Language detection of user input text for online gaming

A method of training a machine learning model to identify candidate languages of a user query for online games, wherein the method comprises:

• receiving text data comprising first text data, associated language labels for the first text data and second text data unlabeled for languages;
• determining associated language labels for the second text data, based on a threshold number of a plurality of non-specific language detectors identifying at least one language;
• creating training text data by separating the text data into n-grams;
• and
• training the machine learning model to generate at least one candidate language identifier of the user query and a respective language confidence score, wherein the training comprises:
• training with the training text data without the associated language labels to obtain current predicted language labels by the machine learning model;
• and
• retraining the machine learning model with discrepancy information between the current predicted language labels and the associated language labels, to update the current predicted language labels.

A system comprising a computing device for detecting language of a user query for online games, wherein the computing device comprises:

• one or more processors;
• and
• logic encoded in one or more tangible media for execution by the one or more processors and when executed operable to perform operations comprising:
• receiving the user query;
• separating the user query into a plurality of n-grams;
• identifying at least one candidate language of the user query, by applying a machine learning model to the plurality of n-grams, wherein the machine learning model is trained with at least one multilingual text corpus and game-related data,
• determining, by applying the machine learning model, a respective confidence score for each of the at least one candidate language;
• finding one or more response matches to the user quay in language-specific game databases associated with each of the at least one candidate language that has a confidence score meeting a confidence threshold;
• and
• determining a respective match score for each of the one or mom response matches;
• and providing search results including game information associated with particular response matches, based, at least in part, on respective weighted scores of at least the respective confidence score and the respective match score.