Presentation skills and Technical seminars for engineers

PRESENTATION SKILLS &
TECHNICAL SEMINAR
Dr. James Adu Ansere
Senior Lecturer, EEED

Effective
Preparations
Clear
Communications
Engaging
Delivery
UNDERSTANDING PRESENTATION

BASIC PRESENTATION
SKILLS
1.Accuracy and Time: Deliver the presentation on
time.
2.Body Language: Proper training for body language
skills reduces performance anxiety, giving the
audience a sense of expertise about the presented
topic.
3. Voice Tone: Speak with passion and knowledge but
under-checked VOICE TONES

HOW TO DELIVER
IMPACTFUL PRESENTATIONS
1. Thorough Research:
1.Start by thoroughly researching your topic to ensure a deep understanding.
2.Gather relevant and up-to-date information from reputable sources.
2. Clear Structure:
1.Organize your presentation with a clear introduction, body, and conclusion.
2.Use a logical flow of ideas and transitions between sections.
3. Engaging Opening:
1.Begin with a compelling introduction to capture your audience's attention.
2.Consider using a relevant quote, anecdote, or a thought-provoking
question.

WHAT MAKES A POWERPOINT
PRESENTATION EFFECTIVE?
1. Prepared to WIN!
2. Designed Correctly
3. Practiced to Perfection
4. Deliver with Confidence and
Calmness
5. Free from Mistakes

Types
of
Presentation
Persuasive Presentations
Instructional Presentations
Inspirational Presentations
Informative Presentations

Presentation skills and Technical seminars for engineers

Summary
Preparing a Technical Session Presentation Slides
The PowerPoint presentation template can be found in the author kit. Please create
the following slides as a part of your presentation:
Slide 1 | Introductory slide
•Include your paper number and title.
•Include your author and company name and/or logo information.
Note: This should be the only slide to contain your company name/logo.
Slide 2 | Information slide
•Main content of your presentation in a One-Column or Two-Column Format.
•Enter Paper #, Paper Title, and Presenter Name at the bottom of the slide.
•Copy and insert this slide as many times as needed for your content.
Slide 3 | Acknowledgement, thank you, questions in a One-Column Format
This slide should be displayed during your Q&A time

Summary
Speaking Tips
•Do not read the presentation. Practice the presentation so you can speak from
bullet points. The text should be a cue for the presenter rather than a message for
the viewer.
•Give a brief overview at the start, present the information and wrap up by reviewing
important points.
•Use a wireless mouse/remote or pick up the wired mouse so you can move around
as you speak.
•If sound effects are used, wait until the sound has finished before speaking.
•Do not turn your back to the audience.
•Do not include judgmental remarks or opinions about the technical competence,
personal character, or motivations of any individual, company, or group. Any material
that does not meet these standards will be returned with a request for revision
before the conference.

TECHNICAL SEMINAR
 A technical seminar refers to a presentation or
lecture focused on a SPECIFIC TECHNICAL
TOPIC or subject within a particular field of
study or industry.

•Title: Artificial Intelligence in Cybersecurity
•Subtitle: Safeguarding the Digital Frontier
•Presenter's Name and Affiliation
INNOVATIVE EXAMPLES OF
TECHNICAL SEMINAR

By:
John Smith
Sunyani Technical University
Artificial Intelligence in
Cybersecurity
Title:

Chapter 1: Introduction
Chapter 2: The Cybersecurity Challenge
Chapter 3: Role of AI in Cybersecurity
Chapter 4: AI-Powered Threat Detection
Chapter 5: Adaptive Security Measures
PRESENTATION OUTLINE

 Brief definition of Artificial Intelligence (AI)
 Overview of the increasing significance of AI in the cybersecurity
landscape
 Highlight the growing threats and challenges in the digital realm
 Statistics or examples of recent cybersecurity incidents
Chapter 3: Role of AI in Cybersecurity
 Brief explanation of how AI is transforming cybersecurity
 Emphasize AI's ability to detect, prevent, and respond to cyber threats
DETAILED OUTLINE

 Brief definition of Artificial Intelligence (AI)
 Overview of the increasing significance of AI in the
cybersecurity landscape

 Highlight the growing threats and challenges in the digital realm
 Statistics or examples of recent cybersecurity incidents

Example of
Presentation
Final PhD Thesis Defense

Respondent: James Adu Ansere
ENERGY EFFICIENCY OPTIMIZATION IN
INTERNET OF THINGS NETWORKS
Pre-Defense of Doctoral
Dissertation
Supervisor: Prof. Guangjie Han
Date: 2020.09.00
Address: College of IoT, RM. 807

30
Presentation Structure
01 Introduction
02 Related Works
03 Proposed Method 1
07 Conclusion and Future Directions
References
(Chapter 1 – 5)
05 Proposed Methods

32
The Internet of Things
 The Internet of Things (IoT) is a network of
connected smart devices that uses sensors, actuator
and software to communicate and exchange data
through internet-enabled infrastructure.
 Statistical data shows that more than 50 billion
devices would be connected via IoT in 2020 and the
number is growing continuously

33
Applications of IoT Technology
IoT
Systems
M2M
Communications
Security and
Surveillance
Telemedicine
& Healthcare
Smart Cities
and Homes
Figure 1. 1: Applications of IoT Technologies

34
The Explosive Growth of IoT Device
Connectivity projected from 2010-2025
5.8
18.2
50.1
75.4
2010 2015 2020 2025
0
10
20
30
40
50
60
70
80
Connected
IoT
devices
(billions)
Year
• Source: IoT Statistics
Figure 1.3: Growth of IoT device connections

35
Research Motivations
High energy consumption is a major concern for wireless
networks and telecommunication network providers. Some of the
reasons include:
 Non-renewable power sources
 Environmental concerns
 Economic and Financial Impact
 Massive IoT connectivity
Key Motivation: how to design an energy efficient IoT networks
to minimize the energy consumption.

36
Research Questions
While efforts to reduce energy consumption have covered different
aspects of IoT, many important issues remain untouched. This thesis
aims to address the following questions:
1. What are the causes of energy consumption and wastage in IoT
networks?
2. Does the spectrum channel accessing affects energy
consumption?
3. How to design efficient resource allocation methods to
maximize energy efficiency performance in IoT networks?
4. What makes the energy resource optimization methods
different from the other existing methods?
5. How to evaluate and benchmark the proposed energy resource
optimization methods?

37
Research Objectives
1. To identify the various causes of energy consumption in
IoT networks (Chapter 1).
2. To review the state-of-the-art of energy consumption in
IoT networks (Chapter 1).
3. To design a resource allocation models to maximize
energy efficiency (Chapter 2-5).
4. To propose a novel resource allocation algorithms to
improve energy efficiency performance and facilitate the
practical implementation in IoT networks (Chapter 2-5).
5. To perform simulations and testing of the proposed
resource allocation algorithms and compared with the
existing models (Chapter 2-5).

39
Comparison with existing works
Ref. Proposed Method Contributions Difference
[1] Gu et al. [1] proposed a robust
cooperative spectrum sensing
scheduling optimization for CR-IoT
Minimize sensing time
overhead.
Under perfect
spectrum
sensing
[2] A relay opportunistic spectrum
sharing scheme [2] was presented to
improve channel access and sensing
time minimization.
Maximize energy
efficiency and spectrum
efficiency
[3] A dynamic compressive wide-band
spectrum sensing based on channel
energy reconstruction was proposed
for cognitive IoT networks
Reduce energy
consumption

Dynamic Spectrum Sensing for Cognitive Radio IoT Networks
Dynamic Spectrum Sensing for
Cognitive Radio IoT Networks
Proposed Method 1
03

Optimal Resource Allocation in Energy Efﬁcient Internet of Things Networks with Imperfect
CSI
Optimal Resource Allocation in
Energy Efficient IoT Networks with
Imperfect CSI
Proposed Method 2
04

，
Joint Power Allocation and User
Allocation Algorithm for Data
Transmission in Internet of Things
Networks
Proposed Method 3
05

，
Joint Resource Allocation
Optimization for IoT Networks with
Large-Scale BS Antennas
Proposed Method 4
06

， 07
Conclusion and Future
Directions

45
Conclusion
 In this thesis, we have examined resource allocation algorithms
designed for energy efficient IoT networks. We have addressed the
problem of energy efficiency maximization taking into account
transmit power allocation and different QoS requirements for users
under channel uncertainties.
 In method 1, we examined a dynamic spectrum sensing method to
enhance radio spectrum utilization and minimize energy
consumption for data transmission in IoT networks, where
secondary users opportunistically sensed available spectrum
channels without interference.
 From method 2, we investigated joint optimization of user
selection, power allocation and the number of activated BS
antennas to maximize energy efficiency in the IoT networks.

46
Future Directions
In the future research, the proposed methods in this thesis can be
implemented in the following areas:
1. Energy Eﬃciency in 5G-enabled IoT Networks
2. Green Cognitive Vehicular Networks
3. Energy Eﬃcient Power Control in 5G Femtocells and Interference
Management networks
4. Green 5G Heterogeneous Small-Cell Network
5. Optimizing Green Energy Utilization for IoT Networks with
Hybrid Energy Supplies

47
References
[1] J. Gu, W. S. Jeon, and J. M. Kim, “Proactive frequency-hopping dynamic spectrum
access against asynchronous interchannel spectrum sensing,” IEEE Transactions on
Vehicular Technology, vol. 62, no. 8, pp. 3614–3626, 2013.
[2] W. Zhang, R. K. Mallik, and K. B. Letaief, “Cooperative spectrum sensing optimization
in cognitive radio networks,” 2008 IEEE International Conference on Communications, pp.
3411–3415, 2008.
[3] R. Zhu, Y. Li, F. Gao, J. Wang, and X. Xu, “Relay opportunistic spectrum sharing based
on the full-duplex transceiver,” IEEE Transactions on Vehicular Technology, vol. 64, no. 12,
pp. 5789–5803, 2015
[4] T. Van Chien, E. Bj ̈ornson, and E. G. Larsson, “Joint power allocation and user
association optimization for massive MIMO systems,” IEEE Transactions on Wire-less
Communications, vol. 15, no. 9, pp. 6384–6399, 2016.
[5] Y. Lin, Y. Wang, C. Li, Y. Huang, and L. Yang, “Joint design of user association and
power allocation with proportional fairness in massive MIMO HetNet, IEEE Access vol.5,
pp. 6560-6569, 2017.
[6] D. Zhai, R. Zhang, L. Cai, B. Li, and Y. Jiang, “Energy-efficient user scheduling and
power allocation for NOMA-based wireless networks with massive IoT devices,” IEEE
Internet of Things Journal, vol. 5, no. 3, pp. 1857–1868, 2018

THANKS FOR YOUR ATTENTION
Respondent: James Adu Ansere
End of Presentation

Quantum Deep Reinforcement Learning for
Dynamic Resource Allocation in Mobile Edge
Computing-based IoT Systems
By:
James Adu Ansere

Outline
 Overview
 System model and Problem formulation
 Quantum-empowered Deep Reinforcement Learning Approach
 Performance evaluation
 Conclusion and Future Works

Outline
• Overview
• System model and Problem formulation
• Quantum-empowered Deep Reinforcement Learning Approach
• Performance evaluation
• Conclusion and Future Works

Massive IoT connectivity and Applications
Key design challenges:
1. High transmission latency
2. Significant energy consumption
3. Network dynamics and randomness
Fig. 2: IoT applications
Fig. 1: Forecast of IoT connectivity

System model
Fig. 3: The proposed quantum-enhanced MEC-enabled IoT network.

Real-time multi-user scenario
 We considered an uplink MEC scenario where a set of K IoT
devices will offload computational task to the Edge server.
 The time is divided into timeslot of duration τ
 A subset of IoT devices is scheduled within a
time-slot
Binary variable
 A scheduled device will transmit a task
Amount of bits
Computation
frequency
Latency required
 The task is transmitted using a wireless
channel
Transmitting Power Channel gain Achievable data rate

Optimization problem: Maximizing the Energy efficiency
 We seek to maximize the energy efficiency and formulate the joint
optimization problem as:
• Transmit power constraint
• QoS requirement constraint
• Association policy constraint
• Computational CPU frequency
• Minimum task processing latency
• Offloading binary decision

MDP-based Problem Formulation
 MDP is expressed in five-tuple
 We define the state, action, and reward function as follows:
   
, , max ,
( ) , ( ), , ( ),
l k l k k i k
S s p P f
     
  

1. System state space: At each discrete time step, the DRL agent observes the dynamic environment state by:
2. System action space: Each IoT selects an action randomly as:
   
, , ,
( ) ( ), ( ), , .
oc
k i l k k loc k
A p R
     

 
  
3. System reward function: The agent learns a policy control strategy t by mapping
state to action spaces. Hence, the expected cumulative reward function is given as:
task offloading policy χk,i (τ), transmission power control policy p′l,k(τ), computing resource allocation policy, Rkoc (τ), and local computing resource Dk (τ)
subchannel allocation sl,k (τ), transmit power allocation, available computing resources, and bandwidth allocation state Bk (τ)

State Transition Probability
Fig. 4: State-action pair transitions

Fig. 5: illustrates the data flow using PyTorch to train the RL agent

Iterative Process of Quantum Action Selections 1/2
 Four action selections: up, down, left, and right.
 The IoT has no prior knowledge about the environment and randomly chooses action.
Fig. 8: Deployment of IoT in a complex dynamic environment
IoT
Obstacle/hole
Pathway

Simulations results 1/2
 The system requires a relative small number of
episodes to identify optimal resources to be
allocated.
 The proposed algorithm learns faster to
guarantee an improved exploration and
exploitation trade-off as the number of
episodes increases.
 It uses Grover’s iteration to increase the
learning rates and obtain stochastic optimal
policy.
Fig. 10: IoT resource request per episodes
Fig. 11: Reward versus episode

Conclusion and Future Works
 This paper examined a stochastic computation task offloading to implement a proposed
quantum algorithm using quantum and Grover’s iteration to enhance the learning speed in
practical scenarios.
 Based on the quantum uncertainty, we employed quantum superposition principles, where
eigenstates and eigenfunctions were introduced for updating the multiple state-action pairs
to maximize the expected rewards and value functions.
 The simulation results demonstrated the feasibility of the proposed algorithm and its
exponential quantum learning speed to maximize the network processing efficiency.
 In the future, we will explore quantum algorithms with semantic information extraction,
caching systems, and federated quantum learning processes.

Presentation skills and Technical seminars for engineers

More Related Content

Similar to Presentation skills and Technical seminars for engineers (20)

Recently uploaded (20)

Presentation skills and Technical seminars for engineers