- The models that use  these algorithms starts by making a parse tree for each input sentence ,which is traversed to identify all the predicates.

- For each predicate the SRL iterates  through each node in the parse tree and perform supervised classification to determine whether the node serves as a semantic  role for the specified predicate and what type of role.

Feature-based algorithm for Semantic Role Labelling

Artificial neural network models use bi-directional LSTM which performs BIO tagging.

 Neural Algorithm for Semantic Role Labeling

Precision, recall and  F1 score is used to evaluate whether each word constituent has been assigned the correct argument label.

Evaluation of Semantic Role Labeling

- Sematic Role Labeling(SRL) is the process of finding the semantic role of each argument of each predicate in a sentence.

- Supervised machine learning approaches are used to achieve SRL systems.

- FrameNet and PropBank supplies the train and test datasets to develop SRL models.

San Jose State University

Semantic roles focus about what a word actually means and is actually doing in the semantic context they do not care about the structure of the sentence.

Semantic Roles ( Thematic Roles )

An on-going but a helpful book resource about NLP
https://web.stanford.edu/~jurafsky/slp3/

Speech and Language Processing (3rd ed. draft) 

Commonly Used Semantic Roles

It has proved quite difficult to come up with a standard set of roles, and equally difficult to produce a formal definition of roles like AGENT, THEME, or INSTRUMENT. Possible solutions are:
- Define generalized semantic roles that abstract over the specific thematic roles.
- Define semantic roles that are specific to a particular verb or a particular group of semantically related verbs or nouns.

Problems with Thematic Roles

Semantic Role Labeling

Primitive Decomposition (Componential Analysis) is the idea of decomposing meaning into sets of primitive semantics elements or features. 

For example, we could represent the meaning of "kill" as:
- KILL(x, y)
- CAUSE(x, BECOME(NOT(ALIVE(y))))

Learn Before

Related

Learn After