Lecture 1: Introduction

Prompt Engineering for Language Models

Jee Eun Kim
Mar. 04, 2025
Outline
1. Class Introduction
2. Introduction to:
– Natural Language Processing (NLP)
– Artificial Intelligence (AI)

03/04/2025 Prompt Engineering 2025 2

CLASS INTRODUCTION
Class Outline
▪ Team-teaching Class with 2 Lecturers
1. 1st week ~4th week
• Introduction to Class and NLP/AI
• Linguistics and Promptification
2. 5th week ~
• Prompt Engineering Basics & Implementations
▪ Goal
– To train students to be able to efficiently utilize Generative AI
• To learn effective communication skills with Generative AI, using human
language
– To learn linguistic aspects and AI theories which are required for
manipulating Prompt Engineering
– To practice how to utilize Prompt Engineering for LLMs
03/04/2025 Prompt Engineering 2025 4
Lecturers
1. Jee Eun Kim
– Conducting classes from week 1 to week 4
– Supervising midterm (week 8)
– Contact Info
• e-mail (preferred)
– [email protected]
• Telephone
– 2173-3110 (O)
• Office
– Faculty Building #312
• Office Hour
– Upon Request/Appointment

03/04/2025 Prompt Engineering 2025 5

Lecturers (cont’d)
2. Juae Kim
– Conducting classes from week 5 to week 7 & week 9 to week 13
– Supervising final exam (week 14 or 15)
– Contact Info
• e-mail (preferred)
– [email protected]
• Telephone
– 2170-8830 (O)
• Office
– Faculty Building #622
• Office Hour
– Upon Request/Appointment

03/04/2025 Prompt Engineering 2025 6

Textbook
▪ There is no appropriate textbook available
– Class will provide lecture notes and references
• The lecture notes will be uploaded to e-class before each class
• Will use Chap. 12 of Speech and Language Processing (SLP) as one of
the references
– Daniel Jurafsky & James H. Martin. Prentice Hall. 3rd ed. Jan. 2025. (Online
draft version)
» Uploaded to e-class

03/04/2025 Prompt Engineering 2025 7

e-class
▪ Prompt Engineering
– Folders
• Lecture Notes
– Lecture notes are to be uploaded before class
• Assignments
– Homework to be uploaded
– File format for class notes
• pdf format: to read/view the file
– Go to http://get.adobe.com/kr/reader/
– Install Adobe Reader

03/04/2025 Prompt Engineering 2025 8

Syllabus
▪ Mar. 04
– Introduction to Class and NLP
– Mar. 11 ~ 25
• Introduction to Prompt Engineering from Theoretical perspectives
– Apr. 01 ~
• Introduction to Prompt Engineering from Practical perspectives
▪ Mar. 11
– Linguistics and Prompt Engineering
▪ Mar. 18
– Interface & Prompt
▪ Mar. 25
– Discussion based on Assignments
03/04/2025 Prompt Engineering 2025 9
Syllabus (cont’d)
▪ Apr. 01
– Word Embedding
▪ Apr. 08
– Language Models based on Deep Learning
▪ Apr. 15
– Prompt Engineering Concepts
▪ Apr. 22
– Midterm
▪ Apr. 29
– Prompt Engineering Strategies
03/04/2025 Prompt Engineering 2025 10
Syllabus (cont’d)
▪ May 06
– Holiday
▪ May 13
– Advanced Prompting Techniques
▪ May 20
– Prompt Engineering Practice #1
▪ May 27
– Prompt Engineering Practice #2
▪ Jun. 03
– Prompt Engineering Practice #3
03/04/2025 Prompt Engineering 2025 11
Syllabus (cont’d)
▪ Jun. 10
– A make-up class
– The future of Prompt Engineering
▪ Jun. 17
– Final Exam

03/04/2025 Prompt Engineering 2025 12

Evaluation
▪ Attendance
– 10%
▪ Assignments
– 20% (final assignments: 10%)
▪ Midterm Exam
– 30%
▪ Final Exam
– 40%

03/04/2025 Prompt Engineering 2025 13

INTRODUCTION TO NLP
What Is NLP?
▪ Natural Language Processing
– Having existed for more than 50 years
– Having roots in the field of linguistics

▪ “NLP may be the fastest-developing and most important field

of Artificial Intelligence (AI) & Data Science. If you want to
change the world you will need to understand how machines
read and process natural language text."
– Hobson Lane and Maria Dyshel. 2nd ed. of Natural Language
Processing in Action. Manning Publications Co. June, 2024.

03/04/2025 Prompt Engineering 2025 15

 What Is NLP? (cont’d)
▪ The intersection of computer science, linguistics and machine
learning (ML)
– All about making computers understand and generate human
language referred to as Natural Language (NL)
– The ability of a computer program to understand and interpret human
language as it's spoken and written
▪ A branch of AI
– A technique used in AI algorithms
– Dealing with analyzing, understanding, and generating the languages
that humans use naturally
– Dealing with the interaction between computers and humans using NL
• To interface with computers in both written and spoken contexts
using natural human languages instead of computer languages
03/04/2025 Prompt Engineering 2025 16
 What Is NLP? (cont’d)
▪ While computers excel at processing structured data, such as
spreadsheets or databases, natural language in its
unstructured form (text, speech, etc.) presents a unique
challenge
▪ NLP bridges this gap by allowing machines to process and
understand human languages, making it an essential tool in
modern AI systems

03/04/2025 Prompt Engineering 2025 17

 What Is NLP? (cont’d)
▪ Communication With Machines
– Human language has become a interface between human and
machines

03/04/2025 Prompt Engineering 2025 18

 What Is NLP? (cont’d)
▪ Communication With Machines (cont’d)

03/04/2025 Prompt Engineering 2025 19

 What Is NLP? (cont’d)
▪ One of the hottest areas of AI
– Thanks to applications such as:
• Text Generators that compose coherent essays
• Chatbots that fool people into thinking they’re sentient
• Text-to-image programs that produce photorealistic images of anything
you can describe
– https://openai.com/dall-e-3
– https://www.technologyreview.com/2024/02/15/1088401/openai-amazing-
new-generative-ai-video-model-sora/
– Recent years have brought a revolution in the ability of computers to
understand human languages, programming languages, and even
biological and chemical sequences, such as DNA and protein
structures, that resemble language
– The latest AI models are unlocking these areas to analyze the
meanings of input text and generate meaningful, expressive output
03/04/2025 Prompt Engineering 2025 20
 What Is NLP? (cont’d)

03/04/2025 Prompt Engineering 2025 21

 What Is NLP? (cont’d)

03/04/2025 Prompt Engineering 2025 22

 What Is NLP? (cont’d)
▪ The ultimate objective of NLP
– To develop algorithms and models that enable computers to
understand, interpret, and generate human languages
• Focusing on the interaction between computers and human language
– To read, decipher, understand, and make sense of the human
languages in a manner that is valuable
– To bridge the communication gap between humans and computers
• NLP models are typically trained on tasks such as next word prediction
which allow them to build contextual dependencies and then be able to
generate relevant outputs
▪ NLP plays a crucial role in various industries
– Continuing to advance with the development of more sophisticated
algorithms and technologies
03/04/2025 Prompt Engineering 2025 23
 What Is NLP? (cont’d)
▪ NLP techniques
– Heavily benefited from recent advances in ML, especially from deep
learning (DL) techniques to derive meaning from NLs
• Speech Recognition (SR)
– The translation of spoken language into text
• Natural language understanding (NLU)
– The computer’s ability to understand what we say or write
• Natural language generation (NLG)
– The generation of NL by a computer

03/04/2025 Prompt Engineering 2025 24

 What Is NLP? (cont’d)
▪ Applications of NLP techniques
– Machine translation (Google Translate, Papago, etc.)
– Web Search
– Text-filtering such as Spam filters
– Sentiment Analysis
– Voice assistants such as Amazon’s Alexa and Apple’s Siri
– Chatbots
– etc.
▪ NLP has a variety of real-world applications in numerous fields
– Including medical research, business intelligence, customer service,
content analysis, etc.
03/04/2025 Prompt Engineering 2025 25
 What Is NLP? (cont’d)

03/04/2025 Prompt Engineering 2025 26

Multimodal NLP
▪ Emergence of Multimodal NLP
– By integrating multiple data modalities
– The availability of diverse data types, such as images, audio, and videos,
catalyzed this shift from unimodal to multimodal NLP
– Researchers realized that combining these modalities could significantly
enhance language understanding and context
▪ The modern era
– Multimodal NLP has evolved into a vibrant field where researchers explore
sophisticated models capable of simultaneously understanding, generating,
and processing multiple modalities
– State-of-the-art models leverage neural architectures like Transformers,
which have shown remarkable prowess in handling diverse data types

03/04/2025 Prompt Engineering 2025 27

Multimodal NLP (cont’d)
▪ Involving the combination of different types of information,
such as text, speech, images, and videos, to enhance NLP tasks
– Allowing machines to better comprehend human communication by
taking into account additional contextual information beyond just text
– For example:
• Can be used to enhance machine translation by integrating visual data
from images or videos to provide better translations
• Can also be used to improve sentiment analysis by incorporating non-
textual data such as facial expressions or tone of voice
▪ A growing field of study which is expected to become
increasingly significant as more data becomes available across
multiple modalities
03/04/2025 Prompt Engineering 2025 28
Multimodal NLP (cont’d)
▪ Involving challenges:
1. Multimodal Representation
• Integrating information from different modalities into a cohesive
representation that a machine can understand and process
2. Alignment and Fusion
• Aligning information from different modalities to ensure that the data
from each modality complements and enhances the understanding of the
overall context
3. Multimodal Understanding
• Developing models that interpret and comprehend information from
multiple modalities simultaneously
• For example
– It is understanding both the text and the visual content of an image or video
03/04/2025 Prompt Engineering 2025 29
Multimodal NLP (cont’d)
▪ Involving challenges: (cont’d)
4. Multimodal Generation
• Creating systems capable of generating outputs that encompass
multiple modalities
• For example:
– It generates a description of an image or creates a video based on a textual
description

03/04/2025 Prompt Engineering 2025 30

Multimodal NLP (cont’d)
▪ The Significance of Multimodal NLP
– Lying in its ability to bridge the gap between language and other forms
of information
– By understanding images, videos, and audio alongside textual data,
machines can achieve a more holistic comprehension of the world,
mirroring human-like understanding
– This evolution from text-centric NLP to multimodal approaches
signifies a paradigm shift in how machines perceive and interpret
information, setting the stage for more contextually aware and
comprehensive AI systems

03/04/2025 Prompt Engineering 2025 31

Multimodal NLP (cont’d)
▪ The Significance of Multimodal NLP (cont’d)

03/04/2025 Prompt Engineering 2025 32

Multimodal NLP (cont’d)
▪ Applications
– Visual Question Answering (VQA)
– Automatic Image and Video Captioning
– Sentiment Analysis in Multimedia
– Multimodal Translation and Transcription
– Enhanced Content Recommendation Systems
– Multimodal Assistive Technologies
– Context-aware conversational AI
– Healthcare Applications
▪ As research in this field progresses, the potential for innovative
applications continues to expand, promising more
sophisticated and context-aware AI solutions
03/04/2025 Prompt Engineering 2025 33
Multimodal NLP (cont’d)
▪ Implementation
1. Data Preparation
• Gather a multimodal dataset of relevant data, such as text-image pairs,
audio-text transcripts, or video-text descriptions.
• Ensure the data is clean and well-structured, handling missing values,
inconsistencies, and data quality issues.
• Divide the data into training, validation, and testing sets for model
evaluation.

03/04/2025 Prompt Engineering 2025 34

Multimodal NLP (cont’d)
▪ Implementation
2. Feature Extraction
• Use appropriate algorithms to extract meaningful features from each
modality (text, images, audio, video).
– For text, use text processing techniques such as tokenization, stemming,
lemmatization and word embedding to represent text as numerical vectors
– Use image processing techniques like Convolutional Neural Networks
(CNNs) to extract visual features for images
– For audio, use audio processing techniques like Mel-frequency cepstral
coefficients (MFCCs) to extract acoustic features.
– For video, use video processing techniques like 3D convolutional neural
networks (3D CNNs) to extract temporal and spatial features.

03/04/2025 Prompt Engineering 2025 35

Multimodal NLP (cont’d)
▪ Implementation
3. Fusion
• Combine the extracted features from different modalities into a unified
representation
• Common fusion techniques include:
– Early Fusion: Concatenate or average features from different modalities at
the input layer.
– Late Fusion: Pass features through separate unimodal models and combine
their outputs at a later stage.
– Attention Mechanisms: Use attention mechanisms to weigh and combine
features based on their relevance dynamically.

03/04/2025 Prompt Engineering 2025 36

Multimodal NLP (cont’d)
▪ Implementation
4. Model Training
• Choose an appropriate deep learning architecture for your multimodal
task, such as CNN-RNN, Transformer, or a combination of both
• Train the model on the prepared multimodal data using a suitable loss
function and optimization algorithm
• Monitor the model’s performance on the validation set to prevent
overfitting
5. Evaluation and Deployment
• Evaluate the trained model on the testing set to assess its generalizability
and performance on unseen data
• Deploy the model to production in the desired environment, such as a
web application or mobile app
• Monitor & refine the model to adapt to new data & improve performance
03/04/2025 Prompt Engineering 2025 37
NLP Trends For 2025
▪ By leveraging NLP, we not only break the language barrier but
also push machines to an extent where they can understand
the intent behind the query without much of an explanation
– https://www.shaip.com/blog/nlp-trends-2025/

1. Real-Time Language Translation

– Based on current advancements in
NLP, these models can achieve up
to 98% accuracy when translating
spoken and written languages

03/04/2025 Prompt Engineering 2025 38

NLP Trends For 2025 (cont’d)
2. Deep Learning Models for Specialized Tasks
– We are witnessing Transformers models like GPT-4 and BERT are
achieving excellent accuracy and in 2025, they will surely reach new
dynamics of possibilities
– In our testing, we saw that these models can now handle niche tasks
like drafting legal contracts and analyzing medical records of patients
with close to human-like precision
– Furthermore, when fine-tuned, you can customize them for industries
like finance and law

03/04/2025 Prompt Engineering 2025 39

NLP Trends For 2025 (cont’d)
3. Better Emotional Intelligence
– Modern AI models now go beyond merely identifying positive or
negative sentiments
– They can detect a wide range of emotions such as anger, joy,
frustration, and more
– This capability allows for a deeper understanding of human
interactions

03/04/2025 Prompt Engineering 2025 40

NLP Trends For 2025 (cont’d)
4. Better Healthcare
– Hospitals with NLP can extract data
from unstructured sources like clinical
notes and medical reports
• Doctors can identify patterns in
patients’ clinical history, predict diseases, and suggest treatments
5. Conversational AI Gets Even Better
– Apple integrated ChatGPT into Siri
– Google integrated Gemini to Google
Assistant
– These chatbots will be capable
enough to distinguish between
sarcasm and genuine requests
03/04/2025 Prompt Engineering 2025 41
NLP Trends For 2025 (cont’d)
6. Ethical AI will be Prioritized more than Ever
– As NLP becomes more and more powerful, it will raise concerns about
biases and privacy
– It will eventually raise concerns as models trained over biased data
will discriminate in hiring and lending
– To solve this, we might witness the formation of multiple regulatory
authorities to mandate transparency, forcing companies to disclose
training data sources

03/04/2025 Prompt Engineering 2025 42

NLP Trends For 2025 (cont’d)
7. E-Commerce gets Personalized
– Companies would be able to use NLP
to analyze browsing patterns & provide
tailored recommendations to the user
• Using semantic search and
personalized suggestions
8. The Age of Hybrid AI Systems
– NLP once matured enough will be
integrated into computer vision
applications like automated medical
report generation and real-time image
captioning

03/04/2025 Prompt Engineering 2025 43

NLP Trends For 2025 (cont’d)
9. Multilingual Model Support
– As of now, NLP systems can handle 300+ languages and with
initiatives like Google’s Universal Speech Model (USM), the aim is to
cover 1,000 languages
• Currently, USM supports 400+ languages including some low-resource
languages like Amharic and Assamese, enhancing accessibility in
regions like Africa and South Asia

03/04/2025 Prompt Engineering 2025 44

What’s Special about Language?
▪ A human language is a system specifically constructed to
convey the speaker/writer’s meaning
– Not just an environmental signal, it’s a deliberate communication
– Using an encoding which little kids can quickly learn
▪ A human language is a discrete/symbolic/categorical
signaling system
– Rocket = ; violin =
– With very minor exceptions for expressive signaling
• eg., I loooove it

03/04/2025 Prompt Engineering 2025 45

What’s Special about Language? (cont’d)
▪ The categorical symbols of a language can be encoded as a
signal for communication in several ways:
– Sound
– Gesture
– Images (writing)
▪ The symbols is invariant across different encodings!

03/04/2025 Prompt Engineering 2025 46

What’s Special about Language? (cont’d)
▪ A human language is a symbolic/categorical signaling system
▪ However, a brain encoding appears to be a continuous pattern
of activation, and the symbols are transmitted via continuous
signals of sound/vision
▪ We will explore a continuous encoding pattern of thought
▪ The large vocabulary, symbolic encoding of words creates a
problem for machine learning - Sparsity

03/04/2025 Prompt Engineering 2025 47

Why is Language Hard?
▪ Ambiguities on many levels
▪ Rules, but many exceptions
▪ No clear understand how humans process language

03/04/2025 Prompt Engineering 2025 48

What Makes Language Hard?
▪ The little words
– I read a book in Chinese.
– I read a book on Chinese.
▪ How we say it
– Different meaning
• The white house is on the corner.
• The White House is on the fence.
– Focusing attention on corrections
• We will send the response to Donald Metzger.
• No that’s Ronald Metzger vs. No, that’s Donald Fetzger.
– Emphasizing important information
• I need to be in Boston no later than 9 am.
03/04/2025 Prompt Engineering 2025 49
Ambiguity
▪ A fundamental problem of NLP
– Researchers in NLP are obsessed with ambiguity
▪ Resolving ambiguity is a crucial goal
▪ Ambiguity is Pervasive

03/04/2025 Prompt Engineering 2025 50

Ambiguity (cont’d)
▪ Where are the ambiguities?
– Part of speech ambiguities
• saw as a tool or a verb form
– Word sense ambiguities & homographs
• bat (baseball vs. mammals)
• The ink is in the pen. vs. The pig is in the pen.
– Syntactic attachments could be complex and lead to ambiguities
• I saw the man with the telescope.

▪ Why is NLP difficult?

– Ambiguities in NL
• I saw a man with a telescope.
• 여기 물이 참 좋군!
03/04/2025 Prompt Engineering 2025 51
Ambiguity (cont’d)
▪ Automatic disambiguation
– Part-of-speech tagging
– Word sense disambiguation
– Syntactic disambiguation
– Speech act interpretation

03/04/2025 Prompt Engineering 2025 52

Variation
▪ Lexical Variation
– The different words and phrases we use to refer to the same entity
– e.g.,
• color & colour
• PRC &China
• truck & lorry, elevator & lift, sofa & settee
• trousers & pants, sofa & couch
• beautiful, beauty, beautify & beautifully
• etc.

03/04/2025 Prompt Engineering 2025 53

Variation (cont’d)
▪ Syntactic Variation
– The study of sentence structure and how sentence structure interacts
– aka, grammatical variation
– Dealing with the ways in which sentences are structured and
constructed
– e.g.,
• I am going shopping later. I am going to go shopping later.
• Anybody want no tea. Nobody wants tea. or There isn’t anybody who
wants tea.
• She gave him the book. He was given the book by her. The book was
given to her.
• etc.

03/04/2025 Prompt Engineering 2025 54

Problems in Processing Natural Language
▪ When people see text, they understand its meaning (by and
large)
▪ When computers see text, they get only character sequences
(and perhaps HTML tags)
▪ We'd like computer agents to see meanings or be able to
intelligently process text
▪ Why is Natural Language Understanding (NLU) so complex?

03/04/2025 Prompt Engineering 2025 55

Why is NLU difficult?
1. Infinite diversity of sentences
a. The vocabulary is not completely known (out-of-vocabulary problem
OOV)
b. The set of constructions is itself not completely predetermined
c. The set of senses attributed to each word is also not completely
predetermine

2. Tolerance of errors (robustness)

03/04/2025 Prompt Engineering 2025 56

Why is NLU difficult? (cont’d)
3. Implicit elements
a. Contraction
• « Anne didn’t promised that… »
b. Anaphoric reference
• « Mr Major arrived in France today. The prime minister will meet the
President tomorrow. The Conservative leader will then travel to Moscow
where he will meet Mr Gorbachev. Mrs Major will join her husband in
Russia, where this son of a circus artist is relatively unknown figure. »
c. Ellipses
• « John is having dinner with Mary tomorrow night, and Paul with Susan. »

03/04/2025 Prompt Engineering 2025 57

Why is NLU difficult? (cont’d)
4. The hidden structure of language is highly ambiguous
Structures for
– Times flies like an arrow.

03/04/2025 Prompt Engineering 2025 58

Why is NLU difficult? (cont’d)

▪ But that’s what makes it fun!

03/04/2025 Prompt Engineering 2025 59
Why is NLU difficult? (cont’d)
▪ Making progress on the problems
– The task is difficult! What tools do we need?
• Knowledge about language
• Knowledge about the world
• A way to combine knowledge sources
– How we generally do this:
• Probabilistic models built from language data
– P(“maison” → “house”) high
– P(“L’avocat général” → “the general avocado”) low
• Luckily, rough text features can often do half the job

03/04/2025 Prompt Engineering 2025 60

Conversational Agents
▪ Any dialogue system
– Conducting NLP
– Responding automatically using human language
– Representing the practical implementation of computational
linguistics
▪ A virtual agent
– Usually deployed as chatbots and virtual or AI assistants
– Used to communicate with a human in natural language
– simulating human-to-human interaction
– understanding context and meaning just as humans do

03/04/2025 Prompt Engineering 2025 61

Conversational Agents (cont’d)
▪ Chatbot
– A software program designed to simulate human conversation
– Used to provide information, answer questions, perform tasks, and
make purchases
– Implemented with a rule-based approach
• Meaning that they can only respond based on specific text or button
inputs
– More narrowly focused on their objectives
• For example, they're either employed to resolve particular customer
queries
– such as looking up an insurance policy or help with the e-commerce
checkout process

03/04/2025 Prompt Engineering 2025 62

Conversational Agents (cont’d)
▪ Conversational agents
– Programs that use NLP and NLU technology to converse with humans
– Understanding human emotions, answer basic questions, respond to
commands, and interact through natural language conversations
– Often used to automate customer support and marketing campaigns
– Their platforms construct and deploy chatbots and virtual assistants
that allow users to interact with them through natural language
conversations in order to provide a solution

▪ Conversational agents vs. chatbots

– Chatbots are conversational agents
– But conversational agents are NOT necessarily chatbots
03/04/2025 Prompt Engineering 2025 63
Conversational Agents (cont’d)
▪ Switchboard Receptionists
– Implications
• Need to know how a native or non-native speaker of English is likely to
pronounce a random American name
▪ Airline and other simple dialogue systems
– Lots of fabulously fun pragmatic issues
• Grounding
– The process of adding to common ground between conversational
participants
• Error handling
▪ In general
– Now that there are commercial installations
– Big incentive to use linguistic or any other kind of knowledge if it
actually improves performance
03/04/2025 Prompt Engineering 2025 64
 Modern Conversational Agents
▪ Conversational agents contain:
– Speech recognition
– Language analysis
– Dialogue processing
– Information retrieval
– Text to speech

03/04/2025 Prompt Engineering 2025 65

 Modern Conversational Agents (cont’d)
▪ ChatGPT
– A new framework for creating conversational agents based on
generative pre-trained transformer (GPT) models
• Large neural networks to generate natural language texts from a given
text input or prompt (benefits of a conversational agent)
– Leveraging the power of GPT models to create engaging and natural
chatbots that can handle a variety of tasks and domains
– Not replacement for conversational AI platforms
• ChatGPT is a specific ML model that is designed to generate human-like
text based on a given prompt or context (not having enough features)
• Conversational AI platforms typically include a range of tools and
features that are necessary for any virtual assistant to function
effectively from building, designing and deploying a bot to including bot
life cycle management, NLP algorithms, ML models, and enterprise
integrations
03/04/2025 Prompt Engineering 2025 66
Question Answering
▪ Sample questions
– What does “door” mean?
– What year was Abraham Lincoln born?
– How many states were in the United States when Lincoln was born?
– Was there a military draft during the Hoover administration?
– What do US scientists think about whether human cloning should be
legal?

03/04/2025 Prompt Engineering 2025 67

Question Answering (cont’d)
▪ What’s needed?
– Speech recognition and synthesis
– Knowledge of the English words involved
• What they mean
– How groups of words clump
• What the clumps mean
– Dialog
• It is polite to respond, even if you’re planning to kill someone.
• It is polite to pretend to want to be cooperative
– I’m afraid, I can’t…

03/04/2025 Prompt Engineering 2025 68

Question Answering (cont’d)
▪ Caveat
– NLP has an AI aspect to it
• We’re often dealing with ill-defined problems
• We don’t often come up with exact solutions/algorithms
• We can’t let either of those facts get in the way of making progress

03/04/2025 Prompt Engineering 2025 69

INTRODUCTION TO AI
What Is It?
▪ Definition
– Intelligence
• The capacity to acquire knowledge and apply it to achieve an outcome
– The action taken is related to the particulars of the situation rather than
done by rote
– Artificial Intelligence
• Getting a machine to perform in the manner describe above
• But there is no single or simple definition of AI
1. A computerized system that exhibits behavior that is commonly thought
of as requiring intelligence
2. A system capable of rationally solving complex problems or taking
appropriate actions to achieve its goals in whatever real world
circumstances it encounters
3. etc.
03/04/2025 Prompt Engineering 2025 71
What Is It? (cont’d)
▪ Artificial intelligence (AI)
▪ The simulation of human intelligence in machines that are
programmed to think and act like humans
– Examples of cognitive abilities
• Learning
• Reasoning
• Problem-solving
• Perception
• Language comprehension
• etc.
• https://www.simplilearn.com/tutorials/artificial-intelligence-
tutorial/what-is-artificial-intelligence#what_is_artificial_intelligence

03/04/2025 Prompt Engineering 2025 72

What Is It? (cont’d)
▪ A method of making a computer, a computer-controlled robot,
or a software think intelligently like the human mind
– Accomplished by:
1. Studying the patterns of the human brain
2. Analyzing the cognitive process
– The outcome of these studies develops intelligent software and
systems
– For example:
• NLP uses AI to analyze and interpret human language in text or speech
– It enables applications like chatbots and virtual assistants to understand
user queries, extract meaningful information, and deliver accurate,
context-aware responses, transforming unstructured communication into
actionable insights

03/04/2025 Prompt Engineering 2025 73

What Is It? (cont’d)
▪ AI & Generative AI

https://blog.dataiku.com/large-language-model-chatgpt

03/04/2025 Prompt Engineering 2025 74

What Is It? (cont’d)

03/04/2025 Prompt Engineering 2025 75

What Is It? (cont’d)
▪ Types of AI
– The emergence of artificial superintelligence will change humanity
• But it’s not happening soon

03/04/2025 Prompt Engineering 2025 76

Best AI Jobs in 2025
1. AI/ML Engineer
– Focusing on designing, building, and maintaining AI systems that
automatically learn and improve from experience
– utilizing machine learning frameworks like TensorFlow and PyTorch to
develop predictive models with various applications, from
recommendation systems to autonomous vehicles
– Qualifications
• Requiring a degree in Computer Science, Statistics, or a related field
• Strong programming skills in Python, R, or Java and an understanding of
ML algorithms are essential

03/04/2025 Prompt Engineering 2025 77

Best AI Jobs in 2025 (cont’d)
2. Data Scientist
– Analyzing vast amounts of raw information to find patterns that
streamline a company’s processes
– Using statistical tools and algorithms to generate insights that drive
strategic business decisions
– Qualifications
• Requiring a degree in data science, statistics, computer science, or a
related field
• Proficiency in SQL, Python, R, and specialized data analytics tools like
Tableau or SAS

03/04/2025 Prompt Engineering 2025 78

Best AI Jobs in 2025 (cont’d)
3. AI Research Scientist
– Developing new approaches to AI technology
– Involving creating innovative machine-learning techniques or cognitive
computing systems
– Qualifications
• Often requiring a PhD in a related discipline, such as computer science,
cognitive science, or neural networks
• Extensive knowledge of multiple AI disciplines, including machine
learning, deep learning, and computational statistics, is essential

03/04/2025 Prompt Engineering 2025 79

Best AI Jobs in 2025 (cont’d)
4. AI Ethics Officer
– Ensuring that AI technologies are developed and used in a way that is
ethical and compliant with existing laws and regulations
– working on guidelines that help shape the ethical development of AI
applications
– Qualifications
• Requiring a background in ethics/law and additional training in AI or
technology
• Knowledge of current AI technologies and the regulatory landscape is
important
5. Robotics Engineer

03/04/2025 Prompt Engineering 2025 80

Best AI Jobs in 2025 (cont’d)
6. Natural Language Processing (NLP) Engineer
– Developing algorithms that allow computers to understand and
process human languages in a valuable way, enabling applications
such as chatbots and translation services
– Qualifications
• Requiring strong programming skills and a good understanding of
linguistics
• Typically requiring a degree in Computer Science or Computational
Linguistics

03/04/2025 Prompt Engineering 2025 81

Best AI Jobs in 2025 (cont’d)
7. AI Product Manager
8. Computer Vision Engineer
9. AI Safety Engineer
10. Chief AI Officer

03/04/2025 Prompt Engineering 2025 82

What’s Next?
▪ Introduction to Large Language Models (LLMs)
▪ Linguistics and Prompt Engineering (PE)
– Introduction to PE
– Contribution of linguistic knowledge to PE

03/04/2025 Prompt Engineering 2025 83