Index

1. Introduction
2. Syntactic Processing
3. Semantic Analysis
4. Discourse analysis and Pragmatic processing
5. Learning in Natural Language Processing
6. Inductive Learning in NLP
7. Learning Decision Trees
8. Explanation-Based Learning in NLP
9. Learning Using Relevance Information
10. Neural Network Learning in NLP
11. Genetic Learning in NLP
12. Representing and Using Domain Knowledge
13. Expert System Shells
14. Knowledge Acquisition

Introduction

Natural Language Processing (NLP) is a branch of artificial intelligence that focuses on enabling computers to understand, interpret, and generate human language. It bridges the gap between human communication and computer understanding.

Let's dive deeper into the introductory aspects of NLP.

I didn't spend much effort on this watered down version as there is not much to understand based on the syllabus and lack of content in general. But I will soon work on a completely detailed and in-depth material on NLP for learning AI in general, once this exam is done.

1. Definition and Scope of NLP:

• NLP is the technology that allows computers to understand, analyze, manipulate, and potentially generate human language
• It combines computer science, linguistics, and artificial intelligence
• Goals range from simple tasks (like spell checking) to complex ones (like understanding sarcasm or emotion in text)

2. Challenges in Processing Natural Language:

a) Ambiguity:

• Lexical ambiguity (words with multiple meanings)
  Example: "bank" can mean financial institution or river bank

• Structural ambiguity (sentences with multiple interpretations)
  Example: "I saw a man with a telescope" (Who has the telescope?)

b) Context-dependence:

• Same words can mean different things in different contexts
• Understanding references (pronouns, definite articles)
• Cultural and situational context matters

c) Linguistic Variation:

• Different dialects and regional variations
• Informal language, slang, and neologisms
• Non-standard grammar and spelling

3. Applications of NLP:

• Machine Translation (Google Translate)
• Information Extraction
• Text Summarization
• Question-Answering Systems
• Sentiment Analysis
• Speech Recognition
• Text Generation

4. Levels of Language Understanding (Out of context for our syllabus)

a) Phonology:

• Study of sound patterns in language
• Important for speech recognition/synthesis

b) Morphology:

• Study of word formation
• Understanding prefixes, suffixes, root words
  Example: "unhappiness" = un (prefix) + happy (root) + ness (suffix)

c) Syntax:

• Grammar rules and sentence structure
• How words combine to form phrases and sentences

d) Semantics:

• Meaning of words and sentences
• Understanding relationships between concepts

e) Pragmatics:

• Understanding language in context
• Interpreting implied meaning and intentions

5. Major Challenges for NLP Systems:

• Handling multiple languages
• Understanding metaphors and idioms
• Resolving coreference (pronouns referring to previously mentioned entities)
• Dealing with errors in input text
• Understanding humor and sarcasm
• Managing ambiguity at various levels

Syntactic Processing

Syntactic processing is the process of analyzing the grammatical structure of sentences. It involves identifying the parts of speech, phrases, and the relationships between them. This analysis helps computers understand the underlying meaning of a sentence.

Key Concepts in Syntactic Processing

• Part-of-Speech (POS) Tagging:

  • Assigning a grammatical category (noun, verb, adjective, etc.) to each word in a sentence.
  • Example: "The quick brown fox jumps over the lazy dog."
    • POS tags: DT JJ NN VBZ IN DT JJ NN

• Chunking:

  • Identifying phrases (noun phrases, verb phrases, etc.) within a sentence.
  • Example: "The quick brown fox" is a noun phrase.

• Parsing:

  • Analyzing the grammatical structure of a sentence to create a parse tree.
  • Two main types of parsing:
    • Constituency parsing: Breaking down a sentence into its constituent parts.
    • Dependency parsing: Identifying the relationships between words in a sentence.

Challenges in Syntactic Processing

• Ambiguity:
  • A sentence can have multiple possible interpretations. For example, "I saw the man with the telescope."
• Language Variation:
  • Different languages have different grammatical rules and structures.
• Noise and Errors:
  • Real-world text often contains typos, misspellings, and other errors.

Applications of Syntactic Processing

• Machine Translation:
  • Understanding the grammatical structure of a sentence is crucial for accurate translation.
• Text Summarization:
  • Identifying the main ideas and supporting details in a text.
• Information Extraction:
  • Extracting specific information from text, such as names, dates, and locations.
• Natural Language Understanding:
  • Enabling computers to understand the meaning of text.

Parsing Techniques

Parsing is the process of analyzing the grammatical structure of a sentence. It involves breaking down the sentence into its constituent parts and identifying the relationships between them.

Main Parsing Techniques

1. Constituency Parsing:

   • Breaks down a sentence into a hierarchical structure of phrases.

   • Each phrase is a constituent of a larger phrase.

   • Uses phrase structure rules to generate parse trees.

   • Example:

     S -> NP VP
     NP -> DT NN | JJ NN
     VP -> VB NP | VB ADVP NP
     

   • Challenge: Can be computationally expensive, especially for complex sentences.

2. Dependency Parsing:

   • Identifies the grammatical relationships between words in a sentence.
   • Focuses on the dependencies between words, rather than phrases.
   • Produces a dependency tree, where words are nodes and edges represent dependencies.
   • Example:

[The] --det-> [quick] --amod-> [brown] --amod-> [fox] --nsubj-> [jumps] --advmod-> [over] --prep-> [the] --det-> [lazy] --amod-> [dog]

• Advantage: Simpler and more efficient than constituency parsing.

Challenges in Handling Ambiguous Sentences

Ambiguous sentences can pose significant challenges to parsing algorithms. Here are some common types of ambiguity:

1. Lexical Ambiguity:

   • A word can have multiple meanings.
   • Example: "The bank is open." (financial institution or river bank)

2. Structural Ambiguity:

   • A sentence can have multiple possible syntactic structures.
   • Example: "I saw the man with the telescope." (Did I use a telescope to see the man, or was the man holding a telescope?)

3. Reference Ambiguity:

   • Pronouns or other references can be ambiguous.
   • Example: "John said Mary liked him. She smiled." (Who does "she" refer to, John or Mary?)

Techniques for Handling Ambiguity

• Statistical Parsing:

  • Uses statistical models to assign probabilities to different parse trees.
  • More likely parse trees are preferred.

• Semantic Information:

  • Using semantic information can help disambiguate sentences.
  • For example, knowing the meaning of words can help resolve lexical ambiguity.

• Contextual Information:

  • Considering the context of a sentence can help resolve reference ambiguity.
  • For example, the surrounding text can help determine the antecedent of a pronoun.

Semantic Analysis

Semantic analysis is the process of understanding the meaning of text. It involves extracting the underlying meaning of words, phrases, and sentences. This is a crucial step in many NLP applications, such as machine translation, information extraction, and text summarization.

Kinda similar to how semantics are used in parsing for various LR(1) grammars in compiler design.

Key Concepts in Semantic Analysis

• Word Sense Disambiguation (WSD):

  • Determining the correct meaning of a word based on the context.
  • Example: The word "bank" can refer to a financial institution or the side of a river.

• Semantic Role Labeling:

  • Identifying the semantic roles of words in a sentence, such as agent, patient, instrument, etc.
  • Example: In the sentence "The cat chased the mouse with a stick," "cat" is the agent, "mouse" is the patient, and "stick" is the instrument.

• Textual Entailment:

  • Determining whether one text logically implies another.
  • Example: Given the sentences "The sky is blue" and "It is daytime," the first sentence entails the second.

Challenges in Semantic Analysis

• Word Sense Disambiguation:
  • Many words have multiple meanings, and the correct meaning often depends on the context.
• Semantic Role Labeling:
  • Identifying the correct semantic roles can be challenging, especially in complex sentences.
• Textual Entailment:
  • Determining logical relationships between sentences can be difficult, especially when dealing with subtle nuances and implicit meanings.

Techniques for Semantic Analysis

• Knowledge-Based Approaches:

  • Using knowledge bases and ontologies to represent semantic information.
  • Example: WordNet, ConceptNet

• Statistical Approaches:

  • Using statistical models to learn semantic relationships from large amounts of text data.
  • Example: Latent Semantic Analysis (LSA), Word2Vec

• Hybrid Approaches:

  • Combining knowledge-based and statistical approaches to improve accuracy.

Real-world Applications of Semantic Analysis

Semantic analysis plays a crucial role in many real-world applications. Here are a few examples:

1. Search Engines

• Semantic Search: Understanding the underlying meaning of search queries to provide more relevant results.
• Contextual Search: Taking into account the context of a search query to improve accuracy.

2. Information Extraction

• Extracting Key Information: Identifying and extracting relevant information from large amounts of text, such as names, dates, locations, and relationships.
• Knowledge Graph Construction: Building knowledge graphs that represent the relationships between entities.

3. Machine Translation

• Semantic Translation: Translating text while preserving the underlying meaning, rather than just word-by-word translation.
• Contextual Translation: Taking into account the context of the text to produce more accurate translations.

4. Sentiment Analysis

• Understanding Sentiment: Analyzing text to determine the sentiment (positive, negative, or neutral) expressed.
• Identifying Emotions: Detecting more specific emotions, such as anger, joy, sadness, or surprise.

5. Text Summarization

• Extractive Summarization: Identifying and extracting the most important sentences from a text.
• Abstractive Summarization: Generating a concise summary by understanding the underlying meaning of the text.

6. Chatbots and Virtual Assistants

• Understanding User Intent: Interpreting user queries to provide relevant responses.
• Generating Natural Language Responses: Generating human-like text responses to user queries.

Discourse analysis and Pragmatic processing

Discourse Analysis

Discourse analysis is the study of language beyond the sentence level. It examines how language is used in context, and how meaning is created through the interaction of different linguistic elements. It focuses on the structure and meaning of larger units of language, such as conversations, texts, and documents.

Key aspects of discourse analysis:

• Cohesion: The way sentences and paragraphs are linked together through grammatical and lexical devices (e.g., pronouns, conjunctions, reference).
• Coherence: The logical connections between ideas in a text.
• Turn-taking: The rules governing conversation, such as who speaks when and how long they speak.
• Speech Acts: The actions performed through language, such as making requests, giving orders, or asking questions.
• Pragmatics: The study of how language is used in context.

Pragmatic Processing

Pragmatic processing involves understanding the intended meaning of language, considering factors such as:

• Speaker's intent: What the speaker is trying to achieve.
• Context: The situation in which the language is being used.
• World knowledge: General knowledge about the world.

Key aspects of pragmatic processing:

• Implicature: The meaning that is implied, but not explicitly stated. For example, in the sentence "It's cold in here," the speaker might be implying that someone should close the window.
• Presupposition: The assumptions that underlie a statement. For example, in the sentence "John stopped smoking," the presupposition is that John used to smoke.
• Politeness: The use of language to maintain social harmony and avoid conflict.

Challenges in Discourse Analysis and Pragmatic Processing:

• Ambiguity: Language is often ambiguous, and the intended meaning can be difficult to determine.
• Context-dependency: The meaning of language can vary depending on the context.
• Cultural differences: Different cultures have different ways of using language.

Applications of Discourse Analysis and Pragmatic Processing:

• Natural Language Understanding: Enabling computers to understand the meaning of text and dialogue.
• Text Summarization: Identifying the main ideas and arguments in a text.
• Machine Translation: Translating text while preserving the underlying meaning and context.
• Dialogue Systems: Designing dialogue systems that can engage in natural and meaningful conversations.

Real-world Applications of Discourse Analysis and Pragmatic Processing

Discourse analysis and pragmatic processing have a wide range of real-world applications, including:

1. Natural Language Understanding (NLU)

• Chatbots and Virtual Assistants: These systems use discourse analysis to understand the context of a conversation and respond appropriately.
• Sentiment Analysis: Analyzing the sentiment of a text can be improved by considering the discourse structure and pragmatic implications of the language.

2. Text Summarization

• Identifying Main Ideas: Discourse analysis helps identify the main topics and arguments in a text.
• Preserving Coherence: Summarization systems can use pragmatic knowledge to ensure that the summary is coherent and preserves the original meaning.

3. Machine Translation

• Contextual Translation: Discourse analysis helps in understanding the context of a sentence or paragraph, which is crucial for accurate translation.
• Preserving Cultural Nuances: Pragmatic knowledge helps in understanding and translating cultural references and idioms.

4. Information Retrieval

• Semantic Search: Discourse analysis can help in understanding the semantic relationships between words and phrases, leading to more accurate search results.
• Contextual Information Retrieval: Considering the context of a query can improve the relevance of search results.

5. Legal Text Analysis

• Contract Analysis: Analyzing legal contracts to identify key clauses and potential risks.
• Case Law Analysis: Understanding the reasoning and implications of legal judgments.

6. Social Media Analysis

• Sentiment Analysis: Analyzing the sentiment of social media posts to understand public opinion.
• Identifying Trends: Identifying emerging trends and topics by analyzing large volumes of social media data.

Learning in Natural Language Processing

Learning in NLP is a fundamental aspect of developing intelligent systems that can understand and generate human language. Various learning techniques are employed to train models on large amounts of text data.

Types of Learning

1. Supervised Learning:

   • Task: Training a model on labeled data.
   • Examples:
     • Text Classification: Categorizing text into predefined classes (e.g., sentiment analysis, spam detection).
     • Named Entity Recognition (NER): Identifying named entities in text (e.g., persons, organizations, locations).
     • Part-of-Speech Tagging: Assigning grammatical tags to words in a sentence.

2. Unsupervised Learning:

   • Task: Training a model on unlabeled data.
   • Examples:
     • Topic Modeling: Discovering abstract topics present in a document collection.
     • Word Embedding: Learning semantic and syntactic relationships between words.

3. Semi-Supervised Learning:

   • Task: Training a model on a combination of labeled and unlabeled data.

   • Example: Using a small amount of labeled data to improve the performance of an unsupervised model.

4. Reinforcement Learning:

   • Task: Training a model to make decisions by interacting with an environment.

   • Example:

     • Dialogue systems can be trained to generate more engaging and informative responses through reinforcement learning.
     • Interacting with a game environment and learning by interacting within the game environment.

Learning Algorithms

• Statistical Learning:
  • Naive Bayes: A probabilistic classifier based on Bayes' theorem.
  • Support Vector Machines (SVM): A supervised learning model that classifies data points.
  • Hidden Markov Models (HMM): A statistical model for sequential data.
• Neural Networks:
  • Recurrent Neural Networks (RNNs): Model sequential data, such as text.
  • Long Short-Term Memory (LSTM) Networks: A type of RNN that can capture long-term dependencies.
  • Transformer Networks: A powerful neural network architecture for various NLP tasks, including machine translation and text generation.

Challenges in Learning

• Data Quality and Quantity: High-quality, labeled data is essential for training accurate models.
• Overfitting and Underfitting: Finding the right balance between model complexity and generalization.
• Evaluation Metrics: Choosing appropriate metrics to assess the performance of NLP models.

Inductive Learning in NLP

Inductive learning is a machine learning paradigm where a model learns general rules from specific examples. In the context of NLP, this involves training a model on a large dataset of text and labels, and then using the learned patterns to make predictions on new, unseen data.

Key Concepts in Inductive Learning for NLP

• Feature Engineering:

  • The process of selecting and transforming raw data into features that can be used by a machine learning model.
  • In NLP, features can include word frequencies, n-grams, part-of-speech tags, and syntactic dependencies.

• Model Training:

  • The process of fitting a machine learning model to the training data.
  • Common models used in NLP include:
    • Naive Bayes: A probabilistic classifier based on Bayes' theorem.
    • Support Vector Machines (SVM): A supervised learning model that classifies data points.
    • Decision Trees: A tree-like model of decisions and their possible consequences.

• Model Evaluation:

  • The process of assessing the performance of a trained model on a validation or test dataset.
  • Common evaluation metrics include accuracy, precision, recall, and F1-score.

Applications of Inductive Learning in NLP

• Text Classification:

  • Categorizing text documents into predefined classes (e.g., sentiment analysis, spam detection).

• Information Extraction:

  • Identifying and extracting specific information from text (e.g., named entity recognition, relationship extraction).

• Machine Translation:

  • Translating text from one language to another.

• Text Summarization:

  • Generating concise summaries of long documents.

Learning Decision Trees

A decision tree is a tree-like model of decisions and their possible consequences. In the context of NLP, decision trees can be used to classify text documents or to predict the next word in a sequence.

How Decision Trees Work

1. Root Node: The starting point of the tree, representing the entire dataset.
2. Internal Nodes: Represent decisions or tests on features.
3. Branches: Represent the possible outcomes of a decision.
4. Leaf Nodes: Represent the final classification or prediction.

Learning Decision Trees

The process of learning a decision tree involves:

1. Feature Selection: Choosing the best feature to split the data at each node.
2. Node Splitting: Dividing the dataset into subsets based on the selected feature.
3. Stopping Criteria: Deciding when to stop growing the tree, which can be based on factors like maximum depth, minimum number of samples, or information gain.

Challenges in Learning Decision Trees

• Overfitting: A decision tree that is too complex may overfit the training data, leading to poor performance on new, unseen data.
• Feature Selection: Choosing the right features to split the data can be challenging.
• Handling Missing Values: Decision trees can be sensitive to missing values, which can impact their performance.

Applications of Decision Trees in NLP

• Text Classification: Classifying text documents into predefined categories (e.g., sentiment analysis, spam detection).
• Information Extraction: Extracting specific information from text (e.g., named entity recognition, relationship extraction).
• Natural Language Generation: Generating text, such as product descriptions or news articles.

Explanation-Based Learning in NLP

Explanation-based learning (EBL) is a learning paradigm where a system learns by analyzing and understanding examples. In the context of NLP, EBL can be used to learn rules and patterns from text data.

How EBL Works

1. Problem Solving: A system is presented with a problem to solve.
2. Explanation Generation: The system generates an explanation for how to solve the problem, using domain knowledge and reasoning.
3. Learning from Explanation: The system extracts general rules from the explanation.
4. Rule Application: The learned rules are applied to future problems.

Challenges in EBL

• Knowledge Acquisition: Acquiring accurate and comprehensive domain knowledge can be challenging.
• Explanation Generation: Generating correct and informative explanations can be difficult, especially for complex problems.
• Rule Generalization: Generalizing rules from specific examples can lead to overfitting or underfitting.

Applications of EBL in NLP

• Text Understanding: Understanding the meaning of text by analyzing its syntactic and semantic structure.
• Question Answering: Answering questions by reasoning over knowledge bases and text documents.
• Natural Language Generation: Generating text, such as summaries or translations.

Limitations of EBL

• Knowledge Acquisition Bottleneck: EBL relies heavily on domain knowledge, which can be difficult to acquire and maintain.
• Limited Generalization: EBL may struggle to generalize from specific examples to new, unseen situations.

While EBL has the potential to create intelligent NLP systems, it is often combined with other learning techniques to overcome its limitations.

Learning Using Relevance Information

Learning using relevance information is a technique where a machine learning model is trained to improve its performance based on feedback from a user or a predefined relevance criterion. This feedback can be used to refine the model's predictions and adapt to the user's preferences.

Key Concepts in Learning Using Relevance Information

• Relevance Feedback: A user provides feedback on the relevance of retrieved documents or information.
• Query Refinement: Using relevance feedback to modify the original query and improve subsequent search results.
• Learning to Rank: Training a model to rank documents according to their relevance to a query.

Applications of Learning Using Relevance Information

• Information Retrieval: Improving the accuracy and precision of search engines.
• Recommendation Systems: Personalizing recommendations based on user preferences.
• Text Classification: Refining classification models by incorporating user feedback.

Challenges in Learning Using Relevance Information

• User Bias: User feedback can be biased, leading to inaccurate model training.
• Cold Start Problem: The model may perform poorly initially, especially when there is limited user data.
• Computational Cost: Learning from relevance information can be computationally expensive, especially for large datasets.

Techniques for Learning Using Relevance Information

• Query Expansion: Adding relevant terms to the original query based on user feedback.
• Re-ranking: Reordering search results based on user relevance judgments.
• Learning to Rank: Training a model to predict the relevance of documents to a query.

Neural Network Learning in NLP

Neural networks are a powerful class of machine learning models inspired by the human brain. They are composed of interconnected nodes, or neurons, that process information. In the context of NLP, neural networks have revolutionized the field, enabling significant advancements in various tasks.

Types of Neural Networks for NLP

1. Recurrent Neural Networks (RNNs):

   • Designed to process sequential data, such as text.
   • Long Short-Term Memory (LSTM) Networks: A type of RNN that can capture long-term dependencies in text.
   • Gated Recurrent Unit (GRU) Networks: A simplified version of LSTM networks.

2. Convolutional Neural Networks (CNNs):

   • Originally designed for image processing, but can be adapted for text processing by treating words as pixels in an image.

3. Transformer Networks:

   • A powerful architecture that has revolutionized NLP.
   • Attention Mechanism: Allows the model to focus on the most relevant parts of the input sequence.
   • Self-Attention: Enables the model to weigh the importance of different parts of the input sequence.

Applications of Neural Networks in NLP

• Machine Translation: Neural machine translation models can produce more accurate and fluent translations.
• Text Summarization: Generating concise summaries of long documents.
• Sentiment Analysis: Classifying text as positive, negative, or neutral.
• Text Generation: Generating human-quality text, such as poetry, code, or scripts.
• Question Answering: Answering questions based on a given text.

Challenges in Neural Network Learning

• Data Hunger: Neural networks require large amounts of training data.
• Computational Cost: Training neural networks can be computationally expensive.
• Black-Box Nature: It can be difficult to interpret the decision-making process of neural networks.

By understanding the principles of neural networks and their applications, we can build sophisticated NLP models that can achieve state-of-the-art performance on various tasks.

Genetic Learning in NLP

Genetic learning is a machine learning technique inspired by the process of natural selection. It involves creating a population of candidate solutions, evaluating their fitness, and selecting the fittest individuals to produce the next generation.

Key Concepts in Genetic Learning

• Chromosome: A representation of a solution, often encoded as a string of bits or symbols.
• Gene: A single unit of information within a chromosome.
• Fitness Function: A function that evaluates the quality of a solution.
• Genetic Operators:
  • Selection: Choosing individuals for reproduction.
  • Crossover: Combining parts of two parent chromosomes to create offspring.
  • Mutation: Randomly changing genes in a chromosome.

Applications of Genetic Learning in NLP

• Feature Selection: Identifying the most relevant features for a given NLP task.
• Parameter Optimization: Tuning the parameters of machine learning models.
• Grammar Induction: Learning grammar rules from a corpus of text.
• Text Summarization: Generating concise summaries of long documents.

Challenges in Genetic Learning

• Computational Cost: Genetic algorithms can be computationally expensive, especially for large-scale problems.
• Premature Convergence: The algorithm may converge to a suboptimal solution.
• Representation: Choosing a suitable representation for solutions can be challenging.

Representing and Using Domain Knowledge, Expert System Shells, and Knowledge Acquisition

Representing and Using Domain Knowledge

(We already did this in a bit more detail in the last two modules)

Domain knowledge is crucial for building intelligent NLP systems. It helps the system understand the context of the text, identify relevant information, and make informed decisions.

Methods for Representing Domain Knowledge:

• Knowledge Graphs: A graph-based representation of entities and their relationships.
• Ontologies: Formalized representations of concepts and their relationships.
• Frame-based Knowledge Representation: A structured representation of objects and their properties.

Using Domain Knowledge in NLP:

• Semantic Analysis: Understanding the meaning of text by leveraging domain knowledge.
• Information Extraction: Identifying and extracting relevant information from text.
• Question Answering: Answering questions by reasoning over domain knowledge.
• Text Generation: Generating text that is consistent with domain knowledge.

Expert System Shells

An expert system shell is a software tool that provides the infrastructure for building expert systems. It includes a knowledge base and an inference engine.

Key Components of an Expert System Shell:

• Knowledge Base: Stores domain knowledge in a structured format.
• Inference Engine: Uses reasoning techniques to draw conclusions from the knowledge base.
• User Interface: Provides a user-friendly interface for interacting with the expert system.

Applications of Expert System Shells in NLP:

• Medical Diagnosis: Assisting doctors in diagnosing diseases.
• Financial Analysis: Analyzing financial data to make investment decisions.
• Legal Reasoning: Analyzing legal documents to identify relevant information.

Knowledge Acquisition

Knowledge acquisition is the process of acquiring and organizing domain knowledge. It involves:

• Knowledge Elicitation: Interviewing experts to extract knowledge.
• Knowledge Engineering: Structuring and formalizing knowledge.
• Machine Learning: Automatically extracting knowledge from data.

Challenges in Knowledge Acquisition:

• Knowledge Elicitation: Experts may have difficulty articulating their knowledge.
• Knowledge Representation: Choosing the appropriate representation for knowledge can be challenging.
• Knowledge Validation: Ensuring the accuracy and completeness of knowledge.

By effectively representing and using domain knowledge, and by leveraging expert system shells and knowledge acquisition techniques, we can build more intelligent and sophisticated NLP systems.

A proper more in-depth module of NLP will done once this exam is out of the way.