- This is an example of Named Entity Recognition(NER) problem. One can build a sequence model such as an LSTM to perform this task. However, as shown in both the sentences above, forward only LSTM might fail here.
- Using forward only direction LSTM might result in a model which recognises Teddy as a product : “bear”, which is on sale. This is because it does not look at context after the occurrence of the word, as in Teddy Roosvelt.
- Using BiDirectional RNNs can resolve the above problem. As context of “President” in w2 can be transferred back to the word “Teddy” which makes it easier to call it a person and not a product.
Given the following two sentences, how do you determine if Teddy is a person or not? “Teddy bears are on sale!” and “Teddy Roosevelt was a great President!”
Posted on