in a real-time online manner and generating fault brief information; combining the switch information, the protection action information and the fault brief information by a transformer substation for forming transformer substation fault data information; receiving and analyzing uploaded fault data information by a scheduling center, performing real-time diagnosis on the fault equipment of the power grid according to a logical reasoning algorithm, and performing comparison checking on the diagnosis result and the fault diagnosis result of the transformer substation. The electric power equipment fault early-warning method has advantages of acquiring fault information in the power grid equipment in time, shortening accident processing time, ensuring safe and reliable operation of the power grid, ensuring high safety of power grid equipment and preventing abnormity of the power grid.

The invention discloses an end-to-end dialogue system and method based on dialogue state guidance and knowledge base retrieval. The system comprises: a dialogue historical information encoding moduleused for encoding sentence-level semantic information by adopting a Transformer structure and encoding dialogue-level semantic information by adopting a GRU structure so as to output a dialogue-leveldistributed vector according to input dialogue historical information; a dialogue state tracking module used for extracting round-level information, extracting the inter-round information to obtain the probability distribution of the slot value on the current slot, and taking the maximum probability as the final slot value; a knowledge base retrieval module used for determining an accessed database and determining attribute entries in the accessed database; and a dialogue reply decoding module used for introducing a copy mechanism to decode so as to alleviate the OOV problem, and controlling the source of the copy by using a gating mechanism. The system has the advantages of accurate database positioning, accurate reply and logical reasoning capability.

The invention discloses an intelligent equipment fault diagnosis method based on semantic reasoning. According to the method, on the basis that a cloud distributed system is used for storing and rapidly processing mass manufacturing data generated by the intelligent equipment, firstly, a Profee graphical modeling tool is used for building an intelligent equipment semantic model, and the semantic model is loaded and edited through a Jena toolkit; reading the real-time data of the equipment through the JDBC, and associating the real-time data of the equipment in the database with the model; self-defining a logical reasoning rule in Jena, and performing logical reasoning by utilizing a reasoning machine to realize fault diagnosis of the intelligent equipment; finally, the fault diagnosis result is intelligently visualized, and a proper solution is recommended. By adopting the method, the fault diagnosis efficiency and level of the intelligent equipment in the intelligent factory can be remarkably improved, the maintenance time of the equipment is shortened, and the production benefit is maximized.

The invention discloses a complex network link prediction method and system based on logical reasoning and graph convolution. The method comprises the following steps: constructing a knowledge graph corresponding to a complex network, and obtaining a training set; performing relation reasoning on each entity pair in the training set through a first-order logical reasoning network with default, and obtaining a relation confidence coefficient matrix through mapping; based on the relation confidence coefficient matrix, performing iterative training on a graph convolutional neural network based on iterative attention through a centralized training decentralized execution mechanism and a local relation attention mechanism to obtain first probability distribution; calculating second probability distribution according to a relation weight matrix and a relation confidence coefficient matrix output by network iteration; obtaining a Wasserstein distance between the first probability distribution and the second probability distribution according to a joint evaluation function; iteratively updating the two networks according to a Wasserstein distance to obtain a link prediction model; and complementing the knowledge graph according to the link prediction model. The link prediction efficiency is high.

The invention discloses a logical reasoning mechanism being suitable for a legal expert system, effectively overcomes the malady that the current legal expert system is based on regulation, and solves the problems that the majority of legal expert systems can not apply judging knowledge for reasoning, only provides simple explanation through regulation feedback, the problems that uncertainty how to stimulate the intuition, experience and reasoning result and other problems are expressed dissatisfactorily. The reasoning mechanism introduces a correlation theory and a method of fuzzy reasoning, abandons simple positive and negative two-value features of traditional logic, regards the logical world into gray level with continuous changes, allows a proposition to be this or that, and exists partial positive and partial negative. The invention provides a more natural and more reasonable operation environment, can describe target system behavior adopting a fuzzy logic technology, imitates people to judge and control behavior technology by cognition and experience, can transform some natural languages into algorithms of computer operation to achieve the purpose of using a simple program to process complex systems which can not be processed or is difficult to be processed by traditional control method. The fuzzy logic correlation theory and method are applied in the legal expert system, can well overcome the shortcomings of knowledge expression, reasoning mechanism and knowledge acquisition and the like of the traditional method in the legal expert system, are more convenient to express numerous uncompleted and inaccurate knowledge and relationship in legal fields, more appropriately describe uncertainty in legal reasoning, makes up for defects of traditional reasoning method, and are suitable for large-scale popularization.

A system and method defining the twelve modular functions of cognition, and using this information to identify individual work style preferences and to facilitate individual and leadership competence development programs. The system and method associates sets of two related pairs of the cognitive functions with each of the three cognitive sectors of Reasoning logically, Relating to others, and taking Action. A preference survey instrument includes sets of any combination of statements, words, questions, or behaviors related to traits common to each of the functions, and defines a set of six preferred functions based on responses to each item in the instrument.

The embodiment of the invention discloses a knowledge graph mining method, device and equipment and a storage medium. The method comprises the following steps of generating each candidate entity representation pair according to a candidate feature set of an entity and/or each user dialogue corpus; determining the weight of each candidate entity characterization pair according to entity informationand characterization information included in each user dialogue corpus; and determining the key representation of the entity according to the weight of each candidate entity representation pair. According to the embodiment of the invention, the subjective influence factors when a worker determines the entity and the representation relationship are avoided, the labor experience cost for determining the entity and the representation relationship and the complex logical reasoning process are reduced, and the determination efficiency and the accuracy of effective representation under the entity are improved.