Abstract
News editors obtained the following quote from the background information suppli ed by the inventors: “Machine-learning models, such as neural networks, are trai ned based on training data. Training data that is used for a machine-learning mo del is generally in the same domain as the problem that the machinelearning mod el is solving. For example, a machine-learning model that generates search resul ts generally must be trained based on search logs that describe how users have i nteracted with search results in the past. However, when a model is used to gene rate output for a new vertical of an online system, the model generally struggle s to generate optimal output because of a lack of training data. For example, a machinelearning model that is trained to provide content suggestions alongside search results that are responsive to a search query may have less training data than the model that provides search results because users may be more inclined to interact with search results than content suggestions. Thus, a content sugges tion model may be less well trained than a machine-learning model presenting sea rch results.
NETL
NSTL