Scribd
Upload
25 views2 pages

Speech Neuromorphic Computing

Uploaded by

soumiaoukhira66
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views2 pages

Speech Neuromorphic Computing

Uploaded by

soumiaoukhira66
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

speech Neuromorphic computing

problematic: Traditional computers


Traditional computers are built on the Von Neumann architecture, where tasks follow three
key steps: fetching instructions and data from memory, processing them in the CPU, and
storing results back in memory. This structured process happens billions of times per
second, making them excellent at handling well-defined tasks like calculations, running
software, and managing databases.

1. Slow Data Transfer: The CPU retrieves the image from memory, processes it step by step,
and sends the results back. This constant back-and-forth slows things down and uses a
lot of energy.
2. Struggles with Patterns: To recognize the cat, the computer follows strict rules (e.g., "look
for whiskers or ears"). If the photo is blurry or the angle is different, the rules might fail
because the computer isn’t naturally good at spotting patterns.
3. No Learning: If you show it a new photo with a different type of cat, it won’t recognize it
unless you rewrite the rules. It can’t adapt or learn from experience.
4. High Energy Use: Processing thousands of images like this requires massive energy,
making it costly and inefficient for large tasks like training AI models.

so how can neuromorphic adress these challenges?

Key finding: about challenges, and benefits


Neuromorphic computing is witnessing substantial growth in recent years. The market,
valued at USD 86.9 million in 2023, is projected to reach USD 9.36 billion by 2032, growing at
a compound annual growth rate (CAGR) of 68.27% from 2024 to 2032. This surge is driven by
neuromorphic systems' unique capabilities such as energy efficiency, adaptability, and
parallel processing, which make them particularly suited for AI-driven applications, robotics,
and autonomous vehicles​

IBM - United States

The image you see illustrates the growing market trend, with hardware, software, and services
all contributing to the overall expansion. It shows that, by 2032, neuromorphic computing
could represent a major sector of the tech industry, offering advanced solutions that could
revolutionize fields requiring fast, real-time decision-making, such as cybersecurity and
machine learning

Frontiers IBM - United States


Explain how does neuromorphic chips works (spiking neural
networks)
To understand how neuromorphic chips works: Imagine you're in a room full of people, and
you're trying to find your friend. In a traditional computer, it would be like checking each
person one by one, which could take a lot of time if there are many people. In contrast, a
neuromorphic system works like the brain: it looks at all the people at once, finding the right
person much faster because it’s processing everything in parallel.

In more technical terms, neuromorphic chips use spiking neurons, which only "fire" when
necessary. These spikes carry information, similar to how our brain cells communicate. The
chips also store data in the same way, processing and storing it simultaneously instead of
moving it back and forth between separate memory and processing units, which is what
causes bottlenecks in traditional computers.

These chips offer much lower energy consumption, as they only activate when needed,
making them highly efficient for applications like AI, robotics, and autonomous systems.

You might also like