Introduction to Natural Language Processing (Part 1)

Language: Broadly, language is a communication system that uses a set of common processing rules to express instructions, which can be conveyed through visual, auditory, or tactile means. It can be categorized into: natural language, animal language, and computer language.

Natural language: An information exchange method developed by humans, including spoken and written forms, reflecting human thought. It naturally evolves with culture and serves as a medium for human communication, transmitted through visual, auditory, or tactile means.

• Turing Test: Proposed by Alan Turing, used to assess whether a machine exhibits human-level intelligence. In the test, the tester interacts with a participant (a human and a machine) through a keyboard. If the machine leads an average participant to make more than 30% incorrect judgments after multiple tests, it is considered to possess human intelligence.

• Ultimate Goal: Requires computers to have human language capabilities, meaning they must understand (listen, read) and generate (speak, write) language to achieve natural language communication between humans and machines.

Image source: 2019_knowledge_guided_nlp_cn(tsinghua.edu.cn)

Natural language processing broadly includes natural language understanding and natural language generation. Historically, more research has been conducted on natural language understanding, while natural language generation has received less attention, though this situation is changing.

a. What is Natural Language Processing

• Definition: Natural language processing is a branch of computer science and artificial intelligence aimed at developing technologies that can understand and generate human language. It involves analyzing, understanding, and generating natural language through formal computational models.

b. Ultimate Goal: To enable computers to have natural language communication abilities

• Natural Language Input: The input stage receives user natural language data.

• Computer: The core processing unit responsible for processing the input natural language, including understanding and generating language.

• Natural Language Output: The output stage generates natural language responses or actions based on processing results.

c. Two Core Issues: NLU, NLG

• Natural Language Understanding (NLU): The goal is for machines to have the ability to understand human language like humans do.

• Natural Language Generation (NLG): To generate natural language as a response. (Converting non-linguistic data into human language format to facilitate human-machine interaction)

• Branch of Artificial Intelligence: Natural language processing is a subfield of artificial intelligence, alongside machine learning, data mining, computer vision, and robotics.

• Cross-disciplinary Field: Natural language processing is an interdisciplinary field involving computer science (tools for NLP research), linguistics (the subject of processing), mathematics (providing theoretical foundations), cognitive science, psychology, and philosophy.

Text Mining (Data Mining)

• Text mining primarily focuses on extracting useful information and knowledge from text data, typically involving techniques such as information retrieval, text classification, text clustering, sentiment analysis, and topic modeling.

• The goal of text mining is to convert unstructured text into structured data for further analysis.

Speech Recognition (Electronic Engineering)

• Speech recognition involves parsing and converting human speech into machine-readable formats. This complex process includes acoustic models and language models. Speech recognition technology enables devices to understand and execute voice commands, widely used in smart assistants and automated customer service.

Natural Language Processing (NLP) and Computational Linguistics are closely related but distinct fields. Both involve using computer technology to process and understand human language, but their research focuses, methods, and applications differ.

Natural Language Processing (NLP)

Natural language processing is a branch of computer science and artificial intelligence focused on enabling computers to understand, interpret, and generate human language.

• Research Focus: NLP emphasizes engineering and technical aspects, aiming to develop practical technologies and systems capable of efficiently performing language processing tasks, such as speech recognition, machine translation, and sentiment analysis.

• Methods: NLP typically employs machine learning techniques, including deep learning, to process large-scale language data and learn how to solve specific language processing problems.

• Applications: NLP has wide-ranging applications, including chatbots, voice assistants, and automatic summarization systems, with the goal of creating systems that work effectively in real-world environments.

Computational Linguistics

Computational linguistics is a branch of linguistics that uses computational methods to study language. It focuses on understanding the fundamental properties of language, including computational models of language structure.

• Research Focus: Computational linguistics is more concerned with establishing theoretical models, including the formal and structural aspects of language, such as grammar, semantics, and phonetics.

• Methods: Computational linguistics may employ formal methods, such as statistical models, algorithms, and theoretical analysis, to explore language rules and structures.

• Applications: Applications of computational linguistics are often academic, focusing on enhancing understanding of how language works to improve and guide the development of NLP technologies.

Differences and Connections

• Differences: NLP focuses more on technology and applications, aiming to develop practical language processing tools; computational linguistics, on the other hand, emphasizes theory, aiming to understand the workings and structures of language.

• Connections: Despite their differing focuses, the two fields are closely linked. Many of the techniques and algorithms used in NLP are based on the theoretical foundations established by computational linguistics. Research outcomes in computational linguistics can guide and improve the design and implementation of NLP systems.