Nvidia's new AI could make it easier to find a Black Friday deal and an RTX 3080 - Technology News

Breaking News

Nvidia's new AI could make it easier to find a Black Friday deal and an RTX 3080

It sounds like Nvidia has had enough of pandemic stock shortages as it has just announced new efforts to leverage its AI in routing goods from factories to stores and homes - just in time for this year's Black Friday deals.

Nvidia’s newly debuted ReOpt AI software combines the company’s pre-existing machine learning software with pre-existing data and algorithms specific to local regions to determine the most optimized route delivery vehicles should travel.

As we head into Black Friday 2021 - which already looks set to be impacted by stock shortages that have been keeping the latest and greatest gadgets out of our hands for months - Nvidia suggests its new tool could be a blessing for both consumers and companies. Manufacturers and stores can ensure supply chains run as smoothly as possible and that stock is where it needs to be, while customers can relax a little more knowing that the top deals won’t run out quite as quickly as they might otherwise.

Given how near we are to the sales event this year (early deals are already starting to appear) it may be a little late for Nvidia to support Black Friday this year. But Nvidia’s tool could be used to help shoppers secure last-minute holiday gifts in December and take advantage of the January sales we typically see in the New Year.

That said, some companies have already adopted the system; Domino's Pizza has been using Nvidia’s ReOpt AI to deliver orders - so if you feel like your takeaway arrived a little quicker than it has before you can thank the graphics card manufacturer.


Analysis: rendering our future

The computer science nerds among you will instantly recognize this as a company’s latest efforts to tackle the well-known traveling salesman problem - where the goal is to be able to find the most efficient route between multiple different points. While it is fairly trivial for a system with few destinations, as you increase the situation’s complexity and add different routes and modes of transport between locations it becomes something a human would struggle to work out quickly.

For a computer, the problem is much less of a challenge, and Nvidia’s graphics cards experience will be a huge benefit to its efforts here. 

A GPU renders a scene by completing lots of very simple calculations. The difficulty of this task and the need for very high-end components arises from there being a ton of these calculations to perform at once - especially as you introduce higher-quality graphics and effects like ray tracing.

In much the same way, measuring the efficiency of any one route from point A to B isn’t difficult, but the traveling salesman problem exists because you have to measure many thousands of routes between many thousands of points. If you’re after the most useful data you also have to do this as near to real-time as possible - much like how a GPU has to render an image in real-time as the game progresses and reacts to the player’s input.

This property of GPUs is also why they are so sought after by cryptocurrency miners - mining requires you to perform lots of very simple steps. That's another reason why you haven't been able to find the RTX 3080 you so desperately want.

If you aren’t bothered about logistics, though (in fairness it is a pretty bland subject) you should still keep an eye on Nvidia’s AI. The travelling salesman problem also has applications in DNA sequencing, and Nvidia has announced that a drug discovery start-up (Entos) will use its machine learning software to find next-generation treatments that work better than what we have currently.

Similarly, Nvidia has explained that its AI could also help scientists better track the effects of climate change, and make more accurate models to understand what weather disasters we might see so that communities have more time to prepare for what lies ahead.



from TechRadar - All the latest technology news https://ift.tt/3wHDQYx

No comments