Connect with us

Hi, what are you looking for?


Hard science is the next domain that machine learning wants to capture.

Hard science

Hard sciences are also being revolutionized by machine learning

Particle physicists, the leaders in the hard sciences, have historically been early adopters of technology, if not its founders, from email to the Internet. It is therefore not surprising that scientists started teaching computer models to tag particles in the chaotic jets generated by collisions as early as 1997. Since then, these models have laboured along and have gotten better and better, though not everyone has been happy with this advancement. Over the past ten years, parallel to the greater deep-learning revolution, particle physicists have taught algorithms to solve previously unsolvable problems and take on completely new tasks.

“I felt really scared by machine learning.” He claims that at first, he believed that it imperiled his ability to characterize particle jets using human judgment. Thaler, however, has now come to accept it and has used machine learning to solve a number of issues in particle physics. Machine learning is a partner.

Jesse Thaler, a theoretical particle physicist at the Massachusetts Institute of Technology

The data used in particle physics is very different from the traditional data used in machine learning, to start. Convolutional neural networks (CNNs) are excellent at classifying images of everyday objects like trees, cats, and food, but they struggle to deal with particle collisions. The problem, according to particle physicist Javier Duarte of the University of California, San Diego, is that collision data from sources like the Large Hadron Collider isn’t by definition an image. Flashy depictions of LHC collisions could trick the detector into believing it is full. In actuality, the millions of inputs that aren’t actually registering a signal are represented by a white screen with a few black pixels. Although the resulting image is substandard due to the weakly supplied data, a more recent architecture known as graph neural networks can make good use of it (GNNs).

Innovation is needed to overcome additional particle physics problems.

“We’re not merely importing hammers to smash our nails.” We need to create new hammers because there are strange new types of nails. The enormous volume of data generated at the LHC—roughly one petabyte every second—is one peculiar nail. Only a limited amount of high-quality data is saved from this large volume. Researchers seek to teach a sharp-eyed algorithm to sort better than one that is hard coded in order to develop a better trigger system, which saves as much good data as possible while getting rid of low-quality data. The intention is not to connect the device or the experiment to the network and have it publish the articles without keeping them informed, according to Whiteson.

Daniel Whiteson, a particle physicist at the University of California, Irvine

He and his colleagues are attempting to have the algorithms deliver feedback in terms of what people can comprehend, but it’s possible that other individuals have communication duties as well.

But for such an algorithm to be effective, according to Duarte, it would have to run in a matter of microseconds. To speed up their algorithms and address these problems, particle physicists are pushing the bounds of machine techniques like pruning and quantization. Because the LHC has to store 600 petabytes of data over the course of the next five years of data collection, researchers are exploring for ways to compress the data.

It becomes essential to ask, “Are they practising physics or computer science?” when some particle physicists delve deeper into machine learning. This is a difficult subject. Coding already has a bad reputation and is sometimes derided as “not real physics,” and machine learning is also fraught with similar concerns. One worry is that machine learning may muddy the physics and turn analysis into a complex system of automated processes that is hard for people to understand. In contrast, Thaler contends, “We simply need to learn how to think a little bit more like a computer.” On the one hand, he says, “We’d want to have a machine learn to think more like a physicist.” We need to get better at speaking each other’s languages.

For More IT Articles Click Here

Click to comment

Leave a Reply

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

You May Also Like

Trending News

Multimodal generative AI is already here and now; it is no longer in the future. In recent months, generative AI models have become widely...


Quis autem vel eum iure reprehenderit qui in ea voluptate velit esse quam nihil molestiae consequatur, vel illum qui dolorem eum fugiat.

Featured News

Levi Ray & Shoup, Inc. (LRS) announced today that Shell plc (“Shell”) has selected the LRS® Enterprise Cloud Printing Service, a fully managed service provided by...

Featured News

The first Social Listening Solution from Digimind integrates two potent AI engines to give users a thorough view of their online presence. The combination...