British Telecommunications Public Limited Company

British Telecommunications Corporation

Robert A Ghanea-Hercock

Alexander L Healing

Hakan Duman

Fang Wang

Yaoru Sun

A computer implemented method of executing a plurality of discrete software modules each including a machine learning algorithm as an executable software component configurable to approximate a function relating a domain data set to a range data set; a data store; and a message handler as an executable software component arranged to receive input data and communicate output data for the module, wherein the message handler is adapted to determine domain parameters for the algorithm based on the input data and to generate the output data based on a result generated by the algorithm, the method including providing a communication channel between modules in order that at least part of output data for a first module constitutes at least part of input data for a second module so as to create a network of modules for combining machine learning algorithms to refine the approximation of the function.

Interoperation of machine learning algorithms

A computer implemented method of executing a plurality of discrete software modules each including a machine learning algorithm as an executable software component configurable to approximate a function relating a domain data set to a range data set; a data store; and a message handler as an executable software component arranged to receive input data and communicate output data for the module, wherein the message handler is adapted to determine domain parameters for the algorithm based on the input data and to generate the output data based on a result generated by the algorithm, each module having associated a metric of resource utilization by the module, the method including receiving a request for a machine learning task; and selecting a module from the plurality of modules for the task based on the metric associated with the module.

Load balancing of machine learning algorithms

A computer implemented method of executing a software module includes a machine learning algorithm as an executable software component configurable to approximate a function relating a domain data set to a range data set; a data store; and a message handler as an executable software component arranged to receive input data and communicate output data for the module, wherein the message handler is adapted to determine domain parameters for the algorithm based on the input data and to generate the output data based on a result generated by the algorithm, the method including generating a message as input data for the module, the message including instructions for execution by the module to effect a modification of the machine learning algorithm of the module.

Adaptation of machine learning algorithms

A computer implemented method of validating use of a computing resource by a an executing requester software module from a plurality of discrete software modules, the method including validating a characteristic of the requester software module; generating a first transaction defining criteria for consumption of the computing resource by the requester software module, the first transaction being encrypted with a private key from a public key/private key pair and being added as part of a block of transactions to a blockchain data structure; generating a subsequent encrypted transaction corresponding to a request of the requester software module to consume the computing resource, the subsequent transaction referring to the first transaction, wherein the subsequent transaction is validated by a transaction miner computing component from a plurality of miners by authenticating the transaction using the public key and verifying compliance with the criteria defined in each transaction.

Management of interoperating machine learning algorithms

A computer implemented method for identifying a signal in a sequence of numerical values, the method including: converting the sequence of numerical values into a vector sequence by converting each value in the sequence to a normalized vector representation; generating a vector field comprising the vector sequence as a multi-dimensional data structure such that vectors in the vector sequence are plotted in sequence in the vector field; modeling a simulated flow of particles through the vector field such that a flow of each particle is influenced by vectors in the vector field; generating one or more static representations of the particles, each representation being generated at a particular point in time; and performing feature detection on each of the one or more static representations to identify features in the representation corresponding to the signal.

Signal identification

A method of manufacturing an article with integral active electronic component uses an additive manufacturing process to: a) form a non-electrically conductive substrate; b) form a non-electrically conductive perforated layer having an aperture; c) form electrically conductive anode and cathode elements spaced in the aperture; d) deposit a conductive electrical connection to each of the elements suitable for imparting an electrical potential difference between the elements; e) form a non-electrically conductive sealing layer atop the perforated layer so as to retain and seal the aperture in the perforated layer.

Printed active device with a 3D thermionic electronic component

A data source software component generator apparatus for generating a representation of one or more data sources for selection from a plurality of data sources to satisfy a data dependency of a software service, each data source including a definition of at least one semantic identifier corresponding to data accessible via the data source, the data sources being represented organized into clusters of multiple data sources based on a semantic association between semantic identifiers of data sources in a cluster, each cluster being represented as one or more data structures, and the data dependency being defined by a specification including one or more semantic identifiers corresponding to data required for execution of the software service, the apparatus comprising: a data source encapsulator unit adapted to encapsulate each cluster as an executable software component; a semantic identifier selection unit adapted to select, from a set of semantic identifiers for all data sources represented in a cluster of a software component, a proper subset of the set of semantic identifiers based on at least one predetermined semantic identifier selection criterion; a software component configuration unit adapted to configure a software component to advertise semantic identifiers to components external to the software component, and provide an interface accessible by components external to the software component, the software component being adapted to deliver data from data sources in the cluster of the software component via the interface, such that, in operation, the apparatus generates and configures executable software components for selection of one or more software components to provide data for the software service based on the advertised semantic identifiers so as to satisfy at least part of the data dependency of the software service.

Selection of data sources for software services

A method of manufacturing an article with integral active electronic component includes using an additive manufacturing process to: a) form a non-electrically conductive substrate; b) form a non-electrically conductive perforated layer having an aperture; c) form electrically conductive anode and cathode elements spaced in the aperture; d) deposit a conductive electrical connection to each of the elements suitable for imparting an electrical potential difference between the elements; and e) form a non-electrically conductive sealing layer atop the perforated layer so as to retain and seal the aperture in the perforated layer.

Method of manufacturing an article with integral active electronic component

An additively manufactured apparatus having a gas filled sealed cavity containing at least two additively manufactured cathodes and an additively manufactured anode spaced from the cathodes such that a continuous electric discharge of the gas stimulated between at least one of the cathodes and the anode provides a Boolean function output at the anode corresponding to electrical input signals at two of the cathodes.

Printed logic gate

A network of devices permits data to be stored on the devices and subsequently searched and accessed from any other one of the devices. A plurality of channels are defined to distribute a plurality of access points throughout the network. A mechanism securely assigns authorizations to users to read or write from or to specified ones of the channels, which authorizations are capable of being checked by each of the access points. To write into a channel, a request is made to one of the access points which checks if the requesting user is authorized to write onto the requested channel. If appropriate, the data is associated with the requested channel. To search for data from a particular channel or group of channels, a search request is made to one of the access points which first checks the requesting user is authorized to read from the requested channel or group of channels. If appropriate it issues a corresponding search request or requests to a subset of the devices which checks to see if stored data satisfying the request exists there and, if so, the data is transmitted to the requesting user. Otherwise the search request is forwarded to another one or more of the devices.

Method and apparatus for communicating information between devices

The present invention relates to distributed systems in which resource utilisation decisions depend upon the semi-automatic categorisation of resource descriptions stored in the distributed system. In the principal embodiment, the resource descriptions are web service descriptions which are augmented with tags (i.e. descriptive words or phrases) entered by users and/or by web service administrators. The initial use of automatic categorisation of these descriptions, followed by a user-driven fine-tuning of the automatically-generated categories enables the rapid creation of reliable categorisation of the resource descriptions, which in turns results in better resource utilisation decisions and hence a more efficient use of the resources of the distributed system.

Distributed system

A station used as a node in a peer-to-peer communications system keeps a list of stations to which requests for resources may be sent. The list includes for each member a performance parameter. When a station receives a request for resources it selects at least one station from the list and sends a resource request message to it. The selection of the station from the list may be performed by a random or pseudo-random process such that the probability of selection of any station is a function of the stored parameter. A station receiving such a request, for resources that it can itself supply, sends a reply to the originator. When the originator receives a reply it updates the respective performance parameter: the degree of adjustment may vary according to the quality of reply received.

Peer-to-peer networks

Security is provided in a network system. A message is received from a user, which message requires authentication of the user. An authentication message is sent indicating the identity of the user to an initial software security agent. The software security agent, on receipt of the authentication message, determines whether information relating to the user is stored on a security database associated with the software security agent, and, if so, the software security agent adds an authentication key to the authentication message. The authentication message is sent on to one or more further software security agents. The prior noted steps are repeated with the further software security agent(s) and, if user-related stored security information is found, adding an authentication key to the authentication message and sending the authentication message on to one or more further software security agents until the number of keys associated with the authentication message equals a predetermined number N. When the number of associated keys equals N, the authentication message is sent to the initial software security agent which then grants the required permission to the user.

Network security system

A computer security system uses a plurality of co-operating software agents to protect a network against attack. Individual agents at each node the network co-operatively act to detect attacks and to share attack signatures and solutions via a message exchange mechanism. A global internal measurement of the overall health of the group of agents may be used as an indicator of a possible attack. In larger networks, the agents may be formed a plurality of separate autonomous groups, with a common group identity being automatically maintained by the message passing mechanism. Individual groups may be located by a system designer in separate cells or domains within the network, so that if one cell becomes compromised the rest of the network is not affected.

Growing community of inventors

Ipswich, United Kingdom

Robert A Ghanea-Hercock