6th March 2024 (11 Topics)
1
Mar
2
Mar
4
Mar
5
Mar
6
Mar
7
Mar
8
Mar
9
Mar
11
Mar
12
Mar
13
Mar
14
Mar
15
Mar
16
Mar
18
Mar
19
Mar
20
Mar
21
Mar
22
Mar
23
Mar
25
Mar
26
Mar
27
Mar
28
Mar
29
Mar
30
Mar
AI’s Carbon Footprint
Context
The integration of artificial intelligence (AI) in addressing the climate crisis poses a paradox: while AI holds promise for solving environmental challenges, the energy demands of AI models contribute to carbon emissions, particularly through the infrastructure supporting data centers.
1: Dimension: Scope of the problem
- Source of emission: The emissions come from the infrastructure associated with AI, such as building and running the data centres that handle the large amounts of information required to sustain these systems.
- To put things in perspective, training GPT-3 (the precursor AI system to the current ChatGPT) generated 502 metric tonnes of carbon, which is equivalent to driving 112 petrol powered cars for a year.
- GPT-3 further emits 8.4 tonnes of CO? annually due to inference.
2: Dimension- Technological approach to reduce emission
- Different technological approaches to build AI systems could help reduce its carbon footprint. Two technologies in particular hold promise for doing this: spiking neural networks and lifelong learning.
- Spiking neural networks (SNNs) and lifelong learning (L2), have the potential to lower AI’s ever-increasing carbon footprint, with SNNs acting as an energy-efficient alternative to Artificial neural networks (ANN).
- Artificial neural networks (ANNs), which learn patterns from data to make predictions, require significant computing power due to their reliance on decimal numbers, leading to increased energy consumption as networks grow larger and more complex, mirroring the brain's structure with billions of interconnected neurons.
- L2 is a set of algorithms aimed at minimizing forgetting in sequentially trained artificial neural networks (ANNs), allowing models to learn new tasks without losing previous knowledge, thereby reducing the need for energy-intensive retraining from scratch.
- Advances in quantum computing could revolutionize training and inference processes in ANNs and SNNs, potentially offering energy-efficient solutions for AI on a much larger scale
More Articles
Courses
Study
Notes
Home
Resources
More
