Unit 6

• Automated Planning, Classical Planning,

Algorithms for Classical Planning, Heuristics
for Planning, Hierarchical Planning, Planning
and Acting in Nondeterministic Domains,
Time, Schedules, and Resources, Analysis of
Planning Approaches, Limits of AI, Ethics of AI,
Future of AI, AI Components, AI Architectures.
• Automated planning is a way of making efficient and
effective decisions in complex systems by achieving
the goal of a decision-processing method that can
work in a constantly changing world.
• Key Components of Automated Planning
• Domain Model: Defines the environment's rules and
the actions' effects within that context. This model is
crucial for understanding how actions change the state
of the world.
• Planner: The algorithmic core that processes input
data (current state and goal) and outputs a plan, which
is a sequence of actions leading to the goal.
• Executor: Implements the plan, often capable of adjusting in real-
time to unforeseen changes in the environment.
• Monitor: Observes the execution and environment to provide
feedback to the planner, facilitating dynamic re-planning if
necessary.
• Techniques in Automated Planning
• Automated planning techniques can be broadly classified into two
categories: deterministic and non-deterministic.
• Deterministic planning assumes a predictable environment where
every action has a guaranteed outcome, suitable for static or highly
controlled environments.
• Non-deterministic (or probabilistic) planning, on the other hand,
deals with uncertainty in action outcomes, requiring more complex
algorithms like Markov decision processes (MDP) or Partially
Observable Markov Decision Processes (POMDP).
 Classical Planning in AI

• AI Classical planning is a key area in Artificial

Intelligence to find a sequence of actions that
will fulfil a specific goal from an exact
beginning point. This process creates methods
and algorithms that allow smart systems to
explore systematically various actions and
their outcomes which eventually lead to the
desired result occasionally from the starting
place.
• Here are some algorithms used in classical
planning in artificial intelligence (AI):
• Reinforcement learning
• A classic AI algorithm that uses deep neural
networks to learn policies for decision problems
•
• Forward planning
• An important tool for exploring unknown spaces
and states, and using the data to solve problems
•
• Bayesian optimization algorithm (BOA)
• An Estimation of Distribution Algorithm (EDA) that uses
Bayesian networks to evolve solution populations

• Other algorithms used in AI include: Unsupervised

learning algorithms and Temporal planning.

• Planning algorithms in AI are used to create strategies

or actions for intelligent agents, autonomous robots,
and unmanned vehicles. These solutions are complex
and must be discovered and optimized in
multidimensional space.
• Classical planning can be formulated as a
pathfinding problem over a directed graph. The
nodes of the graph represent the states of the
system or environment, and the edges represent
the state transitions that the actions make
possible.
•
• Some common AI planning approaches include
rule-based, goal-based, and utility-based. In rule-
based planning, the AI agent uses a set of rules to
make decisions.
 Heuristic for planning
• In artificial intelligence (AI), heuristic planning
methods are shortcuts that use rules of thumb
to find approximate solutions to
problems. Heuristics are based on cognitive
science, the study of how humans think and
solve problems.
• Heuristic planning methods are used when:
• An exact solution is not possible
•
• An exact solution would take too much time or
processing power
•
• A problem is too complex to tackle with a step-by-step
approach
•
• Limited resources are available
•
• Heuristic planning methods are used in many
areas, including:
• Route planning
• Game playing
• Machine learning
• Supply chain planning
• Supply chain logistics
• Product lifecycle management
• Enterprise asset managemen
• Heuristic planning methods prioritize speed
over accuracy, so they are often paired with
optimization techniques to improve
outcomes.
• AI applications that require precision must
carefully consider the trade-off between
speed and accuracy.
 Hierarchical Planning in AI

• Hierarchical Planning in Artificial Intelligence is a

problem-solving and decision-making technique
employed to reduce the computational expense
associated with planning.
• In artificial intelligence (AI), hierarchical planning is
a planning methodology that entails grouping tasks and
actions into several abstraction levels or hierarchies,
with higher-level jobs being broken down into a series
of lower-level tasks.
• It offers a method for effectively using a hierarchy of
goals and sub-goals to reason and plan in complex
contexts.
• Components of Hierarchical Planning
• Artificial intelligence (AI) hierarchical planning usually
entails the following essential elements:
• High-Level Goals: High-level goals provide the initial
direction for the planning process and guide the
decomposition of tasks into smaller sub-goals.
• Tasks: Tasks are actions that need to be performed to
accomplish the high-level goals.
• Sub-Goals: Sub-goals are intermediate objectives that
contribute to the accomplishment of higher-level goals.
Sub-goals are derived from decomposing high-level
goals into smaller, more manageable tasks.
• Hierarchical Structure: Hierarchical planning organizes tasks and
goals into a hierarchical structure, where higher-level goals are
decomposed into sub-goals, and sub-goals are further decomposed
until reaching primitive actions that can be directly executed.
• Task Dependencies and Constraints: Hierarchical planning
considers dependencies and constraints between tasks and sub-
goals. These dependencies determine the order in which tasks need
to be executed and any preconditions that must be satisfied before
a task can be performed.
• Plan Representation: Plans in hierarchical planning are represented
as hierarchical structures that capture the sequence of tasks and
sub-goals required to achieve the high-level goals. This
representation facilitates efficient plan generation, execution, and
monitoring.
• Plan Evaluation and Optimization:
Hierarchical planning involves evaluating and
optimizing plans to ensure they meet the
desired criteria, such as efficiency, feasibility,
and resource utilization. This may involve
iteratively refining the plan structure or
adjusting task priorities to improve
performance.
Hierarchical Planning Techniques in AI

• In section, we are going to discuss the hierarchical

planning techniques that are leveraged for organizing
and executing hierarchical structures:
• 1. Hierarchical Task Networks (HTNs)
• Hierarchical Task Networks are used for representing
and reasoning about hierarchical task decomposition.
HTNs consist of a set of tasks organized into a
hierarchy, where higher-level tasks are decomposed
into sequences of lower-level tasks. HTNs provide a
structured framework for planning and execution,
allowing for the efficient generation of plans that
satisfy complex goals and constraints.
• 2. Hierarchical Reinforcement Learning (HRL)
• Hierarchical Reinforcement Learning is extension of reinforcement
learning, it leverages hierarchical structures to facilitate learning and
decision-making in complex environments. In HRL, tasks are organized into
a hierarchy of sub-goals, and the agent learns policies for achieving these
sub-goals at different levels of abstraction. By learning hierarchies of
policies, HRL enables more efficient exploration and exploitation of the
environment, leading to faster learning and improved performance.
• 3. Hierarchical State Space Search
• Hierarchical state space search is a planning technique that involves
exploring the state space of a problem in a hierarchical manner. Instead of
directly exploring individual states, hierarchical state space search
organizes states into hierarchical structures, where higher-level states
represent abstract representations of the problem space. This hierarchical
exploration allows for more efficient search and pruning of the state
space, leading to faster convergence and improved scalability.
Time, Schedules, and Resources
 What is the Role of Planning in
Artificial Intelligence?
• At the heart of AI systems’ ability to perform tasks
autonomously is AI planning, which is critical in guiding
AI systems to make informed decisions, execute tasks,
and achieve goals efficiently.
• What is Planning in AI?
• In Artificial Intellingence (AI), planning refers to the
process of deciding a sequence of actions that a system
must take to achieve a specific goal. Just like humans
plan their daily tasks with a goal in mind, AI systems
use planning algorithms to break down a problem into
smaller, manageable tasks and determine the best
course of action.
• Planning is crucial for AI applications because it
allows machines to:
• Think ahead: AI systems evaluate multiple
possible outcomes before selecting the optimal
one.
• Adapt to changes: When faced with uncertain
environments, AI systems can revise their plans
based on new information.
• Act autonomously: Whether it’s an autonomous
robot or a smart assistant, AI planning enables
machines to perform tasks independently.
 Types of Planning in Artificial
Intelligence (AI)

• There are several types of planning approaches in AI,

each suited to different tasks and environments:
• 1. Classical Planning
• Classical planning is the traditional form of AI planning
and assumes a static and fully observable environment
where all actions are deterministic. The AI agent has
complete knowledge of the world and operates with a
fixed goal, attempting to find a sequence of actions
that leads from an initial state to a goal state.
• Probabilistic Planning
• Probabilistic planning is designed to handle environments that contain
uncertainty. The AI system must account for the fact that actions may have
different possible outcomes with associated probabilities. Probabilistic
planning often uses models like Markov Decision Processes
(MDPs) or Partially Observable Markov Decision Processes (POMDPs) to
manage this uncertainty.
• Key Features:
– Handling Uncertainty: Takes into account the likelihood of various outcomes
for each action.
– Probabilistic Models: Utilizes models like MDPs and POMDPs for decision-
making under uncertainty.
– Challenges: Requires more complex computations due to the need to consider
all possible action outcomes and their probabilities.
• Example: Autonomous vehicles navigating traffic, where road conditions
and other drivers’ behaviors are uncertain.
• Reactive Planning
• Reactive planning is suitable for highly dynamic and unpredictable
environments. Rather than following a pre-defined plan, the AI
agent continuously reacts to changes in the environment in real-
time. This approach doesn’t rely on creating a full plan ahead of
time but focuses on immediate responses to the current situation.
• Key Features:
– Real-time Adaptation: AI reacts dynamically to changes in the
environment.
– No Pre-computed Plan: Focuses on immediate actions rather than
long-term planning.
– Challenges: May lack long-term strategy or foresight, focusing only on
immediate responses.
• Example: A robot avoiding obstacles in an unknown environment or
video game AI adapting to player actions.
• 4. Hierarchical Task Network (HTN) Planning
• Hierarchical Task Network (HTN) Planning breaks down complex
tasks into simpler, smaller sub-tasks and creates a plan for each sub-
task. This hierarchical approach is especially useful for solving large-
scale problems where goals can be divided into manageable steps.
It often involves decomposing high-level tasks into sequences of
lower-level actions.
• Key Features:
– Hierarchical Decomposition: Complex tasks are broken down into
smaller, more manageable subtasks.
– Structured Approach: Provides a well-structured, layered approach to
planning.
– Challenges: Can become inefficient if the hierarchy of sub-goals is too
complex or poorly defined.
• Example: A robot assembling a product by breaking the task into
subtasks like gathering parts, assembling, and testing the product.
• Challenges in AI Planning
• Despite its importance, AI planning presents several challenges:
• Computational Complexity: Planning, especially in complex
environments, can be computationally expensive. Finding the
optimal sequence of actions in large, dynamic systems can take a
significant amount of processing power and time.
• Handling Uncertainty: In uncertain or unpredictable environments,
creating a plan that can handle every possible outcome is
challenging. Probabilistic and reactive planning methods aim to
address this, but it remains a difficult problem.
• Scalability: As the size of the problem or task increases, so does the
difficulty of planning. Scaling up planning algorithms to handle large
datasets or environments with numerous variables is a technical
hurdle.
 what are limits of ai
• Artificial intelligence (AI) has many limitations, including:
• Lack of common sense: AI systems can't understand the context of a
situation or make decisions based on intuition. For example, an AI might
not understand sarcasm or figurative language.
•
• Limited understanding of the world: AI systems are trained to perform
well in specific domains, but they don't have a deep understanding of the
world.
•
• Lack of creativity: AI can generate content, but it can't create original
thought.
•
• Ethical and moral decision-making: AI systems don't have inherent ethical
frameworks or moral reasoning.
•
• Vulnerability to adversarial attacks: AI systems can be misled if the
input data is intentionally manipulated.
•
• Data dependency: AI systems are dependent on the quality and
quantity of training data. If the data is biased or incomplete, the AI's
decisions may be incorrect or unfair.
•
• Resource intensiveness: Training AI models requires a lot of
computational power and energy.
•
• Limited transfer learning: AI models are good at the tasks they're
trained for, but it's hard to transfer their knowledge to new tasks.
•
• Mathematical paradox: Some computational problems can't be
solved with algorithms.
•
• Unknown unknowns: AI can't answer questions or fill gaps in
understanding that aren't known to exist.
• AI ethics is a set of moral principles that guide the development and
use of artificial intelligence (AI) to ensure it benefits society. AI
ethics considers a range of factors, including:
• Fairness: AI should be fair and avoid bias
•
• Privacy: AI should ensure the privacy of users and their data
•
• Transparency: AI should be transparent and explainable
•
• Accountability: AI should be accountable for its actions
•
• Environmental impact: AI should mitigate environmental risks
•
• Human rights: AI should protect human rights and dignity
•
• Diversity: AI should ensure diversity and inclusiveness
•

• AI ethics in technological advancements can help foster a

world with less bias and more fairness. Here’s what it is and
why it matters.
• As artificial intelligence (AI) becomes increasingly
important to society, experts in the field have identified a
need for ethical boundaries when it comes to creating and
implementing new AI tools. Although there's currently no
wide-scale governing body to write and enforce these rules,
many technology companies have adopted their own
version of AI ethics or an AI code of conduct.
• AI ethics are the moral principles that companies use to
guide responsible and fair development and use of AI. In
this article, we'll explore what ethics in AI are, why they
matter, and some challenges and benefits of developing an
AI code of conduct.
• What are AI ethics?
• AI ethics are the set of guiding principles that
stakeholders (from engineers to government officials)
use to ensure artificial intelligence technology is
developed and used responsibly. This means taking a
safe, secure, humane, and environmentally friendly
approach to AI.
• A strong AI code of ethics can include avoiding bias,
ensuring privacy of users and their data, and mitigating
environmental risks. Codes of ethics in companies and
government-led regulatory frameworks are two main
ways that AI ethics can be implemented. By covering
global and national ethical AI issues, and laying the
policy groundwork for ethical AI in companies, both
approaches help regulate AI technology.
• .
• More broadly, discussion around AI ethics has
progressed from being centered around academic
research and non-profit organizations. Today, big
tech companies like IBM, Google, and Meta have
assembled teams to tackle ethical issues that
arise from collecting massive amounts of data. At
the same time, government and
intergovernmental entities have begun to devise
regulations and ethics policy based on academic
research.
 Future of AI
• Artificial intelligence (AI) is expected to have a significant impact on society in the
future, with applications in many industries:
•
• Education: AI can personalize learning experiences for students of all ages, using
machine learning, natural language processing, and facial recognition.
•
• Healthcare: AI can help doctors and physician assistants make more accurate
diagnoses.
•
• Finance: AI can help banks and financial advisors engage with clients, monitor credit
scores, detect fraud, and more.
•
• Law: AI systems can work with small teams of humans to do the work of 10–20 lawyers,
more quickly and cost-effectively.
•
• Transportation: More autonomous vehicles are expected to be
used in the near future.
•
• Machine learning: Machine learning algorithms are already being
used in many applications, and are expected to become even more
powerful.
•
• Natural language processing (NLP): NLP is the ability of computers
to understand and react to human language, and is expected to
become even more advanced.
•
• AI is also expected to become more integrated into personal and
professional spheres, with user-friendly platforms that allow
nonexperts to use AI. However, AI also faces some challenges,
including the possibility of people becoming overdependent on it
 Components of AI
• 1. Learning
• Learning in the context of AI is similar to how
humans acquire knowledge but implemented
computationally. One fundamental aspect of
AI learning is the trial-and-error method. The
AI system attempts various solutions to a
problem and retains successful strategies in its
database for future use.
• 2. Reasoning and Decision Making
• AI analyzes information and makes decisions
through reasoning. This involves drawing
inferences from given situations, categorized
as inductive or deductive. Deductive
inferences involve providing guaranteed
conclusions, while inductive inferences deal
with situations where outcomes are not
certain.
• 3. Problem Solving
• AI’s problem-solving ability involves techniques
like planning, search, and optimization. Special-
purpose methods tailor solutions to specific
features of a given problem, while general-
purpose methods address a wide range of diverse
issues. Problem-solving in AI includes step-by-
step reduction of differences between the
current state and the goal state.
• Perception
• AI perceives its environment by gathering and
interpreting information through sense
organs, whether artificial or real. The system
internally processes this data to analyze
scenes, recognize objects, and understand
their relationships and features. Perception is
crucial for tasks like computer vision and
speech recognition.
• Language Processing
• Language processing in AI involves understanding
and generating natural language. Techniques like
natural language understanding, machine
translation, and text generation enable AI to
interact with language effectively. This allows
applications like chatbots, language translation
tools, and sentiment analysis to function
seamlessly.
•
• An AI architecture diagram is a visual representation of how the
components of an AI system interact and work together to achieve
specific outcomes.
•
• AI architecture diagrams can help illustrate how AI is used to
connect IoT devices and optimize their operations. They can also
help people understand the structure and organization of a system,
software, or project.
•
• AI systems can generate diagrams using prompts by interpreting
text, raw data, and source code. Some AI-powered tools that can
generate diagrams include:
•
• DiagramGPT: An AI-powered text-to-diagram generator that can
quickly create diagrams
•
• Eraser IO: A whiteboard for engineering teams that includes a
diagram generator and tools for editing diagrams
 Key Elements of AI Architecture

• AI architecture is adaptable to the needs of an organization.

Nevertheless, there are core elements commonly included in most
architectures.
• Purpose
• The primary objective of AI architecture is to align the technical
components of AI—such as machine learning models, data
pipelines, and analytics tools—with the business goals of the
organization.
• Frameworks
• Frameworks play a crucial role in AI architecture. They are the
guiding principles that help in the design, construction, and
deployment of AI systems. Popular frameworks for AI include
TensorFlow, PyTorch, and Scikit-learn, each offering different
capabilities and advantages.
• Components
• Data Layer: This involves the data sources and data storage
systems. It can include databases, data lakes, and cloud
storage solutions.
• Processing Layer: This is where data is transformed,
cleaned, and used for training machine learning models.
Technologies like Apache Spark can be used here.
• Algorithm Layer: This involves the actual machine learning
algorithms that are trained to make predictions or decisions
based on data.
• Interface Layer: This is the application interface where
users interact with the AI system. It could be a web
application, a mobile app, or even an API.