Neuromorphic Computing

(Introduction and Relations with AI)

A Seminar Report
Submitted by:
Shreyas Shridhar Limaye (2211848)
in partial fulfilment for the award of

Diploma
In
Computer Engineering
Under the guidance of:
Prof. Mrunali Jangam
at

Dr. Babasaheb Ambedkar Technological University

Lonere, Tal. Mangaon, Dist. Raigad, Maharashtra (INDIA) - 402103

2023-2024
 CERTIFICATE

This is to certify that the seminar titled “Neuromorphic Computing (History and
Characteristics)” is the work carried out by

Shreyas Shridhar Limaye (2211848)

The student of Diploma in Computer Engineering of Dr. Babasaheb Ambedkar

Technological University, Lonere during the academic year 2023-24, in partial
fulfilment of the requirements for the award of Diploma in Computer Engineering of
Dr. Babasaheb Ambedkar Technological University, Lonere.

Prof. M. A. Jangam Prof. S. M. Sabale

Guide Head

Department of Computer Engineering Department of Computer Engineering

External Examiners:

1.

2.

Place: Lonere

Date:
 ACKNOWLEDGEMENT

I take this opportunity to express my deepest sense of gratitude and sincere thanks to
everyone who helped me to complete this work successfully.

I would like to express my sincere thanks to Department of Computer Engineering,

Dr. Babasaheb Ambedkar Technological University, Lonere. For keeping the
Seminar as a part of syllabus as it helps us to improve our knowledge and skills in the
field of Computer Engineering.

I express my sincere thanks to Prof. S. M. Sabale, Head of Department, Computer

Engineering, for providing me with all the necessary facilities and support.

I would like to place on record my sincere gratitude to my seminar guide Prof. M. A.

Jangam, Assistant Professor, Computer Engineering, for the guidance and
mentorship throughout the course.

Shreyas Shridhar Limaye

(2211848)
 ABSTRACT

This report provides a comprehensive exploration of neuromorphic computing, a

groundbreaking paradigm inspired by the structure and functioning of the human brain.
As the field of artificial intelligence (AI) continues to advance, researchers are
increasingly turning to neuromorphic computing to overcome the limitations of
traditional computing architectures. The report begins with an in-depth introduction to
the fundamental concepts of neuromorphic computing, elucidating its underlying
principles and highlighting its departure from conventional computing models.

The report then delves into the intricate connections between neuromorphic computing
and AI, elucidating how the former serves as a bridge to enhance the capabilities and
efficiency of the latter. Through a synthesis of cutting-edge research and case studies,
the report illuminates the ways in which neuromorphic computing principles can be
applied to bolster AI algorithms, enabling them to perform complex tasks with greater
speed, energy efficiency, and adaptability.

Furthermore, the report investigates current trends and challenges within the field,
offering insights into ongoing research endeavours and potential future developments.
It explores the implications of neuromorphic computing for various AI applications,
such as machine learning, robotics, and natural language processing, and examines the
potential impact on industries ranging from healthcare to autonomous systems.

By providing a comprehensive overview of the symbiotic relationship between

neuromorphic computing and AI, this report aims to foster a deeper understanding of
the transformative potential of these technologies. As the synergy between
neuroscience-inspired computing models and artificial intelligence continues to
evolve, this report serves as a valuable resource for researchers, practitioners, and
stakeholders seeking to navigate the dynamic landscape of emerging technologies and
their impact on the future of computation and intelligent systems.
 TABLE OF CONTENT

Acknowledgment I

Abstract II

List of Figures III

1. von Neumann Architecture 9

2. Neural Networks based on von Neumann Architecture 10
3. Switching Between SIMD Threads 12
1) Introduction and Basics 1
1. What is Neuromorphic Computing? 1
2. Ways of using the Neuromorphic Computing 1
3. Neuromorphic Computing basic implementation 2
2) Advantages 4
1. Energy efficiency 4
2. Parallel processing 4
3. Real-time processing 4
4. Adaptability and Learning 5
5. Cognitive Computing 5
6. Fault Tolerance 5
7. Reduced memory bandwidth requirements 5
8. Brain Inspired Algorithm 5
9. Sensor Integration 6
10. Neuromorphic hardware acceleration 6
3) Disadvantages 7
1. Complexity of Design 7
2. Limited Understanding of the Brain 7
3. Programming Challenges 7
4. Scalability issues 8
5. Lack of Standardization 8
6. High Cost of Development 8
7. Ethical and Privacy Concerns 8
4) Relations with AI 9
1. von Neumann Architecture 9
2. Neural Networks based on von Neumann Architecture 10
3. Mitigating limitations in modern computing systems 11
5) Applications 15
1. Artificial Intelligence (AI) 15
2. Robotics 15
3. Neuromorphic Sensors 15
4. Cognitive Computing 15
5. Edge Computing 15
6. Medical Diagnosis 15
7. Neuromorphic Chips 16
8. Brain Machine Interfaces (BMI) 16
9. Autonomous Vehicles 16
10. Security Systems 16
6) Conclusion 17
7) Reference 19
List of Abbreviations IV

Sr. No. Abbreviation Full Form

01 AI Artificial Intelligence

02 IBM International Business Machine

03 AGI Artificial General Intelligence
04 VLSI Very Large-Scale Integration
05 IoT Internet of Things
06 RAM Random Access Memory
07 DRAM Dynamic Random Access Memory
08 ANN Artificial Neuron Network
09 CPU Central Processing Unit
10 GPU Graphical Processing Unit
11 SIMD Single Instruction Multiple Data
12 AMD Advanced Micro Devices
13 HBM High Bandwidth Memory
14 TPU Tensor Processing Unit
15 BMI Brain Machine Interfaces
Introduction and Basics December2023

Chapter 1: Introduction and Basics

1.1 What is Neuromorphic Computing?

Neuromorphic computing is a method of computer engineering in which elements of

a computer are modelled after systems in the human brain and nervous system. The
term refers to the design of both hardware and software computing elements.
Neuromorphic computing is sometimes referred to as neuromorphic engineering.

Neuromorphic engineers draw from several disciplines - including computer science,

biology, mathematics, electronic engineering, and physics -International Business
Machine (IBM) to create bio-inspired computer systems and hardware. Of the brain's
biological structures, neuromorphic architectures are most often modelled after neurons
and synapses. This is because neuroscientists consider neurons the fundamental units
of the brain.

Neurons use chemical and electronic impulses to send information between different
regions of the brain and the rest of the nervous system. Neurons use synapses to connect
to one another. Neurons and synapses are far more versatile, adaptable, and energy-
efficient information processors than traditional computer systems.

Neuromorphic computing is an emerging field of science with no real-world

applications yet. Various groups have research underway, including universities; the
U.S. military; and technology companies, such as Intel Labs and IBM.

1.2 Neuromorphic technology is expected to be used in the following

ways:

• Deep learning applications.

• Next generation semiconductors.
• Transistors.
• Accelerators.
• Autonomous systems, such as robotics, drones, self-driving cars, and artificial
intelligence (AI).
Diploma in Computer Engineering, DABTU – Lonere 1
Introduction and Basics December2023

1.3 Neuromorphic Computing basic implementation:

Some experts predict that neuromorphic processors could provide a way around the
limits of Moore's Law.

The effort to produce Artificial General Intelligence (AGI) also is driving neuromorphic
research. AGI refers to an AI computer that understands and learns like a human. By
replicating the human brain and nervous system, AGI could produce an artificial brain
with the same powers of cognition as a biological one. Such a brain could provide
insights into cognition and answer questions about consciousness.

Neuromorphic computing implements aspects of biological neural networks as

analogue or digital copies on electronic circuits. The goal of this approach is twofold:
Offering a tool for neuroscience to understand the dynamic processes of learning and
development in the brain and applying brain inspiration to generic cognitive computing.
Key advantages of neuromorphic computing compared to traditional approaches are
energy efficiency, execution speed, robustness against local failures and the ability to
learn.

Neuromorphic computing is an approach to computing that is inspired by the structure

and function of the human brain. A neuromorphic computer/chip is any device that uses
physical artificial neurons to do computations. In recent times, the term neuromorphic
has been used to describe analog, digital, mixed-mode analog/digital Very Large-Scale
Integration (VLSI), and software systems that implement models of neural systems (for
perception, motor control, or multisensory information). The implementation of
neuromorphic computing on the hardware level can be realized by oxide-based
memristors, spintronic memories, threshold switches, transistors, among others.

Training software-based neuromorphic systems of spiking networks can be achieved

using error backpropagation, e.g., using Python based frameworks such as snnTorch,
or using canonical learning rules from the biological learning literature, e.g., using
BindsNet.

A key aspect of neuromorphic engineering is understanding how the morphology of

individual neurons, circuits, applications, and overall architectures creates desirable

Diploma in Computer Engineering, DABTU – Lonere 2

Introduction and Basics December2023

computations, affects how information is represented, influences robustness to damage,

incorporates learning and development, adapts to local change (plasticity), and
facilitates evolutionary change. Neuromorphic engineering is an interdisciplinary
subject to design artificial neuron system, such as vision systems, head-eye systems,
auditory processors, and autonomous robots, whose physical architecture and design
principles are based on those of biological nervous systems. One of the first applications
for neuromorphic engineering was proposed by Carver Mead in the late 1980s.

Diploma in Computer Engineering, DABTU – Lonere 3

Advantages December2023

Chapter 2: Advantages

Neuromorphic computing is a branch of artificial intelligence (AI) that draws

inspiration from the architecture and functioning of the human brain. This approach
aims to develop computer systems that mimic the neurobiological structure and
functioning of the brain. Here are some advantages of neuromorphic computing:

1. Energy Efficiency:
Neuromorphic computing architectures are designed to be highly energy-
efficient, often mimicking the low-power characteristics of the human brain.
This can lead to significant energy savings compared to traditional computing
architectures, making them suitable for various applications, including mobile
devices and Internet of Things (IoT) devices.

2. Parallel Processing:
Neuromorphic chips are inherently parallel in nature, allowing them to process
multiple tasks simultaneously. This parallelism is inspired by the way the
human brain handles multiple tasks concurrently. This can lead to faster and
more efficient processing of complex information.

3. Real-time Processing:
The parallel and distributed nature of neuromorphic computing enables real-
time processing of data. This is particularly beneficial for applications where
low-latency responses are crucial, such as robotics, autonomous vehicles, and
certain types of sensor data processing.

Diploma in Computer Engineering, DABTU – Lonere 4

Advantages December2023

4. Adaptability and Learning:

Neuromorphic systems are designed to adapt and learn from their environment,
like the way the human brain learns. This makes them well-suited for tasks that
involve pattern recognition, machine learning, and artificial intelligence
applications where the system can improve its performance over time through
experience.

5. Cognitive Computing:
Neuromorphic computing aims to replicate cognitive functions, such as
perception, reasoning, and decision-making. This makes it suitable for
developing intelligent systems that can emulate human-like cognitive abilities,
leading to advancements in areas like natural language processing and computer
vision.

6. Fault Tolerance:
The distributed nature of neuromorphic systems can provide a level of fault
tolerance. If one part of the system fails, other components may continue to
function, allowing the system to maintain overall performance. This resilience
can be advantageous in critical applications where reliability is essential.

7. Reduced Memory Bandwidth Requirements:

Neuromorphic architectures often rely on local memory storage and processing,
reducing the need for extensive data transfer between components. This can lead
to lower memory bandwidth requirements, which is beneficial for applications
with limited memory resources.

8. Brain-Inspired Algorithms:
Neuromorphic computing allows the implementation of algorithms inspired by
the structure and functioning of the human brain. This can lead to more efficient
and effective solutions for tasks like pattern recognition, complex decision-
making, and information processing.

Diploma in Computer Engineering, DABTU – Lonere 5

Advantages December2023

9. Sensory Integration:
Neuromorphic systems can be designed to integrate information from various
sensors in a way that resembles human sensory integration. This capability is
valuable for applications in robotics, human-computer interaction, and
immersive technologies.

10. Neuromorphic Hardware Acceleration:

As neuromorphic computing gains popularity, dedicated hardware accelerators
are being developed to efficiently run neuromorphic algorithms. This can result
in improved performance and energy efficiency for neuromorphic applications.

While neuromorphic computing shows promise, it is still an evolving field, and

challenges such as scalability and standardization need to be addressed for widespread
adoption.

Diploma in Computer Engineering, DABTU – Lonere 6

Disadvantages December2023

Chapter 3: Disadvantages

Neuromorphic computing, which draws inspiration from the architecture and

functioning of the human brain, offers several advantages, such as energy efficiency
and parallel processing capabilities. However, like any technology, it also comes with
its share of disadvantages. Here are some potential drawbacks of neuromorphic
computing:

1. Complexity of Design:
Building neuromorphic hardware is a challenging task due to the complexity of
mimicking the intricate structure and functioning of the human brain. Designing
and manufacturing neuromorphic chips with the required level of complexity
can be resource-intensive and technically demanding.

2. Limited Understanding of the Brain:

Our understanding of the human brain is still incomplete. Neuromorphic
computing relies on our current knowledge of neuroscience, and there is much
that researchers still do not know about how the brain processes information.
This can lead to limitations in the accuracy and effectiveness of neuromorphic
systems.

3. Programming Challenges:
Developing software for neuromorphic hardware can be challenging.
Traditional programming paradigms may not be well-suited for neuromorphic
architectures, and new programming languages and tools are needed. This can
be a barrier to adoption for developers who are more familiar with conventional
programming methods.

Diploma in Computer Engineering, DABTU – Lonere 7

Disadvantages December2023

4. Scalability Issues:
As of now, neuromorphic systems are often limited in scale compared to
traditional computing architectures. Scaling up neuromorphic hardware to
handle larger and more complex tasks may pose technical challenges and may
not be as straightforward as increasing the number of conventional computing
components.

5. Lack of Standardization:
The field of neuromorphic computing is still evolving, and there is a lack of
standardization in terms of hardware and software interfaces. This lack of
standardization can hinder interoperability and make it difficult for developers
to create applications that work seamlessly across different neuromorphic
platforms.

6. High Cost of Development:

Research and development in neuromorphic computing require significant
financial investments. The specialized hardware and expertise needed for
designing and manufacturing neuromorphic chips can contribute to higher costs
compared to traditional computing solutions.

7. Ethical and Privacy Concerns:

As neuromorphic computing progresses, there may be ethical concerns related
to the potential development of advanced artificial intelligence systems. Issues
such as privacy, security, and the responsible use of AI must be addressed to
ensure the technology is deployed in a manner that aligns with societal values.

Despite these challenges, researchers and engineers are actively working to address
these issues and push the field of neuromorphic computing forward. As the technology
matures, it is likely that some of these disadvantages will be mitigated or overcome.

Diploma in Computer Engineering, DABTU – Lonere 8

Relations with AI December2023

Chapter 4: Relations with AI

The relationship between Neuromorphic Computing and Deep Learning is a

fascinating and rapidly evolving area of research, with the potential to revolutionize the
way we approach artificial intelligence (AI) and machine learning. Neuromorphic
computing, which is inspired by the structure and function of the human brain, offers a
promising alternative to traditional computing paradigms, while deep learning, a subset
of machine learning, has demonstrated remarkable success in tasks such as image and
speech recognition. By exploring the synergy between these two fields, researchers are
hoping to unlock new capabilities and efficiencies in AI systems, ultimately paving the
way for more advanced and human-like intelligence.

4.1 von Neumann Architecture:

Fig 1. von Neumann Architecture

A great majority of the latest AI systems are built by pairing von Neumann computers
and classical neural networks, dating back to the Rosenblatt's perceptron.

The von Neumann architecture separates the memory and the computations (Hennessy
and Patterson, 2017). The computations are executed in the form of programs, which
are sequences of machine instructions. Instructions are performed by a processor. A
processor instruction usually has several arguments that it takes from processor
registers (small but very fast memory cells located in the processor). At that, the

Diploma in Computer Engineering, DABTU – Lonere 9

Relations with AI December2023

instructions and most of the data are stored in the memory separately from the
processor. The processor and the memory are connected by a data bus by which the
processor receives instructions and data from the memory.

The first bottleneck of this architecture is the limited throughput of the data bus between
the memory and the processor. During the execution of a program, the data bus is loaded
mainly by the transfer of processing data from/to Random Access Memory (RAM).
Moreover, the maximum throughput of the data bus is much less than the speed at which
the processor can process data.

Another important limitation is the big difference in the speed of RAM and processor
registers (see Figure 1). This can cause latency and processor downtime while it fetches
data from the memory. This phenomenon is known as the von Neumann bottleneck.

It is also worth noting that this approach is energy-intensive. As argued in Horowitz

(2014), the energy needed for one operation of moving data along the bus can be 1,000
times more than the energy for one computing operation. For example, adding two 8-
bit integer numbers consumes ~0.03 pJ while reading from Dynamic Random Access
Memory (DRAM) consumes ~2.6 nJ.

4.2 Neural networks based on the von Neumann Architecture:

Fig 2. Neural Networks based on von Neumann Architecture

To solve cognitive problems on computers, there was developed the concept of

Artificial Neuron Network (ANN) based on the perceptron and the backpropagation

Diploma in Computer Engineering, DABTU – Lonere 10

Relations with AI December2023

(Rumelhart et al., 1986) method (Goodfellow et al., 2016). Perceptron is a simplified

mathematical model of an artificial neuron network, in which neurons compute a
weighted sum of their input signals and generate an output signal using an activation
function. The process of training the network by the backpropagation method consists
of modifying its weights toward decreasing the error (loss function).

Since most modern neural networks have a layered architecture, the most
computationally intensive operation in these networks is the operation of multiplying a
matrix by a vector y = Wx. To carry out this operation, it is first necessary to obtain data
from the memory, namely, m*n weights of W and n values of vector x. It should be
noted that m*n weights will be used once per matrix-vector multiplication while the
values from the vector x will be reused (Figure 2).

4.3 Mitigating limitations in modern computing systems:

First, let us look at the ways to mitigate the above limitations in modern AI systems.

4.3.1. Central processing unit:

Classically, the problem of memory latency was solved in the Central Processing Unit
(GPU) by using a complex multi-level cache (a high-speed data storage that speeds up
access to data) system (see Figure 1) (Hennessy and Patterson, 2017). In modern
processors, caches can occupy up to 40% of the chip area, providing tens of megabytes
of ultra-fast memory. Usually, the size of practically used neural networks does not
allow to fit all the weights into caches. Nevertheless, the latest processors with
vertically stacked caches technology (e.g., Advanced Micro Devices (AMD) Ryzen 7
5800X3D) can change this situation, providing larger caches.

Other traditional approaches to CPU optimization were speculative execution, branch

prediction, and others (Hennessy and Patterson, 2017). However, in matrix
multiplication, the order of computations is known in advance and does not require such
complex approaches, which makes these mechanisms useless. This means that in the
field of ANNs the CPU can only be suitable for computing small neural networks, and
it is unsuitable for modern large architectures hundreds of megabytes in size.

Diploma in Computer Engineering, DABTU – Lonere 11

Relations with AI December2023

4.3.2. Graphical processing unit:

The Graphical Processing Unit (GPU) is a massively parallel architecture. It consists of

many computing cores combined into streaming multiprocessors. This allows to
execute single instruction thread over multiple data streams Single Instruction Multiple
Data (SIMD) thread. Moreover, GPU executes several SIMD threads.

The GPU uses several strategies to deal with memory latency. The main one is to give
each streaming multiprocessor a large register file that saves the execution context for
many threads and provides quick switching between them. The computation scheduler
uses this feature and, when an instruction with a high latency is executed in one of the
instruction threads [Single Instruction Multiple Data (SIMD) thread 1], for example,
obtaining data from memory, it immediately switches to another instruction thread
(SIMD thread 2), and if a latency occurs in this additional thread, the scheduler begins
a new ready instruction thread (SIMD thread 3). After some time, data for the first
thread arrives and it also becomes ready for execution (see Figure 3). This enables
memory latency to be hidden (Hennessy and Patterson, 2017).

Fig 3. Switching between SIMD threads

However, what is crucial aside from latency is the memory throughput, i.e., the
maximum amount of data that can be received from the memory per unit time. To solve
this problem in GPUs, NVIDIA began adding High Bandwidth Memory (HBM)
starting with the P100 (2016), and this dramatically increased their performance

Diploma in Computer Engineering, DABTU – Lonere 12

Relations with AI December2023

compared to previous generations. In the Volta and Turing architectures, Nvidia

continued to increase the memory throughput, bringing it up to 1.5 TB/s in the A100
architecture (Krashinsky et al., 2020).

4.3.3. Tensor processing unit:

Google announced the first Tensor Processing Unit (TPU)—TPUv1 in 2016 (Jouppi et
al., 2018). It mitigates latency and low memory throughput by using so-called systolic
matrices and software-controlled memory instead of caches. The idea of systolic
computations is to create a large matrix (256 × 256 for TPUv1) of computing units.
Each unit stores a weight and performs two operations. First, it multiplies number x that
has come from the unit above by the weight and adds the result to the number that came
from the unit to the left. Second, it sends the number x received from above to the unit
below, and forwards the received sum to the unit to the right. This is how the TPU
performs matrix multiplications in a pipeline. With a sufficiently large batch size, it will
not have to constantly access the weights in memory, since the weights are stored in the
computing units themselves. At that, having a batch size larger than the width of the
systolic array, the TPU will be able to produce one result of multiplying a 256x256
matrix by a 256-long vector each cycle.

4.3.4. Neuromorphic approach:

Despite significant advances and market dominance of the hardware discussed above,
the AI systems based on them are still far from their biological counterparts. There is a
gap in the level of energy consumption, flexibility (the ability to solve many different
tasks), adaptability and scalability. However, such problems are not observed in the
mammalian brain. In this connection, it can be assumed that, as it already happened
with the principle of massive parallelism (Rumelhart and McClelland, 1988), the
implementation of crucial properties and principles of the brain operation could reduce
this gap. As a response to this need, a neuromorphic approach to the development of
AI systems attempts to use of the principles of organization and functioning of the brain
in computing systems.

The brain is an example of a fundamentally different, non von Neumann, computer.

Unlike classical neural networks executed in modern computing systems, in the brain:

Diploma in Computer Engineering, DABTU – Lonere 13

Relations with AI December2023

• The power consumption of the human brain is only tens of watts. This is in
orders of magnitude less than the consumption of modern AI systems.
• Neurons exchange information using discrete impulses, i.e., spikes.
• All events are transmitted and received asynchronously—there is no single
process that explicitly synchronizes the work of all neurons.
• Learning processes are local and network topologies are non-layered.
• There is no common memory that universal processors work with. Instead,
many simple computational cells (neurons) function in a self-organizing manner

Diploma in Computer Engineering, DABTU – Lonere 14

Applications December2023

Chapter 5: Applications

Neuromorphic computing finds applications in diverse fields, including:

1. Artificial Intelligence (AI):

Enables more efficient and brain-like processing for AI applications, improving
pattern recognition and learning capabilities.

2. Robotics:
Enhances the performance of robots by mimicking the human brain's ability to
process sensory information and make decisions in real-time.

3. Neuromorphic Sensors:
Develops sensors that emulate the human sensory system, improving the
perception abilities of machines in areas like vision and auditory processing.

4. Cognitive Computing:
Facilitates the development of systems that can understand, learn, and adapt to
complex environments, contributing to advancements in natural language
processing and contextual understanding.

5. Edge Computing:
Enables processing at the edge of networks, reducing the need for constant data
transfer to centralized servers and improving efficiency in applications like IoT
(Internet of Things).

6. Medical Diagnosis:
Enhances the analysis of complex medical data, aiding in the diagnosis of
diseases and the development of personalized treatment plans.

Diploma in Computer Engineering, DABTU – Lonere 15

Applications December2023

7. Neuromorphic Chips:
Advances in hardware design, such as neuromorphic chips, offer energy-
efficient solutions for complex computational tasks, which can be crucial for
mobile devices and energy-constrained environments.

8. Brain-Machine Interfaces (BMIs):

Improves the interaction between the human brain and external devices, leading
to advancements in prosthetics and assistive technologies.

9. Autonomous Vehicles:
Enhances the perception and decision-making abilities of autonomous vehicles,
making them more adept at navigating complex and dynamic environments.

10. Security Systems:

Improves the efficiency of biometric recognition systems and enhances the
capability to identify patterns in large datasets, contributing to better security
measures.

11. Neuromorphic computing's ability to mimic the brain's architecture and

functioning makes it a promising technology for various applications, especially
those requiring real-time processing, energy efficiency, and advanced pattern
recognition

Diploma in Computer Engineering, DABTU – Lonere 16

Conclusion December2023

Chapter 6: Conclusion

In conclusion, this report has provided an in-depth exploration of neuromorphic

computing and its intricate connections with artificial intelligence (AI). Neuromorphic
computing, inspired by the structure and functionality of the human brain, stands as a
promising paradigm for the development of advanced computational systems. This
emerging field leverages the principles of neuroscience to design hardware and
software that mimic the neural architecture, enabling machines to process information
in a manner like the human brain.

The report began with an introduction to neuromorphic computing, elucidating its

origins, principles, and fundamental objectives. It was established that the primary aim
of neuromorphic systems is to emulate the parallelism, plasticity, and efficiency
inherent in biological neural networks. Key components such as spiking neurons,
synapses, and neural networks were explored to illustrate the unique characteristics that
differentiate neuromorphic systems from traditional computing architectures.

The relationship between neuromorphic computing and AI was a central theme

throughout the report. It was demonstrated that the principles of neuromorphic
computing align closely with the aspirations of AI research, particularly in areas such
as pattern recognition, learning, and adaptive decision-making. Neuromorphic systems
hold the potential to address some of the limitations of conventional AI approaches,
offering enhanced efficiency, adaptability, and energy efficiency.

Furthermore, the report highlighted real-world applications where neuromorphic

computing and AI intersect, including robotics, sensory processing, and cognitive
computing. The symbiotic relationship between these two fields is expected to drive
innovation and breakthroughs in developing intelligent systems capable of complex,
human-like tasks.

In conclusion, the journey into neuromorphic computing has uncovered a realm of

possibilities for advancing AI capabilities. As researchers and engineers continue to
explore and refine neuromorphic architectures, the synergy between neuromorphic

Diploma in Computer Engineering, DABTU – Lonere 17

Conclusion December2023

computing and AI promises to reshape the landscape of intelligent computing, ushering

in an era where machines can learn, adapt, and perform tasks with unprecedented
efficiency and sophistication. The future holds exciting prospects for the fusion of
neuroscience-inspired computing and artificial intelligence, propelling us closer to the
realization of truly intelligent machines.

Diploma in Computer Engineering, DABTU – Lonere 18

References December2023

Chapter 7: References

1. Frontiers:

https://www.frontiersin.org/articles/10.3389/fnins.2022.959626

2. TS2 Space:

https://ts2.space/en/the-relationship-between-neuromorphic-computing-and-
deep-
learning/#:~:text=By%20combining%20the%20strengths%20of,advanced%20
and%20human%2Dlike%20intelligence.

3. Nature Journal:

https://www.nature.com/articles/s43588-021-00184-y

4. Tech Target:

https://www.techtarget.com/searchenterpriseai/definition/neuromorphic-
computing#:~:text=Neuromorphic%20computing%20is%20a%20method,hard
ware%20and%20software%20computing%20elements.

5. Wikipedia:

https://en.wikipedia.org/wiki/Neuromorphic_engineering

6. Intel:

https://www.intel.com/content/www/us/en/research/neuromorphic-
computing.html

7. Geeks for Geeks:

https://www.geeksforgeeks.org/neuromorphic-computing/

Diploma in Computer Engineering, DABTU – Lonere 19