The world is facing serious problems, such as natural disasters, endangered species and deforestation. Unsurprisingly, artificial intelligence could help solve some of these problems, but the people who need it most often don’t have access to powerful computers. As a result, IBM (IBM) has released smaller, free versions of its advanced AI models, called Prithvi and TerraMind, that can run on regular laptops and even smartphones. In fact, these versions are open source and designed to work offline, allowing scientists and everyday users to study and protect the planet in real time.
Elevate your investment strategy:
- Enjoy TipRanks Premium at 50% off! Benefit from powerful investing tools, advanced data, and insights from expert analysts to help you invest with confidence.
Small models are especially useful in places where internet or electricity is limited, such as disaster areas, remote areas, or even in outer space. Interestingly, TerraMind, built by a European space team, is the most successful model for Earth monitoring tasks. Meanwhile, Prithvi-EO 2.0, developed with NASA, helped track long-term environmental changes and even won an award. However, earlier versions of these tools were too large and required powerful hardware.
IBM solved this problem by using a technique called “frozen encoder,” which reduces models without losing much accuracy. This makes them up to 120 times smaller and perfect for mobile devices. Additionally, thanks to their compactness, these AI models can be used directly inside satellites, drones and even phones. For example, future satellites could be updated with specific tasks while in orbit. And because these templates are free, anyone with a basic laptop can now train or customize them.
Is IBM a Buy, Sell or Hold?
As for Wall Street, analysts have a Moderate Buy consensus rating on IBM shares based on six buys, six holds and one sell assigned over the past three months, as shown in the chart below. Additionally, IBM’s average price target of $287.25 per share implies that shares are trading close to fair value.
See more IBM analyst notes
Disclaimer & DisclosureReport a Problem
