How much is rate of recall of linguistic model of rich full small artificial intelligence

Recall leads be in the light of original sample book character, its meaning is to be in it is actually be forecasted to be in example the probability of example. To the linguistic model, recall rate can measure a model to be able to identify correctly when handling all sorts of language tasks related to recall the ability of information. If the recall rate of a linguistic model is high, mean it to be able to find wants information well and truly, avoid to omit serious content. The recall rate that raises a linguistic model can pass a variety of means to come true.

For example, can add the training data bulk of the model, let a model learn more language mode and semantic information, raise its to be opposite thereby the identifying of different type text ability. In addition, return the algorithm that can optimize a model and parameter, make its can take the crucial news in text better, raise the accuracy of recall. However, to artificial intelligence linguistic model of Fu Manwei, because lack specific data, we cannot determine the particular numerical value that its recall leads.

Recall is led (Recall) also call really positive rate, hit the target rate, its mirrorred classification implement or the model is forecasted correctly the ability that example spends completely, namely example is forecasted to be example is occupied total the scale of example. The computational formula that recall leads is: Recall leads = Tp / (Tp + Fn) , among them Tp expresses to be actually example and be forecasted to be the example number of example, fn expresses to be actually example but the amount that is calculated to be negative example.

The recall rate of the linguistic model can get the influence of a variety of elements, for example setting of the framework of the model, quality that trains data and amount, application and the union with other technology.

If you are opposite,the recall rate of linguistic model of rich full small artificial intelligence is interested, documentation of the pertinent information that the proposal consults the company offers Fu Manwei, product or technology report, in order to get more detailed with accurate assessment result. The professional evaluation that perhaps pays close attention to this domain and user feedback, but the recall that needs to notice different application setting falls is led may somewhat difference. In the meantime, when evaluating a linguistic model, cannot depend on recall to lead this one index only, still need to consider other index integratedly, wait like value of accuracy rate, exact ratio, F1, with understanding the model expression in different respect in the round.