Abstract: An approach for generating synthetic treebanks to be used in training a parser in a production system is provided. A processor receives a request to generate one or more synthetic treebanks from a production system, wherein the request indicates a language for the one or more synthetic treebanks. A processor retrieves at least one corpus of text in which the requested language is present. A processor provides the at least one corpus to a transformer enhanced parser neural network model. A processor generates at least one synthetic treebank associated with a string of text from the at least one corpus of text in which the requested language is present. A processor sends the at least one synthetic treebank to the production system, wherein the production system trains a parser utilized by the production system with the at least one synthetic treebank.

Abstract: One or more computer processors responsive to neural network run-time, reduce one or more sets of maximum activations along a hidden dimension respectively associated with one or more activation tensors and one or more layers of a neural network. The one or more computer processors compute an interquartile range (IQR) clip threshold for each reduced set for each sequence dimension in the neural network. The one or more computer processors clip one or more activations based on respective computed IQR clip thresholds. The one or more computer processors quantize the clipped activations.

Abstract: An approach for generating synthetic treebanks to be used in training a parser in a production system is provided. A processor receives a request to generate one or more synthetic treebanks from a production system, wherein the request indicates a language for the one or more synthetic treebanks. A processor retrieves at least one corpus of text in which the requested language is present. A processor provides the at least one corpus to a transformer enhanced parser neural network model. A processor generates at least one synthetic treebank associated with a string of text from the at least one corpus of text in which the requested language is present. A processor sends the at least one synthetic treebank to the production system, wherein the production system trains a parser utilized by the production system with the at least one synthetic treebank.

Abstract: A method of using a computing device to improve an answer generated by a natural language question and answer system includes receiving, by a computing device, multiple questions in a natural language question and answer system. The computing device further generates multiple answers to the multiple questions. The computing device still further constructs a new training set with the generated multiple answers, where each answer is compared with a corresponding question of the multiple questions. The computing device additionally augments the new training set with one or more tokens delimiting a span of one or more of the generated multiple answers. The computing device further trains a new natural language question and answer system with the augmented new training set.

Abstract: Techniques for translating and presenting text in a structured document are provided. A structured document including text to be translated is received. At least one portion of the text to be translated is translated into at least one representation of translated text of a specific language. The translation is carried out in accordance with a prioritization schedule. Further, the representation of translated text is provided by at least first and second servers, the first server independently providing the representation of translated text of a first quality and the second server independently providing the representation of translated text of a second quality, the first quality being different than the second quality. As the text is translated, the representation of translated text is presented immediately when available.

Abstract: An improved method, apparatus, and computer instructions displaying search results. A query is received in a first language. A snippet in the document containing the word is selected in response to identifying a document in a second language in which the document contains a word matching a query word in the query. This snippet is translated into the first language.

Abstract: Techniques for translating and presenting text in a structured document are provided. A structured document including text to be translated is received. At least one portion of the text to be translated is translated into at least one representation of translated text of a specific language. The translation is carried out in accordance with a prioritization schedule. Further, the representation of translated text is provided by at least first and second servers, the first server independently providing the representation of translated text of a first quality and the second server independently providing the representation of translated text of a second quality, the first quality being different than the second quality. As the text is translated, the representation of translated text is presented immediately when available.

Abstract: A parser is provided that parses a Chinese text stream at the character level and builds a syntactic structure of Chinese character sequences. A character-based syntactic parse tree contains word boundaries, part-of-speech tags, and phrasal structure information. Syntactic knowledge constrains the system when it determines word boundaries. A deterministic procedure is used to convert word-based parse trees into character-based trees. Character-level tags are derived from word-level part-of-speech tags and word-boundary information is encoded with a positional tag. Word-level parts-of-speech become a constituent label in character-based trees. A maximum entropy parser is then built and tested.

Abstract: A system for translating a first word set in a source language into a second word set in a target language, the system comprising: input means for inputting the first word set into the system; tagging means for tagging the first word set input to the system so as to at least substantially reduce non-essential variability in the first word set; translation means including a single a posteriori conditional probability model and a target candidate store for storing target language candidate word sets, wherein the translation means employs the single model to evaluate the target language candidate word sets in order to select the target language candidate word set having a best score with respect to the first word set; and output means for outputting the best scoring target language candidate word set as the second word set in the target language.

Abstract: The invention proposes using statistical methods to do natural language understanding. The key notion is that there are "strings" of words in the natural language, that correspond to a single semantic concept. One can then define an alignment between an entire semantic meaning (consisting of a set of semantic concepts), and the English. This is modeled using P(E,A.vertline.S). One can model p(S) separately. This allows each parameter to be modeled using many different statistical models.