Mediaboss Marketing

Steffi Diem (middle) participating in a panel at the White House Summit on Developing a Bold Decadal Vision for Commercial Fusion Energy. Diem has joined a collaboration across multiple institutions that will use machine learning to better understand magnetic fusion energy.

Researchers at the University of WisconsinMadison are taking part in a new collaboration built on open-science principles that will use machine learning to advance our knowledge of promising sources of magnetic fusion energy.

The U.S. Department of Energy has selected the collaboration, led by researchers at the Massachusetts Institute of Technology, to receive nearly $5 million over three years. The teams including researchers at UWMadison, William & Mary, Auburn University and the HDF group (a non-profit data management technology organization) are tasked with creating a platform to publicly share data they glean from several unique fusion devices and optimize that data for analysis using artificial intelligence tools. Student researchers from each institutionwill also have an opportunity to participate ina subsidized summer program that will focus on applying data science and machine learning to fusion energy.

The data sources will include UWMadisons Pegasus-III experiment, which is centered around a fusion device known as a spherical tokamak. Pegasus-III is a new Department of Energy funded experiment that began operations in summer 2023 and represents the latest generation in a long-running set of tokamak experiments at UWMadison. A primary goal of the experiment is to study innovative ways to start up future fusion power plants.

Im incredibly excited to be a part of projects like this one as we continue to push innovation both in the analysis and development of experimental devices and diverse workforce development initiatives, says Steffi Diem, a professor of nuclear engineering and engineering physics, who leads the Pegasus-III experiment.

Diem is an emerging leader in the fusion research world. In 2022, she was invited to present at the White Houses Bold Decadal Vision for Commercial Fusion Energy that launched several efforts focused on commercializing fusion energy. In a field traditionally dominated by men, Diem is also one of four women leading the new collaboration.

UWMadison researchers are using the new Pegasus-III experiment to study innovative techniques for starting a plasma. Joel Hallberg

Throughout much of my career, I have often been one of the few women in the room, so it is great to be a part of a collaboration where four out of the five principal investigators are women, Diem says.

The collaboration is based around the principles of open science Diem and her colleagues will make the wealth of data coming from Pegasus-III and other fusion experiments more accessible and usable to others, particularly for machine learning platforms.

While this approach is designed to accelerate knowledge of magnetic fusion devices, its also aimed at providing a more accessible path into fusion research programs for students with wider skillsets and backgrounds, particularly in data sciences. Building a more diverse fusion workforce will be tantamount going forward, says Diem.

Fusion isnt just plasma physicists anymore, she says. As fusion moves out of the lab and toward the goal of providing clean energy to communities, it requires an interdisciplinary approach with engineers, data scientists, skilled technical staff, community members and more.

UWMadison is supporting a broader push to diversify the fusion field. Some of the student researchers who will be participating in the new collaboration are part of the student-led Solis group, which provides gender-inclusive support for students studying plasma physics on campus.

The new collaboration fits well with Diems other research, funded through the Wisconsin Alumni Research Foundation, focused on reimagining fusion energy system design. That work centers energy equity and environmental justice early in the design phase to support a just and equitable energy transition.

While there are still many challenges that lie ahead for fusion, the potential benefits are huge as we drive towards a cleaner, more sustainable, equitable and just future, says Diem.

See the original post here:
UWMadison part of effort to advance fusion energy with machine ... - University of Wisconsin-Madison

One of the worlds leaders in computational psychiatry will kick off the upcomingMaury Strauss Distinguished Public Lecture Seriesat the Fralin Biomedical Research Institute at VTC in September.

The public lectures bring innovators and thought leaders in science, medicine, and health from around the globe to the Health Sciences and Technology campus in Roanoke.

Leading the series with a discussion of machine learning and human thought is Read Montague, the Virginia Tech Carilion Vernon Mountcastle Research professor anddirector of theCenter for Human Neuroscience Researchat the Fralin Biomedical Research Institute at VTC.

Montagues research led to the development of the prediction error reward hypothesis among the most influential ideas at the basis for human decision-making in health and in neuropsychiatric disorders and recently to first-of-their-kind observations in the human brain of how the neurochemicals dopamine and serotonin shape peoples perceptions of the world around them.

He will share details of his data-driven neuroscience applications to machine learning to better identify and treat diseases of the brain at 5:30 p.m. on Sept. 28 at the institute.

Montague, who is working with clinicians and research centers worldwide to gather data on brain signaling, is also a professor in the department of physics at Virginia Techs College of Science.

Next in the series is J. Marshall Shepherd, who started his career as a meteorologist and became a leading international expert in weather and climate. He is an elected member of three of the nations influential scientific academies: the National Academy of Sciences, the National Academy of Engineering, and the American Academy of Arts and Sciences.

How is his work part of a series on health? The World Health Organization recognizes climate change as the single biggest health threat facing humanity. Shepherd will address the intersection of climate, risk and perception.

Bookending the series in May 2024 is Rick Woychik, director of the National Institute of Environmental Health Sciences at the National Institutes of Health. The molecular geneticist oversees federal funding for biomedical research related to environmental influences, including climate change, on human health and disease.

Other lectures in the series address Alzheimers disease, infant nutrition, dementia, COVID-19 and cardiovascular outcomes, and locomotor learning in children with brain injury.

We look forward to joining with members of the wider community to better understand these exciting new innovations and insights that are germane to health, said Michael Friedlander, Virginia Techs vice president for health sciences and technology and executive director of the Fralin Biomedical Research Institute. This is an incredible collection of speakers who represent some of the best thinking in science, medicine, and policy in the context of improving health. We are also proud that our own Read Montague is among them, and we look forward to sharing this research with the wider community.

The free public lectures are named for Maury Strauss, a Roanoke businessman and longtime community benefactor who recognized the value of welcoming leaders in science, medicine, and health to share their work. The 2023-24 series, which began in 2011, highlights the research institutes commitment to the community.

The full 2023-24 Maury Strauss Distinguished Public lectures include:

The public is invited to attend the lectures, which begin with a 5 p.m. reception. Presentations begin at 5:30 p.m. in 2 Riverside at the Fralin Biomedical Research Institute.All are free, in person, and open to the public. Community attendance is encouraged. To make the lectures accessible to a wider audience, most are streamed live via Zoom and archived.

In addition to the Maury Strauss Distinguished Public Lectures, the Fralin Biomedical Research Institute also hostsPioneers in Biomedical Research Seminars, theTimothy A. Johnson Medical Scholar Lecture Series, as well as other conferences, programs, lectures, and special events.

Originally posted here:
Machine learning and thought, climate impact on health, Alzheimer's ... - Virginia Tech

The field of drug development stands at a pivotal crossroads, where the convergence of technological advancements and medical innovation is transforming traditional paradigms. At the forefront of this transformation lies artificial intelligence (AI) and machine learning (ML), powerful tools that are revolutionizing the drug discovery and development processes. The seamless integration of AI/ML has the potential to accelerate research and enhance efficiency in a new era of personalized medicine.

Image credit: Tierney | stock.adobe.com

The FDA acknowledges the growing adoption of AI/ML across various stages of the drug development process and across diverse therapeutic domains. There has been a noticeable surge in the inclusion of AI/ML components in drug and biologic application submissions in recent years.

Moreover, these submissions encompass a broad spectrum of drug development activities, spanning from initial drug discovery and clinical investigations to post-market safety monitoring and advanced pharmaceutical manufacturing.1 In a recent reflection paper, the European Medicine Agency acknowledges the rapid evolution of AI and the need for a regulatory process to support the safe and effective development, regulation, and use of human and veterinary medicines.2

AI and ML tools possess the capability to proficiently aid in data acquisition, transformation, analysis, and interpretation throughout the lifecycle of medicinal products. Their utility spans various aspects, including substituting, minimizing, and improving the use of animal models in preclinical development through AI/ML modeling approaches. During clinical trials, AI/ML systems can assist in identifying patients based on specific disease traits or clinical factors, while also supporting data collection and analysis that will subsequently be provided to regulatory bodies as part of marketing authorization procedures.

AI/ML technologies offer unprecedented capabilities in deciphering complex biological data, predicting molecular interactions, and identifying potential drug candidates. These technologies empower researchers to analyze vast datasets with greater speed and precision than ever before. For example, AI algorithms can sift through enormous databases of chemical compounds to identify molecules with the desired properties, significantly expediting the early stages of drug discovery.

One of the critical challenges in drug development is the identification and validation of suitable drug targets. AI/ML algorithms can analyze genetic, genomic, and proteomic data to pinpoint potential disease targets. By recognizing patterns and relationships in biological information, AI can predict the likelihood of a target's efficacy, enabling researchers to make informed decisions before embarking on laborious and costly experimental processes.

The process of screening potential drug candidates involves evaluating their impact on biological systems. AI/ML models can predict the behavior of compounds within complex cellular environments, streamlining the selection of compounds for further testing. This predictive approach saves time and resources, as only the most promising candidates advance to the next stages of development.

AI/ML-driven computational simulations are transforming drug design by predicting the interaction between molecules and target proteins. These simulations aid in designing drugs with enhanced specificity, potency, and minimal adverse effects. Consequently, AI-guided rational drug design expedites the optimization of lead compounds, fostering precision medicine initiatives.

The utilization of AI/ML in clinical trials has immense potential to improve patient recruitment, predict patient responses, and optimize trial designs. These technologies can analyze patient data to identify potential participants, forecast patient outcomes, and tailor treatment regimens for individual subjects. This leads to more efficient trials, reduced costs, and improved success rates.

Although the integration of AI/MI technologies into drug development has the potential to revolutionize the field, it also comes with several inherent risks and challenges that must be carefully considered:

AI and ML are reshaping the drug development landscape, from target identification to clinical trial optimization. Their ability to analyze complex biological data, predict molecular interactions, and expedite decision-making has the potential to accelerate drug discovery, reduce costs, and improve patient outcomes.

As AI/ML continues to evolve, it will undoubtedly play an increasingly pivotal role in driving innovation and transforming the pharmaceutical industry, leading us toward a more efficient and personalized approach to drug development and health care. Although AI and ML hold immense promise in revolutionizing drug development, their adoption is not without risks.

Careful consideration of these challenges, along with robust validation, regulation, and transparent reporting, are essential to harness the benefits of AI/ML while mitigating potential pitfalls in advancing pharmaceutical innovation.

References

See the article here:
Revolutionizing Drug Development Through Artificial Intelligence ... - Pharmacy Times

Advanced Space LLC., a leading space tech solutions company, is pleased to announce that an Advanced Space-led team has been chosen to apply Machine Learning (ML) capabilities to detect, track and characterize space debris for the IARPA Space Debris Identification and Tracking (SINTRA) program.

Space debrisitems due to human activity in spacepresents a major hazard to space operations. Advanced Space and its teammates Orion Space Solutions and ExoAnalytic Solutions are applying advanced ML techniques to finding and identifying small debris (0.1-10 cm) under a new Space Debris Identification and Tracking (SINTRA) contract from Intelligence Advanced Research Projects Activity (IARPA).

Space debris is an exponentially growing problem that threatens all activity in space, which Congress is now recognizing as critical infrastructure, said Principal Investigator Nathan R. The well-known Kessler syndrome will inevitably make Earth orbit unusable unless we mitigate it, and the first step is developing the capability to maintain persistent knowledge of the debris population. Through our participation in the SINTRA program, our team aims to revolutionize the global space communitys knowledge of the space debris problem.

Currently, there are over 100 million objects greater than 1 mm orbiting the Earth; however, less than 1 percent of the debris that could cause mission-ending damage are currently tracked. The Advanced Space teams solutionthe Multi-source Extended-Range Mega-scale AI Debris (MERMAID) systemwill feature a sensing system to gather data; ground data processing incorporating ML models to observe, detect, and characterize debris below the threshold of traditional methods; and a catalog of this information. A key component of this solution is that the team will use ML methods to decrease the Signal-to-Noise-Ratio (SNR) required for detecting debris signatures in traditional optical and radar data.

Advanced Space CEO Bradley Cheetham said, Monitoring orbital debris is critical to the sustainable, exploration, development and settlement of space. We are proud of the work the team is doing to advance the state of the art by bringing scale and automation to this challenge.

ABOUT ADVANCED SPACE:

Advanced Space (https://advancedspace.com/) supports the sustainable exploration, development, and settlement of space through software and services that leverage unique subject matter expertise to improve the fundamentals of spaceflight. Advanced Space is dedicated to improving flight dynamics, technology development, and expedited turn-key missions to the Moon, Mars, and beyond.

Read the original:
Advanced Space-led Team Applying Machine Learning to Detect ... - Space Ref

What are machine learning models?

A machine learning model automates the process of identifying patterns and relationships that are hidden in data. It can use a combination of prelabeled or unlabeled data processed through various machine learning algorithms to determine the best fit for the problem to be solved.

Each machine learning algorithm represents a specific strategy to uncover patterns within a historical data set, according to Michael Shehab, principal and labs technology and innovation leader at PwC. The process of transforming machine learning algorithms into models consists of three components: representing the problem, identifying a specific task and providing feedback to guide the algorithm's quest for a solution. "The resulting model represents a function that has been learned or produced by the machine learning algorithm and is capable of mapping previously unseen examples to an accurate output," Shehab explained.

Selecting the type of model to use is a mixture of art and science. "There is no one-size-fits-all approach to understanding which model works for your organization," said Brian Steele, vice president of product management at customer analytics platform provider Gryphon.ai. Each model type will offer insights and results based on the data type and use cases. Additionally, the type and quality of the input data will drive the selection of certain types of models.

The machine learning field is rapidly evolving. When it comes to describing approaches like those used in generative AI applications, new techniques are blurring the old methods of classifying models.

There's no commonly accepted classification standard as new models are added daily, said Anantha Sekar, AI lead at Tata Consultancy Services. Still, the most common classifications of machine learning models include supervised, semi-supervised, unsupervised and reinforcement learning. These major types should all be considered along with the objective and learning approach being used, Sekar recommended.

A generative AI model, for example, may involve multiple training approaches deployed in succession. It may start with unsupervised learning on a large corpus of data followed by supervised learning to fine-tune the model and reinforcement learning to continuously tune results after deployment. "Discussing types of models is like discussing types of humans," Sekar noted. "Since each one is ultimately unique, the classifications are useful mainly for broad understanding purposes."

Data scientists will each develop their own approach to training machine learning models. Training generally starts with preparing the data, identifying the use case, selecting training algorithms and analyzing the results. Following is a set of best practices developed by Shehab for PwC:

In general, there is no one best machine learning model. "Different models work best for each problem or use case," Sekar said. Insights derived from experimenting with the data, he added, may lead to a different model. The patterns of data can also change over time. A model that works well in development may have to be replaced with a different model.

A specific model can be regarded as the best only for a specific use case or data set at a certain point in time, Sekar said. The use case can add more nuance. Some uses, for example, may require high accuracy while others demand higher confidence. It's also important to consider environmental constraints in model deployment, such as memory, power and performance requirements. Other use cases may have explainability requirements that could drive decisions toward a different type of model.

Data scientists also need to consider the operational aspects of models after deployment, called ModelOps, when prioritizing one type of model over another. These considerations may include how the raw data is transformed for processing, fine-tuning processes, prompt engineering and the need to mitigate AI hallucinations. "Choosing the best model for a given situation," Sekar advised, "is a complex task with many business and technical aspects to be considered."

The terms machine learning model and machine learning algorithm are sometimes conflated to mean the same thing. But from a data science perspective, they're very different. Machine learning algorithms are used in training machine learning models.

Machine learning algorithms are the brains of the models, Steele suggested. The algorithms contain code that's used to form predictions for the models. The data the algorithms are trained on often determines the types of outputs the models create. The data acts as a source of information for the algorithm to learn from, so the models can create understandable and relevant outputs.

Put another way, an algorithm is a set of procedures that describes how to do something, Sekar explained, and a machine learning model is a mathematical representation of a real-world problem trained on machine learning algorithms. "So, the machine learning model is a specific instance," he said, "while machine learning algorithms are a suite of procedures on how to train machine learning models."

The algorithm shapes and influences what the model does. The model considers the what of the problem, while the algorithm provides the how for getting the model to perform as desired. Data is the third relevant entity because the algorithm uses the training data to train the machine learning model. In practice, therefore, a machine learning outcome depends on the model, the algorithms and the training data.

Follow this link:
What are Machine Learning Models? Types and Examples - TechTarget