STEAM Sustainable Horizons

Tues, Apr. 16, 2024   |   2 – 3:30 pm   |   Virtual via ZOOM
PRE-REGISTER BY APR. 15.  Registrants will receive a meeting link.

Ethics in Machine Learning

The topic of artificial intelligence, or AI, has dominated the news since late 2022, when Generative AI became accessible through the software ChatGPT.  Its ability to create original content—and interact with humans—has provoked both excitement and fear worldwide.  Generative AI's ability to exploit human language has understandably raised serious ethical implications, but is it overshadowing other categories of AI that deserve our attention? Consider Machine Learning, the most widely utilized form of AI today.

Machine Learning is shaping the way we analyze data.  Through algorithms, Machine Learning (ML) identifies data patterns to predict outcomes or develop solutions. What makes this remarkable?  ML enables computers to make decisions, and perform complex tasks, without explicit programming. Whether you're applying for a car loan, or using model training to design a trial for a new pharmaceutical compound, ML can impact any arena where large amounts of data are used to form decisions.  

Important Questions for Discussion How do we ensure the right data is analyzed?  What is owed to the people who provided the data?  How do existing legal standards impact the use of Machine Learning?  Following the talk, you'll be invited to participate in a discussion related to the roles ML should play in research modeling, and the profound effects it may present.

This lecture will help you understand the implications of Machine Learning, data science, and professional ethics on:

    PRE-REGISTER BY APRIL 15      

The Speaker

Based in Albuquerque, Jay Lofstead, Ph.D. represents the Center for Computer Research, at Sandia National Laboratories. A principal member of technical staff within the Scalable System Software department, his work focuses on infrastructure to support all varieties of simulation, scientific, and engineering workflows, with emphasis on:

A co-founder of the IO500 storage list, he was a recipient of a 2013 R&D 100 award for his work on the ADIOS IO library. 

Throughout the year, he works extensively to support student mentoring and diversity programs for high school and college students. Dr. Lofstead graduated with a BS, MS, and Ph.D. in Computer Science from Georgia Institute of Technology.

Tues, Mar. 5, 2024   |   3 – 4:30 pm   |   Virtual via ZOOM
PRE-REGISTER BY MAR. 4.  Registrants will receive a meeting link.

Projecting the Future: Quantifying Uncertainties
in How the Climate Will Change

Record-breaking heat waves, drenching rains, severe floods, years-long droughts, devastating hurricanes, and extreme wildfires...events like these have become more frequent, and intense. Over the next century, societies and ecosystems are expected to endure more extreme weather, as Earth's climate evolves from continued fossil fuel consumption. Explore a burgeoning research field at the forefront of interdisciplinary science!

How certain are we in projections about the distant future?  The scientific question is not whether the climate will change, but how likely it will change by a given amount.  The greatest challenge?  Every prediction of a physical theory comes with error bars—arising from data limitations, and the need for approximations.

Learn how researchers are estimating possible climate futures.  This talk will discuss the general mathematical formalism of statistical uncertainty quantification, and how it is being applied within the U.S. Department of Energy, and other research institutions, to estimate the plausible range of possible climate futures.

Attending this lecture will improve your familiarity with research methods like:
•  Bayesian parameter estimation for nonlinear regression
•  Monte Carlo sampling
•  Sensitivity analysis
•  Model averaging
•  Surrogate modeling, or emulation of expensive computer simulations

You'll learn how these research methods can be applied to our understanding of:
•  Probabilistic predictions of global warming
•  Climate feedback simulations
•  Sea level rise from Antarctic ice sheet disintegration
•  Coastal flooding

   PRE-REGISTER BY MARCH 4      

Dr. Urban will discuss related academic paths, and career opportunities, if time allows.  This engagement is made possible by the Phoenix College CRLC, Computational Research Leadership Council.

The Speaker

Nathan Urban, Ph.D. serves as group leader of the Applied Mathematics group at Brookhaven National Laboratory's Computational Science Initiative (CSI).  He's previously conducted research at Los Alamos National Laboratory, Princeton University; and Penn State University, where he also earned a doctorate in Matter Physics.
 

Using Supercomputers to Understand the World

Tues, Apr. 11, 2023   |   3 – 4:30 pm   |   Virtual via ZOOM
PRE-REGISTER BY APR. 10.  Registrants will receive a meeting link.
 

Throughout history, humankind's insatiable curiosity about the surrounding world has been a constant source of innovation.  A relatively recent development, the supercomputer, has become integral to our investigations of the unknown, through the process of:

     •  Expressing what we want to know through a mathematical model; 
        then, devising methods for solving the model
     •  Translating those methods into scientific software to run on supercomputers
     •  Analyzing the results for insight

Learn about the development of FLASH, a dynamic software that evolved from its original intent—to investigate supernovas—to demystifying a diverse range of phenomena, from the flow of blood through the Pacemaker, to the impact of earthquakes on nuclear reactor rods.

Cinematic Visualization: Applying Hollywood to Science

Tues, Mar. 7, 2023   |   3 – 4:30 pm   |   Virtual via ZOOM
PRE-REGISTER BY MAR. 6.  Registrants will receive a meeting link.

Cinematic Visualization, also known as Impactful Visualization, uses techniques from movies and games to provide rich story-telling experiences.

As educators and professionals work to restore trust in science, Cinematic Visualization presents major possibilities for communicating ideas to the public. Its ability to quickly engage wide audiences, and hold that attention through advanced rendering, lighting, and animation, can be applied to scientific data.  Learn how these design methodologies, developed for gaming and movies, differ from traditional scientific visualization.

Attending this lecture will improve your familiarity with:
    •  Computer Science; Data Science
    •  Scientific Visualization, Ray Tracing, Monte Carlo Methods, Advanced Lighting and Shading
 

Mark Bolstad, Ph.D.

Dr. Bolstad has practiced in a wide range of scientific disciplines, from aerospace to environmental protection to biomedical and defense research. He's led three different visualization labs and HPC teams.

After more than 20 years in visualization and computer graphics, he returned to school to pursue a doctorate in Computer Science. Dr. Bolsted's thesis focused on Rendering Massive Models, asking the poignant question: how do designers render models that exceed 4 billion elements?