Abstract: A computer-implemented method for explaining an image classifier, the method comprising: receiving an initial image, the initial image having been wrongly classified by the image classifier; receiving an initial gradient of a function executed by the image classifier generated while classifying the initial image, the function being indicative of a probability for the initial image to belong to an initial class; converting the initial image into a latent vector, the latent vector being a representation of the initial image in a latent space; generating a plurality of perturbation vectors using the initial gradient of the function executed by the image classifier; combining the latent vector with each one of the plurality of perturbation vectors, thereby obtaining a plurality of modified vectors; for each one of the plurality of modified vectors, reconstructing a respective image, thereby obtaining a plurality of reconstructed images; transmitting the reconstructed images to the image classifier; for each one o

Abstract: The arrangement includes a latch pivotally mounted to cooperate with a catch, a lock member including a locking ratchet arranged to cooperate with the latch between initial and intermediate locking configurations of the latch, following at least one movement control action of the ratchet by an actuator, an ejector arranged to drive, with the ejection thereof, the pivoting of the latch and the lifting of the hood, the latch and the ejector forming an ejection member of the striker. During the movement of the ratchet from the initial configuration to the final configuration, the lock member is arranged to pivot the ejection member in an ejection direction against the resultant torque of the weight of the hood exerted on the ejector via the striker in order to enable the lifting of the hood.

Abstract: A method for strengthening metal structure, including steps: i) mixing two-component epoxy resin composition; ii) applying composition to metal structure surface, or onto strengthening element, or both; iii) contacting applied epoxy resin composition with strengthening element surface and/or metal structure surface wherein composition forms interlayer between metal structure and strengthening element, and iv) curing epoxy resin composition at 100° C. or below; wherein two-component epoxy resin composition contains: first component between 10-50 wt.-% of at least one epoxy resin contains on average more than one epoxy group per molecule; second component between 5-30 wt.-% of hardener for epoxy resins; between 3-25 wt.-% of at least one impact strength modifier in one or both components; between 15-80 wt.-% of at least one filler in one or both components; and wherein two-component epoxy resin composition exhibits after curing Elastic Modulus at least 2500 MPa, and Impact Peel Strength at least 15 N/mm.

Abstract: A method for training a machine learning localization model to localize objects belonging to a given class within an image, the method comprising: receiving images each comprising objects of the given class; and for each image: receiving a heat map generated using the machine learning localization model; identifying proposals each corresponding to a potential object, each proposal having associated thereto an initial probability that the proposal corresponds to the potential object; for each proposal, correcting the initial probability using the heat map; selecting given ones of the proposals having a greatest corrected probability, thereby identifying object candidates; and calculating a loss for the machine learning localization model based on a location of the object candidates within the training image and the heat map; and providing the calculated loss to the machine learning localization model.

Abstract: A compatibility score generator implementing a neural network is trained for assessing compatibility of items. Elements of a feature vector representing each item and of a compatibility data structure indicating items considered compatible are retrieved. The neural network is trained using training data corresponding to the items and indicating compatibility between pairs of items. The compatibility data structure is modified by removing indications that items of a pair of items are compatible. An encoding function generating encoded representations for the items based on the compatibility data structure is evaluated. Encoded representations are provided to a decoder that learns a likelihood that the indication had been removed when modified. The neural network and the decoder are optimized based on a loss function that reflects the decoder's ability to correctly determine whether the indication had been removed. The encoded representations generate a compatibility score for at least two items of interest.

Abstract: Persistent storage contains configuration items representing computing hardware and software, wherein each configuration item is respectively associated with a set of attributes, and wherein pairwise relationships are defined between some of the configuration items. One or more processors are configured to: select a subset of the configuration items that are connected by way of a subset of the pairwise relationships; form a graph representation in which the subset of the configuration items is represented as nodes and the subset of the pairwise relationships is represented as edges between pairs of the nodes; train a graph neural network with k layers on the graph representation, wherein training the graph neural network involves sequentially generating k embeddings for the sets of attributes associated with the nodes, wherein the embeddings are in an f-dimensional feature space; and based a kth of the embeddings, determine that a particular node of the nodes is anomalous.

Abstract: There is provided a method and system for training an embedding model to perform relation predictions in a knowledge hypergraph to output a trained embedding model. A training dataset comprising tuples representing relations between entities in the knowledge hypergraph are received. The embedding model is trained to perform relation predictions for each given tuple from a subset of tuples in the training dataset by generating a respective entity vector for each entity and a respective relation matrix representing relations between the entities. The entity vectors and relation matrix are split into a plurality of windows, and interaction values between elements in each window are calculated. A relation score indicative of the relation in the given tuple being true is calculated. Parameters of the embedding model are updated based on the relation scores for the subset of tuples. The trained embedding model is then output.

Abstract: Disclosed herein a system, a method and a device for generating a voice model for a user. A device can include an encoder and a decoder to generate a voice model for converting text to an audio output that resembles a voice of the person sending respective text. The encoder can includes a neural network and can receive a plurality of audio samples from a user. The encoder can generate a sequence of values and provide the sequence of values to the decoder. The decoder can establish, using the sequence of values and one or more speaker embeddings of the user, a voice model corresponding to the plurality of audio samples of the user.

Abstract: Low-cost and robust platforms are key for the development of next-generation 3D micro- and nanodevices. To fabricate such platforms, nanocrystalline diamond (NCD) is a highly appealing material due to its biocompatibility, robustness, and mechanical, electrical, electrochemical, and optical properties, while glass substrates with through vias are ideal interposers for 3D integration due to the excellent properties of glass. A low-cost process—free of photolithography and transfer printing—for fabricating arrays of TGVs that are sealed with suspended portions of an ultra-thin NCD film on one side is presented. These highly transparent structures may serve as a platform for the development of microwells for single-cell culture and analysis, 3D integrated devices such as microelectrodes, and quantum technologies. It is also possible to replace the NCD with silicon nitride or silicon carbide, allowing for the development of complex heterogeneous structures on a small scale.

Abstract: A computer-implemented method for explaining an image classifier, the method comprising: receiving an initial image, the initial image having been wrongly classified by the image classifier; receiving an initial gradient of a function executed by the image classifier generated while classifying the initial image, the function being indicative of a probability for the initial image to belong to an initial class; converting the initial image into a latent vector, the latent vector being a representation of the initial image in a latent space; generating a plurality of perturbation vectors using the initial gradient of the function executed by the image classifier; combining the latent vector with each one of the plurality of perturbation vectors, thereby obtaining a plurality of modified vectors; for each one of the plurality of modified vectors, reconstructing a respective image, thereby obtaining a plurality of reconstructed images; transmitting the reconstructed images to the image classifier; for each one o

Abstract: A method and a system for joint data augmentation and classification learning, where an augmentation network learns to perform transformations and a classification network is trained. A set of labelled images is received. During an inner loop iteration, an augmentation network applies a transformation on a given labelled image of the set to obtain a transformed image. The classification network classifies the transformed image to obtain a predicted class, and a training loss is determined based on the predicted class and the respective label. The parameters of the classification network are updated based on the classification loss. During an outer loop iteration, the classification network classifies another labelled image of the set to obtain another predicted class, and a validation loss is determined based on the other predicted class and the respective label. The parameters of the augmentation network are updated based on the validation loss.

Abstract: A method and a system for generating training images for training an instance segmentation machine learning algorithm (MLA). A set of image-level labelled images are received, where a given image is labelled with a label indicative of a presence of an object having an object class in the image. A classification MLA detects the object having the object class in each image. A class activation map (CAM) indicative of discriminative regions used by the classification MLA for detecting the object in each image is generated. A region proposal MLA is used to generate region proposals for each image. A pseudo mask of the respective object is generated based on the region proposals and the CAM, where a pseudo mask is indicative of pixels corresponding to the respective object class. The pseudo masks are used as a label with the image-level labelled images for training the instance segmentation MLA.

Abstract: A method and a system for joint data augmentation and classification learning, where an augmentation network learns to perform transformations and a classification network is trained. A set of labelled images is received. During an inner loop iteration, an augmentation network applies a transformation on a given labelled image of the set to obtain a transformed image. The classification network classifies the transformed image to obtain a predicted class, and a training loss is determined based on the predicted class and the respective label. The parameters of the classification network are updated based on the classification loss. During an outer loop iteration, the classification network classifies another labelled image of the set to obtain another predicted class, and a validation loss is determined based on the other predicted class and the respective label. The parameters of the augmentation network are updated based on the validation loss.

Abstract: A method and a system for generating training images for training an instance segmentation machine learning algorithm (MLA). A set of image-level labelled images are received, where a given image is labelled with a label indicative of a presence of an object having an object class in the image. A classification MLA detects the object having the object class in each image. A class activation map (CAM) indicative of discriminative regions used by the classification MLA for detecting the object in each image is generated. A region proposal MLA is used to generate region proposals for each image. A pseudo mask of the respective object is generated based on the region proposals and the CAM, where a pseudo mask is indicative of pixels corresponding to the respective object class. The pseudo masks are used as a label with the image-level labelled images for training the instance segmentation MLA.

Abstract: A method for strengthening metal structure, including steps: i) mixing two-component epoxy resin composition; ii) applying composition to metal structure surface, or onto strengthening element, or both; iii) contacting applied epoxy resin composition with strengthening element surface and/or metal structure surface wherein composition forms interlayer between metal structure and strengthening element, and iv) curing epoxy resin composition at 100° C. or below; wherein two-component epoxy resin composition contains: first component between 10-50 wt.-% of at least one epoxy resin contains on average more than one epoxy group per molecule; second component between 5-30 wt.-% of hardener for epoxy resins; between 3-25 wt.-% of at least one impact strength modifier in one or both components; between 15-80 wt.-% of at least one filler in one or both components; and wherein two-component epoxy resin composition exhibits after curing Elastic Modulus at least 2500 MPa, and Impact Peel Strength at least 15 N/mm.

Abstract: Disclosed herein a system, a method and a device for generating a voice model for a user. A device can include an encoder and a decoder to generate a voice model for converting text to an audio output that resembles a voice of the person sending respective text. The encoder can includes a neural network and can receive a plurality of audio samples from a user. The encoder can generate a sequence of values and provide the sequence of values to the decoder. The decoder can establish, using the sequence of values and one or more speaker embeddings of the user, a voice model corresponding to the plurality of audio samples of the user.

Abstract: A compatibility score generator implementing a neural network is trained for assessing compatibility of items. Elements of a feature vector representing each item and of a compatibility data structure indicating items considered compatible are retrieved. The neural network is trained using training data corresponding to the items and indicating compatibility between pairs of items. The compatibility data structure is modified by removing indications that items of a pair of items are compatible. An encoding function generating encoded representations for the items based on the compatibility data structure is evaluated. Encoded representations are provided to a decoder that learns a likelihood that the indication had been removed when modified. The neural network and the decoder are optimized based on a loss function that reflects the decoder's ability to correctly determine whether the indication had been removed. The encoded representations generate a compatibility score for at least two items of interest.

Abstract: A system for controlling the operation of utilities such as power, water, gas and communications in a building. The system includes a software component that is uploaded onto a mobile communications device such as a cell phone, smart phone, tablet etc. that communicates with at least one utility control devices that enable and disable the utility service to the building based upon the mobile communications device being in proximity to the building.

Abstract: A radiographic apparatus having a laser positioning system comprising three independent laser units for positioning a patient in three-dimension for the purpose of taking x-rays. Each laser unit is orthogonally oriented in respect to the others, such that the first laser unit projects a laser beam aligned with the x-ray beam in said first lateral direction, and the second laser unit projects a laser beam in a second lateral direction that is orthogonal to the first lateral direction, and the third laser unit projects a laser beam in a vertical direction that is orthogonal to both the first and second lateral directions. With the patient properly oriented by the laser system, x-ray films are taken that can be sequenced for radiographic animation analysis.

Abstract: A sealing apparatus is disclosed. In some embodiments, the sealing apparatus includes a shaft assembly that couples a laterally expandable plug and a longitudinally expandable member. The plug and the expandable member are captured within the ends of the shaft assembly. The shaft assembly is configures so that longitudinal expansion of the expandable member results in lateral expansion of the plug. Thus, such a device can be inserted into a pipe and the plug can be expanded within the pipe to provide a seal about the shaft assembly. In some embodiments, the shaft assembly can define a longitudinally extending hollow pathway that allows flow between the two ends of the shaft assembly. Examples of how such hollow pathway and the plug operation can be implemented are disclosed. Example applications, such as pressure measurements, are also disclosed.