Scribd
Upload
6 views

4 th unit notes NLP

The document discusses predicate-argument structure and semantic role labeling (SRL) in linguistics, focusing on how to identify participants and their roles in events represented by predicates. It also covers various meaning representation systems, including Abstract Meaning Representation (AMR), Semantic Networks, Conceptual Dependency Theory, Frame Semantics, and Distributional Semantics, highlighting their approaches to capturing semantic relationships. The goal of these systems is to enable natural language processing tasks to better understand and represent the meaning of sentences.

Uploaded by

frag3676
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

4 th unit notes NLP

The document discusses predicate-argument structure and semantic role labeling (SRL) in linguistics, focusing on how to identify participants and their roles in events represented by predicates. It also covers various meaning representation systems, including Abstract Meaning Representation (AMR), Semantic Networks, Conceptual Dependency Theory, Frame Semantics, and Distributional Semantics, highlighting their approaches to capturing semantic relationships. The goal of these systems is to enable natural language processing tasks to better understand and represent the meaning of sentences.

Uploaded by

frag3676
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

IV UNIT :Predicates Predicate-Argument Structure :

Predict- argument Structure :


A thing that represent a type of event (or) State
 The process of identifying various arguments of predicates in a sentence
we will call semantic labelling
 Identify the participants of the entity in the event
 Resolving the arguments structure of predicate in a sentence is where we
identify which entity .what part of which event
 Semantic representation that illsulate who did ,what to,whom ,when,
why and how ,common structure of semantic labelling

Example : Bell Atlantic corp said it will acquire one of control data corps
computer-maintenance business
It will represent of who did what to whom ,when, where, why, and how

In this above example : said and acquire are predicates


1 Shallow semantics parsing or semantic role labelling, is the process of identifying the
various arguments of predicates in a sentence.

2 In linguistics, predicate refers to the main verb in the sentence. Predicate takes arguments.

3 The role of Semantic Role Labelling (SRL) is to determine how these arguments are
semantically related to the predicate.

4 Consider the sentence "Mary loaded the truck with hay at the depot on Friday".

5 'Loaded' is the predicate. Mary, truck and hay have respective semantic roles of loader,
bearer and cargo. Mary->loader, truck->bearer hay->cargo

6 We can identify additional roles of location (depot) and time (Friday).

7. The job of SRL is to identify these roles so that NLP tasks can "understand" the sentence.

8 Often an idea can be expressed in multiple ways. Consider these sentences that all mean
the same thing:

 "Yesterday, Kristina hit Scott with a baseball";


 "Scott was hit by Kristina yesterday with a baseball";
 "With a baseball, Kristina hit Scott yesterday";
 "Kristina hit Scott with a baseball yesterday".

9 Either constituent or dependency parsing will analyze these sentence syntactically.

10. But syntactic relations don't necessarily help in determining semantic roles.

2 Topic : Meaning Representation Systems

Abstract Meaning Representation (AMR): AMR aims to create a


single, consistent semantic representation for English sentences,
focusing on "who did what to whom, when, where, and why." It
represents the meaning as a rooted, labeled, directed graph where nodes
represent concepts and edges represent semantic relations.

Semantic Networks: These use graphs to represent concepts and the


semantic relations between them. Nodes in the graph represent
concepts, and edges represent relationships like "is-a" (e.g., "cat is-a
mammal"), "part-of" (e.g., "wheel part-of car"), or more specific
relations defined by the system. WordNet is a well-known example of
a lexical semantic network.

Conceptual Dependency (CD) Theory: This theory, developed by


Roger Schank, aims to represent all actions and states in terms of a
small set of primitive conceptual categories and dependencies. For
instance, "John gave Mary a book" might be represented using
primitives like ATRANS (abstract transfer) with specifications for the
actor, object, recipient, and the direction of transfer.
Frame Semantics: This approach, associated with Charles Fillmore,
describes the meaning of words and sentences in terms of "frames" –
schematic representations of situations or concepts. A frame for
"buying," for example, would include roles like Buyer, Seller, Goods,
and Money. Understanding a word within a sentence involves
activating the corresponding frame and filling in the roles.

Distributional Semantics (Word Embeddings): In more recent


years, with the rise of deep learning, distributional semantics has
become incredibly influential. This approach represents the meaning of
words (and sometimes larger units of text) as dense vectors in a high-
dimensional space. The idea is that words that appear in similar
contexts tend to have similar meanings, and this similarity is captured
by the proximity of their vectors. Word2Vec, GloVe, and more
advanced transformer-based embeddings like BERT and RoBERTa fall
under this category.

You might also like