Connect with us

AI Research

UND, partners lead Arctic research supercomputing initiative

Published

on









Research partners unite to develop predictive analytics platform for extreme cold weather operations

In the same way a helmsman guides a ship through turbulent waters, a team of researchers led by the University of North Dakota (UND) is developing an AI-powered system to navigate the growing complexities of the Arctic.

At the core of this initiative is a powerful Kubernetes (K8s) cluster — named after the Ancient Greek term for “helmsman” — which will drive the Arctic Knowledge-Based System (A-KBS), an advanced computational platform designed to support decision-making in cold weather environments using real-time data analytics and long-term forecasting. (Editor’s note: As Google explains, a Kubernetes cluster is a single, unified system composed of multiple computers that cooperate to execute applications.)

Why it matters

The Arctic is a challenging environment for operations.  Advancements in AI and machine learning are enabling the analysis of vast amounts of satellite, lidar, geological, and geospatial data to enhance situational awareness leading to more efficient and resilient operations.  Supported by a research contract from the US Army Corps of Engineers Research and Development Center, the A-KBS marks the first deployment of a Kubernetes cluster with supercomputing capabilities at UND.

“Challenges in the Arctic require supercomputing-level support for advanced situational awareness, real-time monitoring, and predictive analytics,” said Principal Investigator Timothy Pasch, professor of Communication (UND College of Arts and Sciences), associate director of the UND AI Research (AIR) Center and director of the UND ARCTIC Lab.

“As the Arctic increasingly becomes a global focal region, leveraging AI-driven modeling, Earth-scale data science, and remote sensing technology is essential for maintaining stability, sovereignty, and resilience.”

Co-leading the project is Naima Kaabouch, Chester Fritz Distinguished Professor of Computer Science and director of the UND AI Research (AIR) Center in the College of Engineering and Mines, with Aaron Bergstrom, senior computational scientist at UND’s Computational Research Center (Office of the Vice President for Research and Economic Development) managing the University’s High Performance Computing Cluster.

A computing powerhouse — accessible from anywhere

The A-KBS will handle an unprecedented amount of data including satellite imagery, LiDAR scans of Arctic infrastructure, and global-scale geospatial datasets as inputs into AI-driven decision support tools. For some of the most complex datasets, a Science Gateway developed in collaboration with the National Center for Ecological Analysis and Synthesis (NCEAS) led by Matt Jones together with Ian Nesbitt, will deploy the Globus Compute system to securely transmit data to the San Diego Supercomputing Center (under the supervision of Rick Wagner, chief technology officer) for large-scale processing with the EXPANSE Supercomputer, where Pasch has been allocated NSF ACCESS computational credits. https://www.sdsc.edu/systems/expanse/

“This advanced cyberinfrastructure enables geospatial forecasting extending 5, 10, or even 20 years — critical for Arctic infrastructure investment and national preparedness,” said Bergstrom. “These predictive analytics reduce costs and improve resilience in an extreme cold weather environment changing at extraordinary speed.”

A national collaboration for Arctic resilience

The UND-led initiative brings together researchers from across the country. A $100,000 subaward supports collaborative research with SUNY Stony Brook, where Dr. Dilip Gersappe, chair of Materials Science and Chemical Engineering, is developing new sustainable freeze-resisting materials, called hydrogels, aiming to mitigate extreme cold weather damage to infrastructure and helping to reduce the billion-dollar costs of Arctic degradation and repair.

Sherif Abdelaziz, professor of Engineering at Virginia Tech, is also a major collaborator on the A-KBS project, focusing on engineering resilient extreme cold weather infrastructure, including extreme cold temperature effects on soils. He will work with students and faculty partners leveraging UND computation to help analyze permafrost stability, soil-structure interactions and long-term Arctic infrastructure resilience.

UND team and research integration

At UND, graduate students Andrew Wilcox (MS, Earth Systems Science) Mya Geisinger and Sheridan Parker (MA/PhD, Communication), Mary Soaper and Hayden Patterson (BS/MS Geology), and Research Engineer II Aymane Ahajjam (PhD, Computer Science), alongside faculty Xudong Yu (Comm), DjeDje Kossu Zahui (CEM) and Marcus Algaier (Physics) collaborate closely with Pasch. Computer Science PhD students Saisri and Srinilla Pogalla, Software Engineers Stephen Miller and Walker McKee, CI Engineer Brad Traver, and System Administrator Robert Peterson collaborate on A-KBS computational development with Aaron Bergstrom.

Computer Science graduate students William Valentine, Meissam Shayeghmoradi, Kyle Foerster, David Jumar Bacallo, David Pappe, Howard Hottinger, and Nafiul Nawjis work closely with Kaabouch. Partners from UND’s Research Institute for Autonomous Systems (RIAS) including Scott Keane, Emmanuel Chukwuemeka, Zach Reeder and others are contributing to data collection efforts in Alaska, leveraging a combination of aerial and terrestrial sensors. Project Manager Kyle Buzek ensures integration and optimization of all research.

Pasch emphasized the importance of collaboration, citing UND’s Grand Challenges of Computational and Data Science and National Security and Space: “We are incredibly enthusiastic about this research,” he said. “The scope of the Kubernetes Cluster we are constructing has the speed and capability to provide real-time situational awareness and decision support — allowing us to predict and visualize Arctic geospatial change with unparalleled accuracy in one of the world’s most rapidly evolving frontiers.”



Source link

Continue Reading
Click to comment

Leave a Reply

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

AI Research

Artificial intelligence is a commodity, but understanding is a superpower

Published

on


As a developer and a human being, you want to push yourself as much as possible to incorporate the intention of things into your practice. By insisting on understanding a project’s intention and uniting it with your own understanding of the particulars of implementation, you become far more valuable. AI then makes it easier to magnify your intentions into automated activity.

We can speculate that AI will get better at this middle ground in the future, but it will never actually have intention. It will only ever move under human direction. Resist becoming just a connector or interpreter of intention to implementation. Keep on working to develop and contribute your own unique understanding. Implementation can be automated, but the unique qualities of understanding cannot.

Why LLMs will not replace higher-level languages

If you follow the hype cycle, it might seem that AI’s ability to mass produce code to meet requirements makes understanding the intention of that code less important. I’d say it makes it less necessary up front. There may even come a time when AI’s natural language interface is something like what fourth-generation languages are today. I can see a possible future where languages like JavaScript and Python are a layer below the AI interface, akin to how C is today. But if that is the analogy we’re using, then it seems clear we will always need people who deeply understand that layer, just as today we still need people who understand C, assembly machine code, and chip wafers.



Source link

Continue Reading

AI Research

LG AI Research Institute has released a new model of Exemplary Pass, a precision medical artificial ..

Published

on


LG AI Research Institute has released a new model of Exemplary Pass, a precision medical artificial intelligence (AI) model.

On the 9th, LG AI Research Institute unveiled ‘EXAONE Path 2.0’, a next-generation precision medical AI model.

Bio is one of the ABCs (AI, Bio, and Clean Tech) that LG Group Chairman Koo Kwang-mo cited as a future growth engine. Exempathy 2.0 is expected to play an important role in the group’s growth strategy.

Exempathy 2.0 is an upgrade of the 1.0 model released in January last year and the 1.5 model released last month, and has learned much higher quality data than the existing model. Pathologic tissue images (WSI) can precisely analyze and predict genetic variation and expression form, microscopic changes in human cells and tissues, and structural features.

Exemplary Pass 2.0 learned multiomics information such as DNA and RNA containing WSI and genetic information together. WSI is a high-resolution digital image taken during the pathologic diagnosis process of observing a patient’s tissue sample under a microscope. However, if AI simply analyzes this, there is a high possibility that a “characteristic collapse phenomenon” will occur, which will lead to poor prediction accuracy.

Exemplary Pass 2.0 applied a new technology that learns from small units to WSI, increasing the accuracy of predicting genetic mutations to 78.4%, the world’s highest level.

In addition, gene activity can be predicted only by image analysis without expensive dielectric tests.

Park Yong-min, leader of the AI business team at LG AI Research Institute, said, “If Exemplar Pass 2.0 is used, the genetic test time that took more than two weeks can be shortened to less than a minute, helping to secure golden time for treatment of cancer patients. If doctors and pharmaceutical companies use Exemplar Pass 2.0, they can identify target treatments in a short time.” The LG AI Research Institute also released additional models specializing in specific diseases such as lung and colon cancer.

LG AI Research Institute also announced a plan to cooperate with Vanderbilt University Medical Center in the United States. It is joined by a research team led by Korean scholar Hwang Tae-hyun, who leads the stomach cancer project of Cancer Moonshot, a cancer conquest project led by the U.S. government. Professor Hwang said, “Our goal is to create an AI platform that can help medical staff treat and treat patients in the actual medical field. The AI platform will not be just a diagnostic tool, but a game changer that innovates the entire process of new drug development.”

Starting with the field of cancer, LG AI Research Institute and Professor Hwang’s research team plan to expand the scope of research to rejection of transplants, immunology, and diabetes in the future. LG AI Research Institute is planning to introduce Exemite Pass 2.0 at the ‘LG AI Talk Concert 2025’ on the 22nd.

[Reporter Lee Deok-ju]



Source link

Continue Reading

AI Research

LG AI Research launches upgraded AI model Exaone Path 2.0 to improve cancer diagnosis, treatment options

Published

on


Bae Kyung-hoon, head of LG AI Research, speaks at a conference held at LG Science Park in Magok-dong, western Seoul, on July 19, 2023. [LG]

 
LG AI Research on Wednesday unveiled Exaone Path 2.0, its upgraded artificial intelligence (AI) model designed to enhance cancer diagnosis and drug development. The move aligns with LG Group Chairman Koo Kwang-mo’s vision of making AI and biotechnology core growth engines.
 
Exaone Path 2.0 learns from higher-quality data than the 1.0 version, which was launched in August last year, according to LG AI Research.
 
It can precisely analyze and predict not only genetic mutations and expression patterns but also subtle changes in human cells and tissues. The institute says this could enable earlier detection of cancers, forecast disease progression and support new drug discovery and personalized treatments.
 
A key breakthrough comes from new technology that trains the AI not only on small pathology image patches but also on whole-slide imaging. This pushed genetic mutation prediction accuracy to a globally leading level of 78.4 percent.
 
LG AI Research expects it will help secure the critical “golden hour” for cancer patients by cutting gene test times from over two weeks to under a minute. The institute also unveiled models tailored to specific diseases, including lung and colorectal cancers.
 

Dr. Hwang Tae-hyun, an expert in AI-driven research in precision oncology, immuno-oncology, cellular therapy and 3D/4D molecular modeling, is a professor at Vanderbilt University Medical Center. [LG]

Dr. Hwang Tae-hyun, an expert in AI-driven research in precision oncology, immuno-oncology, cellular therapy and 3D/4D molecular modeling, is a professor at Vanderbilt University Medical Center. [LG]

 
LG is bolstering this initiative through a partnership with Dr. Hwang Tae-hyun at Vanderbilt University Medical Center, a leading expert in biomedicine. Hwang, a prominent Korean scientist, heads a U.S. government-backed “Cancer Moonshot” project targeting gastric cancer. 
 
LG AI Research and Hwang’s team plan to jointly build a multimodal medical AI platform that utilizes real clinical tissue samples, pathology images and treatment data from cancer patients in clinical trials. They believe this will usher in an era of personalized, precision medicine.
 
Their collaboration also underscores Chairman Koo’s push to position AI and bio as technologies that transform customers’ lives. LG AI Research and Hwang’s team see this platform as the world’s first attempt to implement clinical AI in this way.
 
Starting with oncology, the team will expand into transplant rejection, immunology and diabetes research.
 
“Our goal isn’t just to develop another AI model. We want to create a platform that actually helps doctors treat patients in real clinical settings,” Hwang said. “This won’t just be a diagnostic tool — it has the potential to become a game changer that transforms the entire process of drug development.”
 

Performance level of LG AI Research's precision medical AI model EXAONE Path 2.0 [LG]

Performance level of LG AI Research’s precision medical AI model EXAONE Path 2.0 [LG]

Translated from the JoongAng Ilbo using generative AI and edited by Korea JoongAng Daily staff.
BY NA SANG-HYEON [[email protected]]





Source link

Continue Reading

Trending