Neuromorphic computing is a new computing paradigm inspired by the workings of the human brain.
It involves the use of artificial neural networks that mimic the structure and function of biological neurons.
These networks are implemented in specialized hardware that is designed to optimize the performance of neural computations.
2. THE TECHNICAL SEMINAR ON
SEMINAR TOPIC : NEUROMORPHIC
COMPUTING
Submitted by
Komal S Pawooskar
USN:2VD19CS018
Guided by
Prof.Yasmeen Shaikh
Prof. Kartik K
3. INTRODUCTION
1.Neuromorphic computing is a new computing paradigm inspired by the
workings of the human brain.
2.It involves the use of artificial neural networks that mimic the structure
and function of biological neurons.
3.These networks are implemented in specialized hardware that is
designed to optimize the performance of neural computations.
4.Neuromorphic computing is particularly well-suited for tasks such as
pattern recognition, classification, and sensory processing.
5.Unlike traditional digital computing, which relies on fixed algorithms,
neuromorphic computing is more flexible and adaptable, making it ideal
for handling complex and dynamic data.
4. DEFINITION
Carver Mead introduced the term Neuromorphic Computing to
describe A new field in Technology whose design principles and
architecture are biologically Inspired.
Neuro: " to do with neurons i.e. neurally inspired"
Morphic: “ structure or form”
Emulates the functional structure of neurobiological systems
5. PRINCIPLES OF NEUROMORPHIC
TECHNOLOGY
• Build machines that have similar perception
capabilities as human perception
• Adaptable and self organizing
• Robust to changing environments
Realisation of future "THINKING machines
(intelligent and interactive systems)
7. NEUROMORPHIC ARCHITECTURE
• Computer architectures that are similar to biological brains;
computer architectures that implement artificial neural networks in
hardware.
• Functional units are composed of neurons, axons, synapses, and
dendrites.
• Synapses are connections between two neurons
Remembers previous state, updates to a new state, holds the
weight of the connection
• Axons and dendrites connect to many neurons/synapses, like long-
range buses.
8. IMPLEMENTING NEUROMORPHIC
COMPUTING
• Choose a neuromorphic hardware platform
• Develop a neural network model
• Map the neural network model onto the hardware platform
• Implement the neural network model
• Test and optimize the implementation
9. HOW THEY WORK
• Neural networks work based on a spike from a neuron.
• Neurons send a spike once a threshold voltage is meant.
• The voltage is added to or subtracted from neurons by
• what is passed through synapses.
• Signals are sent to a neuron through dendrites and sent out through the
axon.
• No "program" to execute
o Uses two phases of input: Learning and Operating
o Learning is done by changing what the synapses lets through i.e. the
weight of the synapse.
o Operation results on spikes from neurons based on specific inputs.
10. PROPERTIES
• Multistate behavior
• Sensitivity external stimuli
• Threshold behavior
• Fault tolerance
• Nonvolatility
• Temperature window
• Insensitivity to noise
• Low energy
• Compatibility
11. APPLICATION OF NEUROMORPHIC
COMPUTING
• Driverless cars
• Drones
• Robots
• Smart home devices
• Natural language, speech, and image processing
• Data Analytics
• Process optimization
12. CHALLENGES OF NEUROMORPHIC
COMPUTING
• Power Consumption: Neuromorphic computing systems require a lot
of power
• Hardware Limitations: Neuromorphic computing systems require
specialized hardware
• Programming Complexity: Neuromorphic computing systems require
specialized programming.
13. FUTURE OF NEUROMORPHIC SYSTEMS
• Implantable medical electronics
• Increased human-computer interaction
• Intelligent transportation systems
• Learning, pattern recognition
• Robot control(self motion estimation)
• Learning higher-order perceptual computation
14. CONCLUSION
Neuromorphic computing is an exciting field that has the
potential to revolutionize many fields. It aims to replicate the structure
and function of the human brain. Neuromorphic computing systems
can be used in robotics, medical diagnosis, and autonomous vehicles.
However, there are several challenges that researchers face in
neuromorphic computing. These challenges include power
consumption, hardware limitations, and programming complexity.
Despite these challenges, neuromorphic computing has the potential
to change the world in ways we can't even imagine.