Sunday , May 16 2021

Deep Learning breakthrough made by Rice University scientists, Ars Technica


      running hal on a beowulf cluster –


Rice University’s MACH training system scales further than previous approaching.


          -Dec (**************************************************************, ******************************************** 6: pm UTC



in an earlier deep learning

article, we talked about how inference workloads —The use of already-trained neural networks to analyze data — can run on fairly cheap hardware, but running the training workload that the neural network “learns” on is orders of magnitude more expensive.

In describing the scale of extreme classification problems, Medini refers to online shopping search queries, noting that ” there are easily more than 353 million products online. ” This is, if anything, conservative — one data companyclaimed (Amazon US alone sold) ************************************************ million separate products, with the entire company offering more than three billion products worldwide. Another companyreckonsthe US product count at 353 million. Medini continues, “a neural network that takes search input and predicts from 353 million outputs, or products, will typically end up with about 2, 000 parameters per product. So you multiply those, and the final layer of the neural network is billion billion … [and] I’m talking about a very, very dead simple neural network model. “

At this scale, a supercomputer would likely need terabytes of working memory just to store the model. The memory problem gets even worse when you bring GPUs into the picture. GPUs can process neural network workloads orders of magnitude faster than general purpose CPUs can, but each GPU has a relatively small amount of RAM — even the most expensive Nvidia Tesla GPUs only have GB of RAM. Medini says, “training such a model is prohibitive due to massive inter-GPU communication.”

Instead of training on the entire million outcomes — product purchases, in this example — Mach divides them into three “buckets,” each containing (***********************************************************. 3 million randomly selected results. Now, MACH creates another “world,” and in that world, the million results are again randomly sorted into three buckets. Crucially, the random sorting is separate in World One and World Two — they each have the same million outcomes, but their random distribution into buckets is different for each world.

With each world instantiated, a search is fed to both a “world one” classifier and a ” world two “classifier, with only three possible outcomes apiece. “What is this person thinking about?” asks Shrivastava. “The most probable class is something that is common between these two buckets.”

At this point, there are nine possible outcomes — three buckets in World One times three buckets in World Two. But MACH only needed to create six classes — World One’s three buckets(plus) ********************************* World Two’s three buckets — to model that nine-outcome search space. This advantage improves as more “worlds” are created; a three-world approach produces results from only nine created classes, a four-world setup gives results from 12 classes, and so forth. “I am paying a cost linearly, and I am getting an exponential improvement,” Shrivastava says.

Better yet, MACH lends itself better to distributed computing on smaller individual instances. The worlds “don’t even have to talk to one another,” Medini says. “In principle, you could train each [world] on a single GPU, which is something you could never do with a non-independent approach.” In the real world, the researchers applied MACH to a 90 million product Amazon training database, randomly sorting it into , 12 buckets in each of separate worlds. That reduced the required parameters in the model more than an order of magnitude — and according to Medini, training the model required both less time and less memory than some of the best reported training times on models with comparable parameters.

Of course, this wouldn’t be an Ars article on deep learning if we didn’t close It out with a cynical reminder about unintended consequences. The unspoken reality is that the neural network isn’t actually learning to show shoppers what they asked for. Instead, it’s learning how to turn queries intopurchases. The neural network does not know or care what the human was actually searching for; It just has an idea what that human is most likely to buy — and without sufficient oversight, systems trained to increase outcome probabilities this way can end up suggestingbaby products to women who’ve suffered miscarriages, or worse.

(***************************************************************************** (Read More) ************************************************** (**************************************

About admin

Check Also

Scientists say interstellar comet has led a very cold existence, Ars Technica

Scientists say interstellar comet has led a very cold existence, Ars Technica

CO Detector — It's very rich in carbon monoxide. Eric Berger - Apr 20, 2020 9:04 pm UTC Enlarge / An image of interstellar comet  2I/Borisov,Comets are essentially time capsules. Most formed during the early days of our Solar System, amid the disk of dust and gas around the Sun. The majority of this dust…

Leave a Reply

Your email address will not be published. Required fields are marked *