Abstract: Disclosed are systems and methods to incrementally train neural networks. Incrementally training the neural networks can include defining a probability distribution of labeled training examples from a training sample pool, generating a first training set based off the probability distribution, training the neural network with the first training set, adding at least one additional training sample to the training sample pool, generating a second training set, and training the neural network with the second training set. The incremental training can be recursive for additional training sets until a decision to end the recursion is made.

Abstract: Disclosed are systems and methods to incrementally train neural networks. Incrementally training the neural networks can include defining a probability distribution of labeled training examples from a training sample pool, generating a first training set based off the probability distribution, training the neural network with the first training set, adding at least one additional training sample to the training sample pool, generating a second training set, and training the neural network with the second training set. The incremental training can be recursive for additional training sets until a decision to end the recursion is made.

Abstract: A system comprises an input interface to receive input indicating a question, a communication module to establish a communication link with an access network, wherein the communication link provides connectivity to one or more packet data networks (PDNs) via the access network, and a computer coupled to the input interface and the communication module, the computer including a processor and a memory, the memory storing instructions executable by the processor to execute an information retrieval procedure including accessing an open-domain context search space of the one or more PDNs and retrieving, from among a plurality of contexts of the open-domain context search space, a plurality of candidate contexts for answering the question using a question-answering model for open-domain question answering, identify a set of non-answering contexts among the plurality of candidate contexts, wherein each of the set of non-answering contexts is a respective context for which the question-answering model predicts the qu

Abstract: A system comprises an input interface to receive input indicating a question, a communication module to establish a communication link with an access network, wherein the communication link provides connectivity to one or more packet data networks (PDNs) via the access network, and a computer coupled to the input interface and the communication module, the computer including a processor and a memory, the memory storing instructions executable by the processor to execute an information retrieval procedure including accessing an open-domain context search space of the one or more PDNs and retrieving, from among a plurality of contexts of the open-domain context search space, a plurality of candidate contexts for answering the question using a question-answering model for open-domain question answering, identify a set of non-answering contexts among the plurality of candidate contexts, wherein each of the set of non-answering contexts is a respective context for which the question-answering model predicts the qu

Abstract: Disclosed are systems and methods to incrementally train neural networks. Incrementally training the neural networks can include defining a probability distribution of labeled training examples from a training sample pool, generating a first training set based off the probability distribution, training the neural network with the first training set, adding at least one additional training sample to the training sample pool, generating a second training set, and training the neural network with the second training set. The incremental training can be recursive for additional training sets until a decision to end the recursion is made.