Abstract: Techniques for leveraging existing statistical prediction models are provided. A first statistical prediction model is generated for a content item. An instruction is received to create a clone from the content item. In response to receiving the instruction, the clone is created based on attributes of the content item. A second statistical prediction model that is different than the first statistical prediction model is generated for the clone. In response to receiving a request for content, the clone is identified as relevant to the first request. A similarity between (1) first content of the content item and (2) second content of the clone is determined. If the similarity exceeds a similarity threshold, then the first statistical prediction model is used to generate a prediction of an entity user selection rate associated with the clone. Otherwise, the second statistical prediction model is used to generate the prediction.

Abstract: Predicting beneficial drug combinations mitigating adverse drug reactions identifies drug combinations and associated target adverse drug reaction from a spontaneous reporting system containing case reports of drugs and associated adverse drug reactions. Each drug combination comprises a first drug and a second drug, and a propensity score is computed for each drug in each group. This propensity score expresses a probability of being exposed to a given drug based on other co-prescribed drugs and reported indications, which reflect patient characteristics. Associations are computed for each drug as well as drug interaction. Among the associations, the sum of the associations of the second drug and the interaction effect represents the predicted beneficial score expressing whether the second drug alters the chance of developing the target adverse drug reaction for patients on the first drug.

Abstract: A computer-implemented method comprising: obtaining text from text-based messages sent between a patient and a therapist providing psychological therapy; determining at least one feature of the text; and determining a characteristic of the patient and/or the therapist using the at least one feature.

Abstract: Mechanisms are provided for obfuscating training of trained cognitive model logic. The mechanisms receive input data for classification into one or more classes in a plurality of predefined classes as part of a cognitive operation of the cognitive system. The input data is processed by applying a trained cognitive model to the input data to generate an output vector having values for each of the plurality of predefined classes. A perturbation insertion engine modifies the output vector by inserting a perturbation in a function associated with generating the output vector, to thereby generate a modified output vector. The modified output vector is then output. The perturbation modifies the one or more values to obfuscate the trained configuration of the trained cognitive model logic while maintaining accuracy of classification of the input data.

Abstract: Various embodiments are generally directed to techniques for embedding a data object into a multidimensional frame, such as for training an autoencoder to generate latent space representations of the data object based on the multidimensional frame, for instance. Additionally, in one or more embodiments latent space representations of data objects may be classified, such as with a machine learning algorithm. Some embodiments are particularly directed to embedding a data object comprising a plurality of object entries into a three-dimensional (3D) frame.

Abstract: Disclosed are systems and methods for autonomous computing replacing or augmenting a human user of computer programs, where access to internal operations of the computer program is not used. An application controller can use the display output of a computer program to determine a current state of the computer program, using the disclosed embodiments. For example, identity of menu options of the computer program can be determined from image frames obtained from the display output of the computer program and used to determine the current state of the computer program. The application controller can provide input commands to the computer program to execute the computer program from the current state to a destination state.

Abstract: A method is provided for rule creation that includes receiving (i) a MDP model with a set of states, a set of actions, and a set of transition probabilities, (ii) a policy that corresponds to rules for a rule engine, and (iii) a set of candidate states that can be added to the set of states. The method includes transforming the MDP model to include a reward function using an inverse reinforcement learning process on the MDP model and on the policy. The method includes finding a state from the candidate states, and generating a refined MDP model with the reward function by updating the transition probabilities related to the state. The method includes obtaining an optimal policy for the refined MDP model with the reward function, based on the reward policy, the state, and the updated probabilities. The method includes updating the rule engine based on the optimal policy.

Abstract: Method and system for performing semi-supervised regression with a generative adversarial network (GAN) that includes a generator comprising a first neural network and a discriminator comprising a second neural network, comprising: outputting, from the first neural network, generated samples derived from a random noise vector; inputting, to the second neural network, the generated samples, a plurality of labelled training samples, and a plurality of unlabelled training samples; and outputting, from the second neural network, a predicted continuous label for each of a plurality of the generated samples and unlabelled samples.

Abstract: The technology relates to predicting that an object is going to enter into a trajectory of a vehicle. This may include receiving sensor data identifying a first location of the object in an environment of the vehicle at a first point in time and receiving sensor data identifying a second location of the object in the environment at a second point in time. In addition, a boundary of the trajectory is determined by defining at least a two-dimensional area through which the vehicle is expected to travel in the future. A first distance between the boundary and the first location and a second distance between the trajectory and the second location are determined. The first distance and the second distance are used to determine that the object is going to enter into the trajectory at a future point in time.

Abstract: A computer-implemented method for training a front-end neural network (“front-end NN”) and a back-end neural network (“back-end NN”) is provided. The method includes combining the back-end neural network with the front-end neural network to form a joint layer to thereby generate a combined neural network. The method also includes training the combined neural network for a speech recognition with a set of utterances as training data, with the joint layer having a plurality of frames and each frame having a plurality of bins, and where one or more specific units in each frame are dropped during the training, each of the specific units being selected randomly or based on a bin number to which the respective unit is set within its frame, with the specific units corresponding to one or more common frequency bands.

Abstract: An information processing device includes: a processor that executes a process, the process including: controlling a recognition process that performs, with respect to input neuron data, a hierarchical neural network operation including a weighting operation using a parameter and that holds the neuron data and the parameter of each layer of the neural network in each of memory areas; and performing, in a learning process of learning the parameter of each layer of the neural network from an error that is obtained from a recognition result, regarding the layer in which the neuron data and the parameter are held in the memory areas, control of calculating an error of the neuron data after calculating an error of the parameter.

Abstract: A method for secure learning of parameters of a convolution neural network, CNN, for data classification includes the implementation, by data processing of a first server, including receiving from a second server a base of already classified learning data, the learning data being homomorphically encrypted; learning in the encrypted domain, from the learning database, the parameters of a reference CNN including a non-linear layer (POLYNOMIAL) operating an at least two-degree polynomial function approximating an activation function; a batch normalization layer before each non-linear layer (POLYNOMIAL); and transmitting the learnt parameters to the second server, for decryption and use for classification.

Abstract: A system for reconciliation comprises a determination engine to determine whether data is structured or unstructured, a data structuring engine to structure the data, and a rule extraction engine to determine relations between pairs of values of a first set and a second set of data. The system further comprises a matching engine to generate a confidence score for each pair of the values, a categorization engine to classify the pairs of values into matched pairs and unmatched pairs, a validation engine to validate matching and classification of the pairs based on a user feedback, and a learning engine to store details pertaining to the validation of the matching and the classification over a period of time. The learning engine forwards the details to the rule extraction engine and the categorization engine to determine the relations between subsequent pairs of values and classify the pairs based on the stored details.

Abstract: Methods, apparatuses, and computer readable mediums for exercise behavior prediction are provided. In a particular embodiment, the prediction evaluation controller is configured to generate an exercise activity pattern based on correlations between scheduling of a user's historical non-exercise events and the user's historical exercise events. In the particular embodiment, the prediction evaluation controller is also configured to generate, based on the generated exercise activity pattern, by the prediction evaluation controller, a future exercise event to correspond with a future non-exercise event scheduled on the user's calendar. In the particular embodiment, the prediction evaluation controller is also configured to provide an indication of the generated future exercise event.

Abstract: A system, method and program product for stratigraphic layer identification using a stratigraphic knowledge base for machine learning. Reservoir data includes seismic data and well log data for a reservoir area. The well log data is processed to identify well stratigraphic layer features and the seismic data is processed to identify seismic stratigraphic layer features. A feature matching algorithm based on a stratigraphic knowledge base is selected to match the well stratigraphic layer features to the seismic stratigraphic layer features. The matched features are used to define a stratigraphic layer interpretation for the reservoir area and the interpretation is presented to a user.

Abstract: A computer-based method includes receiving an input signal at a neuron in a computer-based neural network that includes a plurality of neuron layers, applying a first non-linear transform to the input signal at the neuron to produce a plain signal, and calculating a weighted sum of a first component of the input signal and the plain signal at the neuron. In a typical implementation, the first non-linear transform is a function of the first component of the input signal and at least a second component of the input signal.

Abstract: A facility for processing output from a network of mechanical sensors is described. The facility accesses time-series data outputted by the network of sensors. The facility applies to the accessed time-series data a trained autoencoder to obtain a version of the accessed time-series data in which noise present in the accessed time-series data is at least partially suppressed. The facility stores the obtained version of the accessed time-series data, such as in order to perform human activity recognition against the obtained version of the accessed time-series data.

Abstract: Methods, systems, and devices, including computer programs encoded on computer storage media for transferring a robot customer service to a human customer service are provided. One of the methods includes: obtaining conversation characteristics from at least one round of conversations between the robot customer service and a customer; obtaining state characteristics of the customer; inputting the conversation characteristics and the state characteristics into a confidence score evaluation model to obtain a confidence score evaluation value; and when the confidence score evaluation value meets a robot-to-human intervention condition, transferring the customer to the human customer service. The confidence score evaluation model is a machine learning model, comprising a linear sub-model input with the conversation characteristics and a deep neural network sub-model input with the state characteristics.

Abstract: An acceleration and compression method for a deep convolutional neural network based on quantization of a parameter provided by the present application comprises: quantizing the parameter of the deep convolutional neural network to obtain a plurality of subcode books and respective corresponding index values of the plurality of subcode books; acquiring an output feature map of the deep convolutional neural network according to the plurality of subcode books and respective corresponding index values of the plurality of subcode books. The present application may implement the acceleration and compression for a deep convolutional neural network.

Abstract: Disclosed herein is a heuristically programmable system comprising a web-socket; a backend component that is in operative communication with the web-socket to act as a heuristically programmed algorithm framework that is operative to interact with a user in natural language and in a human-like manner; a main frame computer; a web server that is in operative communication with an internet; where the main frame computer and the web server are in operative communication with the web socket; and a cognitive computing system; where the web socket comprises a computer system configured to provide a voice, motion and/or a graphical interface to a user; and where the web socket is operative to communicate with the cognitive computing system; where the cognitive computing system is operative to convert speech to text and to send this text to the web socket.