I know that NLP algorithms can be trained to recognise the topic of a document or part of a document. I would like to train an algorithm to recognise certain abstract concepts. For example I want to extract sentences or paragraphs referring to a person or a company being good at something or bad at something, I.e. "she is good at", "he is knowledgeable in", "they have superior skills in" or that someone is missing something like " there is a lack of ", "they don't have", etc. Is it possible to train an nlp algorithm for this? I have hundreds of labelled sample sentences for each of these abstract categories that I am interested in. Are there any tutorials or research papers that cover this?