International Business Machines Corporation

Rodrigo Goulart Silva

Shikhar Kwatra

Indervir Singh Banipal

Craig M Trim

Vijay Ekambaram

Zachary A Silverstein

Andrew Ronald Freed

Hemant Kumar Sivaswamy

Gary Francis Diamanti

Ryan R Anderson

A computer-implemented process for training a natural language processing (NLP) agent having a reinforced learning model includes the following operations. A type of document from a document corpus is identified using metadata particularly associated with the document. The NLP agent tokenizes the document to generate a plurality of tokens. Using a schema identified from the type of the document, one of the plurality of tokens is compared to a system of record (SOR) field from the schema. A similarity score between the one of the plurality of tokens with a correct value and a reward based upon the similarity score are generated. A determination is made that an optimum minimum average similarity rate has not been obtained. Based upon the determination, the reinforced learning model is trained using a loss function that includes the reward.

Machine learning for training NLP agent

Mitigating mistranscriptions resolves errors in a transcription of the audio portion of a video based on a semantic matching with contextualized data electronically garnered from one or more sources other than the audio portion of the video. A mistranscription is identified using a pretrained word embedding model that maps words to an embedding space derived from the contextualizing data. A similarity value for each vocabulary word of a multi-word vocabulary of the pretrained word embedding model is determined in relation to the mistranscription. Candidate words are selected based on the similarity values, each indicating a closeness of a corresponding vocabulary word to the mistranscription. The textual rendering is modified by replacing the mistranscription with a candidate word that, based on average semantic similarity values, is more similar to the mistranscription than is each other candidate word.

Audio mistranscription mitigation

A request is received from a user to search a document for at least one search term. The document is analyzed to determine contextual meanings associated with the at least one search term. An ordered set of contexts of the at least one search term is constructed. An option of disambiguation is then prompted to the user to select a context of the ordered set of contexts. Search results are presented to the user associated with the selected context of the at least one search term.

Context-based text searching

A dataset comprising text-based messages can be accessed. Tokens for words and punctuation marks contained in the text-based messages can be generated. Each token corresponds to one word or one punctuation mark. A vector representation for each of a plurality of the tokens can be generated using natural language processing. A sequence of tokens corresponding to the text-based message can be generated for each of a plurality of the text-based messages in the dataset. Ones of the tokens that represent punctuation marks can be identified. An artificial neural network can be trained to predict use of the punctuation marks in sentence structures. The training uses the generated sequence of tokens and the vector representations for the tokens, in the sequence of tokens, that represent the punctuation marks.

Text autocomplete using punctuation marks

Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: obtaining during an interactive user support session user transmitted data including conversation content from a client computing device associated to the user; subjecting the user transmitted data to processing to provide feedback data associated to the conversation content obtained from the client computer device, wherein the feedback data indicates a level of satisfaction of the user, and wherein the processing includes subjecting content of the conversation content to natural language processing to return sentiment data of the user; examining data of the feedback data; performing selecting of a conversation strategy in dependence on the examining; and sending for presentment to the user returned conversation content in accordance with the certain conversation strategy.

Growing community of inventors

Raleigh, NC, United States of America

Rodrigo Goulart Silva