AI THE

AND THE
PRESENT
( UNIT
FUTURE
5)
01
AI THE PRESENT
AND THE FUTURE
NAME- ARYAN
ROLL NO. - AE-1254

SYMBOLIC AI
NAME- JAYA DAS
ROLL NO. - AE-1260

MACHINE LEARNING
FOR AI
NAME- JANISHA SHARMA
ROLL NO. - AE-1223

DEEP LEARNING
FOR
NAME-
AI ABHINAV ANAND
ROLL NO. - AE-1201
 Artificial
Intelligence
• Artificial Intelligence (AI) is a branch of computer
science that focuses on designing and developing
algorithms and computational systems capable of
performing tasks that require human-like
cognitive abilities.
• These include learning, reasoning, problem-
solving, perception, and natural language
understanding.
• AI leverages techniques such as machine
learning, neural networks, probabilistic reasoning,
and optimization to create intelligent systems
that can adapt and improve through experience.
 How Does AI
Work?
Machine Deep
Learning Learning
Machine learning enables Deep learning, a subset of
systems to learn from data machine learning, uses
without explicit programming. artificial neural networks to
Algorithms identify patterns and process information in a
make predictions based on vast hierarchical way, mimicking
amounts of information. the human brain's structure.
AI: The Present and the
Future
Prese
nt
AI is already deeply integrated into our daily lives,
powering everything from search engines and
1
virtual assistants to fraud detection and
personalized recommendations.

Futur
AI e is poised to revolutionize
2
healthcare, transportation, energy,
and countless other sectors, creating
unprecedented opportunities and
challenges.
Overview of AI: From Inception to
Today
1950s: The concept of AI emerges with
Alan Turing's "Imitation Game" and
1980s-90s: The "AI winter" sets in due to
limitations in computing power and data
John McCarthy's Dartmouth Workshop. availability.

1 2 3 4

2000s-present: The rise of machine learning

1960s-70s: Early AI research focuses
and deep learning, driven by advancements
on rule-based systems and expert
in computing and data storage, leads to
systems, with limited success.
significant progress in AI capabilities.
 Current Applications and
Capabilities of AI
Reta Financial Services Healthca
il
AI powers personalized AI is used for credit scoring, re diagnosis,
AI aids in medical
recommendations, inventory fraud detection, risk drug discovery,
management, fraud detection, management, and personalized treatment
and customer service chatbots, automated trading, plans, and robotic surgery,
enhancing customer improving efficiency and advancing healthcare
experiences. accuracy. capabilities.

Insuran Recruitme
ce nt
AI facilitates claims processing, risk assessment, fraud AI helps in candidate screening, talent sourcing,
detection, and personalized pricing, improving and job matching, optimizing recruitment
efficiency and accuracy. processes and improving candidate experience.
Ethical Considerations and
Challenges in AI
Development
Bias and Privacy and Data
Discrimination
AI systems trained on Security
AI raises concerns about data
biased data can perpetuate privacy and security, requiring
existing social inequalities, robust safeguards to protect
requiring careful sensitive information.
consideration of fairness
and inclusivity.
Transparency and Job Displacement and
Explainability Economic Impact
Understanding how AI AI-powered automation can
systems make decisions is lead to job displacement,
crucial for trust and necessitating proactive policies
accountability, requiring to address workforce
efforts to make them transitions and reskilling.
transparent and explainable.
AI-Powered Automation and its Impact on
Increased Jobs
1 Efficiency
AI automation can streamline processes, reducing human error and freeing up employees for
higher-level tasks.

Job
2 Displacement
Certain jobs that involve repetitive tasks are susceptible to automation, potentially
leading to job displacement.

New Job
3 Creation
AI also creates new opportunities in fields related to AI development,
maintenance, and deployment.

Reskilling and
4 Upskilling
Adapting to the changing job market requires reskilling and upskilling,
ensuring a skilled workforce for AI-driven industries.
 What is Symbolic
Definition: AI?
• Symbolic Artificial Intelligence (AI) is a subfield of AI that
focuses on the processing and manipulation of symbols or
concepts, rather than numerical data. The goal of Symbolic AI
is to build intelligent systems that can reason and think like
humans by representing and manipulating knowledge and
reasoning based on logical rules.
• Symbolic AI algorithms work by processing symbols, which
represent objects or concepts in the world, and their
relationships. The main approach in Symbolic AI is to use
logic-based programming, where rules and axioms are used
to make inferences and deductions.
How symbolic AI
works?
• Knowledge Representation: Uses
symbols and logical relationships to
define concepts and facts.
• Reasoning: Applies rules and
axioms to derive conclusions, making
it interpretable and precise.
• System Structure: Combines a
knowledge base (facts and rules) with
an inference engine to process and
analyze data.
 REAL WORLD APPLICATION

• Digital Assistants: Use Symbolic AI to understand

and respond to natural language inputs.

• Medical Diagnosis Systems: Recommend

treatments based on predefined rules and patient
symptoms.

• Autonomous Vehicles: Make decisions like

recognizing stop signs or traffic lights using logical
rules

• Computer Vision: Identify and interpret patterns

in images through symbolic reasoning.
 BENEFITS & LIMITATION

Benefits:
1.Interpretability: Easy to understand and trace how
conclusions are made.
2.Knowledge Representation: Stores knowledge in a
structured, formal manner.
3.Flexibility: Rules can be modified to adapt to various
domains.

Limitations:
4.Incomplete Knowledge: Struggles with undefined or
ambiguous information.
5.Scalability Issues: Large rule sets can be computationally
expensive.
6.Limited Adaptability: Cannot learn new knowledge in real-
time.
7.Ambiguity Handling: Struggles to reason effectively with
 Symbolic AI vs. Other AI Techniques
DATA
DEPENDENC Reasoning Applications
Y

• Symbolic AI uses • Symbolic AI is logic- • Symbolic AI excels in

predefined rules and based and transparent. structured domains.
does not need large • Machine Learning • Machine Learning works
datasets. identifies patterns and is well with unstructured
• Machine Learning often a black-box. and complex data.
depends on vast
amounts of training
data.
 Machine Learning:
Unlocking the
Power of Intelligent
Definition:
Systems
• Machine learning (ML) is a branch of artificial
intelligence (AI) that enables systems to learn
from data and improve their performance over
time. It's about training computers to recognize
patterns and make predictions based on past
experiences.
• Instead of relying on pre-defined rules, machine
learning algorithms learn from examples and
adapt to new data, allowing them to handle
complex and dynamic problems.
How Does Machine Learning Work?
Types of Machine Learning Algorithms
The Role of Big Data in Machine
Learning
Applications of Machine Learning
Real-World Applications of Machine
Learning
Challenges and Limitations of Machine
Learning
 INTRODUCTION TO DEEP
LEARNING

WHAT IS DEEP
LEARNING
• Deep Learning is a? subset of Machine Learning that mimics
the structure and function of the human brain using
Artificial Neural Networks (ANNs).
• It processes large volumes of data to identify patterns and
make predictions, enabling applications such as image
recognition, natural language processing, and self-driving
cars.
• It has become a cornerstone technology for solving
complex problems in various domains.

02
 HOW DEEP LEARNING
• Neurons: The basic computation units in a neural network,
WORKS ?
similar to brain neurons.
• Layers:
⚬ Input Layer: Receives raw data for processing.
⚬ Hidden Layers: Extract and process features through
multiple computations.
⚬ Output Layer: Produces the final result or prediction.
• Activation Functions: Decide which neurons to activate;
examples include ReLU and Sigmoid.
• Backpropagation: Optimizes weights by reducing errors using
gradient descent.

Workflow
1.Input data is fed into the network through the input layer.
2.Hidden layers process and extract meaningful features from the
data.
3.The model learns patterns through iterative optimization and
generates predictions in the output layer. 02
 ROLE OF DEEP LEARNING
IN AI
• Perception: Enables AI systems to interpret visual, audio, and text data using models
like CNNs and RNNs.
• Decision-Making: Helps AI make intelligent decisions by analyzing data patterns and
predictions.
• Language Understanding: Powers AI-based language models like ChatGPT and virtual
assistants for natural language interactions.
• Generative AI: Enables creation of new content, such as images, text, and music, using
models like GANs.

Integration in AI Systems
• Deep Learning provides the foundation for AI to learn autonomously from data,
improving over time without explicit programming.
• Combines with other AI techniques like knowledge graphs and symbolic reasoning for
robust solutions.
 TECHNIQUES OF DEEP
LEARNING
Convolutional Neural Networks
01 Specialized for image data,
(CNNs)
recognizing visual patterns like edges
and shapes.
Recurrent Neural Networks (RNNs):
02 Designed for sequential data like text or
time series, capturing temporal
dependencies.
Transformers:
03 Revolutionized NLP tasks (e.g., GPT, BERT) by
focusing on attention mechanisms for
understanding context.
Generative Adversarial Networks
04 Create synthetic data, often used in image
(GANs):
generation and enhancement.
 BENEFITS OF DEEP
LEARNING
• Automation: Handles repetitive and complex tasks with minimal human input, saving time and
resources.
• Accuracy: Achieves high precision in critical fields such as healthcare for diagnosing diseases.
• Scalability: Effectively manages large-scale data, making it suitable for modern data-driven
environments.
• Versatility: Powers diverse applications, from personalized recommendations to scientific
discoveries.

Real-World Applications
• Autonomous Vehicles: Enables self-driving cars to perceive and navigate the environment.
• Voice Assistants: Powers technologies like Alexa, Siri, and Google Assistant.
• Fraud Detection: Identifies unusual patterns in financial transactions.
• Recommendation Systems: Provides personalized content suggestions on platforms like Netflix
and Amazon.
 CHALLENGES WITH DEEP
LEARNING
• Data Requirements: Requires vast amounts of labeled data, which can be
expensive to obtain.
• Computational Costs: High-performance hardware and energy consumption are
needed for training.
• Black Box Nature: Lack of interpretability makes understanding decision-making
difficult.
• Overfitting: Models can sometimes learn noise in the data instead of genuine
patterns.
• Ethical Concerns: Bias in datasets can lead to unfair or discriminatory outcomes.
Addressing Challenges
• Data Augmentation: Enhances datasets by generating additional samples
through transformations.
• Explainable AI: Develops techniques to make models more interpretable
and transparent.
• Improved Hardware: Advances in GPUs, TPUs, and cloud computing
reduce computational barriers.
Thank
You!