Abstract: A neural transcompilation model is tested with a set of syntax unit tests to determine the syntax elements of a source code program written in a source programming language that fail to translate properly into a target programming language. The syntax elements having a translation defect is identified and ranked according a translation failure rate. The neural transcompilation model is then fine-tuned with training samples of the syntax elements having the highest translation failure rates and their paired correct translation in order to teach the model to learn the association between the syntax elements of a source programming language causing translation defects and its correct translation in a target programming language.

Abstract: Methods for determining a refined title to present in a search results page for a product are described. Components of a server system may receive a set of input titles for a set of listings associated with a product. Components of the server system may receive a listing request including a suggested title for a first listing for the product. The components of the server system may generate a refined title for the first listing based on the set of input titles and the suggested title. The components of the server system may then receive, from a user device, a search query that may be mapped to the product, and the component of the server system may transmit, to the user device, a query response that includes the refined title for the first listing based on the search query.

Abstract: A first variant of a listing and a second variant of the listing are received. The listing describes an item for sale in an electronic marketplace. The first variant describes a different iteration of the item relative to the second variant. First user activity data of the first variant is generated and second user activity data of the second variant is generated. The first user activity data corresponds to user input metrics associated with the first variant. The second user activity data corresponds to user input metrics associated with the second variant. A search engine receives a first query. Based at least in part on the first user activity data relative to the second user activity data, a first search result associated with the first variant is ranked higher than a second search result associated with the second variant.