Scientists have used the world’s fastest supercomputer to develop an AI tool that predicts tsunami flooding in almost real-time.

The system was created by researchers in Japan, one of the most seismically active areas in the world.

Tsunamis have occurred so frequently in the archipelago’s history that the Japanese invented a word to describe the phenomenon: “tsu” meaning harbour, and “nami” meaning wave.

In 2011, the country was hit by a devastating earthquake and tsunami that caused around 20,000 deaths and the worst nuclear distance since Chernobyl .

The tragic event exposed many shortcomings in Japan’s tsunami prediction method. The new AI model was designed to provide a more effective approach.

The system’s developers first high-res tsunami simulations on the Fugaku supercomputer, which generated 20,000 possible outcomes.

This data was used to train the AI to predict flooding conditions before a tsunami strikes land by analyzing offshore waveforms at the time of an earthquake.

Notably, the model can also run on ordinary PCs, which will make it easier to use in prediction systems,

The researchers tested the system on a theoretical tsunami in Tokyo Bay. They say it produced highly accurate predictions in a manner of seconds for a variety of different tsunami scenarios.

They envision using the model to more accurately and rapidly obtain detailed flooding forecast data for specific areas. This could provide vital insights into the effects of tsunamis on buildings and roads in coastal urban areas.

Ultimately, it could give disaster management teams a powerful tool for planning their disaster mitigation and evacuation strategies.

DARPA to test AI-controlled jets in live-fly dogfights after successful simulations

DARPA‘s mission to develop AI fighter jets has moved a step closer to take-off.

The military research agency’s algorithms took down an Air Force pilot in a virtual dogfight last year. In February, the Pentagon’s “mad science” unit tested how they’d perform as a team.

The battle pitted two friendly F-16s against a single enemy aircraft. Each fighter jet was equipped with a gun for short-range engagements and a missile for more distant targets.

Colonel Dan “Animal” Javorsek, program manager in DARPA’s Strategic Technology Office, said testing multiple weapons and aircraft introduced new dynamics to the trials:

DARPA is also assessing how much pilots trust the systems. Javorsek said they’ve installed sensors in a jet to measure physiological responses, such as where a pilot’s head is pointing and where their eyes are moving:

They now plan to test out AI on real-world aircraft. To do this, DARPA is creating an aero-performance model of an L-39 jet trainer, which the algorithm will use to make predictions and maneuver decisions.

Once the model is complete, the agency will begin modifying the aircraft so the algorithm can control it. The Pentagon plans to test them in live-fly dogfights in late 2023 and 2024.

Critics, however, have questioned the value of the trials. The rules-based nature of air-to-air combat is a good fit for algorithmic decisions and the “perfect information” supplied by the simulators isn’t available in the field.

Even if the AI worked equally well in reality, there’s been only one dogfight involving a US aircraft in the last 20 years.

A more pressing concern involves the rush to develop autonomous weapons. An AI arms race could encourage countries to cut corners on safety considerations, and even trigger an accidental war .

HT – The Drive

Greetings Humanoids! Did you know we have a newsletter all about AI? You can subscribe to it right here .

Google is testing Smart Compose on WhatsApp and Telegram

It looks Google’s exploring new ways to turn us all into robots. Users of Gboard beta have spotted Smart Compose popping up in Google Messages, Telegram, and WhatsApp, 9to5Google reports .

The feature’s been making writing suggestions as you type for a couple of years now, but never before in messaging apps. Google initially launched it for Gmail, where it now recommends closing messages, like “Have a great weekend!” or “Please leave me alone, you’re scaring me.” Well, maybe not the latter — but it should do.

It was later added to Google Docs for G Suite, and it now seems like it’s coming to messaging services too. So far, the feature has only been spotted in the snappily-named version 9.5.12.317844448 update to Gboard. But if the trials work, it might be meddling with all of our messages soon.

In fairness to Google, Smart Compose still gives humans the final say. If you like its suggestion, you can swipe right on the text to accept it. If you don’t, just keep typing your terrible prose.

And despite my concerns about becoming a cyborg, I value Smart Compose’s support when penning pointless emails. Its suggestions are also surprisingly smart, although that might just be a reflection of my generic writing.

Only a limited number of users have got the feature in messaging apps for now. If you want to join them, you could try signing up for Gboard beta .