Alien Elements

Alien Elements is an activity developed by Learning Undefeated to help students explore the field of data science and how it can be applied to the studies of chemistry.

Students will use the physical and chemical properties of alien elements to group and sort them into a table that conveys information about relevant trends and patterns, very similar to Periodic Table. Students will also be introduced to ideas of how data science can be predictive through the use of certain analytical tools. In this session, students will model a generative adversarial network (GAN), a type of machine learning that utilizes two neural networks. Students will act as the generators and interact with a discriminator, a computer program that has been trained to identify patterns in a given data set about molecular compounds. Students will use the feedback from the discriminator to propose possible compounds that could exist but are not present in the discriminator’s data set. This is an example of synthetic data generation, a tool that is becoming more and more important in science today.

While GANs have many uses in science, it’s more commonly known for its applications with images and videos, specifically with deepfakes. Students will view examples of deepfakes and discuss potential dangers of this technology as well as discuss how government agencies, like DARPA and the NSA, are working to detect and label deepfakes to prevent misinformation.

Learning Objectives

Standards Alignments & Connections

Next Generation Science Standards Connections

HS-PS1-1: Use the periodic table as a model to predict the relative properties of elements based on the patterns of electrons in the outermost energy level of atoms.

HS-PS1-2: Construct and revise an explanation for the outcome of a simple chemical reaction based on the outermost electron states of atoms, trends in the periodic table, and knowledge of the patterns of chemical properties.

Texas Essential Knowledge and Skills Connections

CHEM.5.A: Explain the use of the chemical and physical properties in the historical development of the Periodic Table.

CHEM.5.C: Interpret periodic trends, including atomic radius, electronegativity, and ionization energy, using the Periodic Table

Virginia Science Standards of Learning Connections

CH.2C: Trends within groups and periods including atomic radii, electronegativity, shielding effect, and ionization energy.

Activity Components

Pre-Laboratory Engagement

In the activity, students will interact with data in the Common Online Data Analysis Platform (CODAP). To reduce time needed to explain the mechanisms within CODAP, we recommend having students go through the following 60-minute exercise prior to the lab experience. CODAP is freely accessible and is browser based.

A Tool for Doing Data Science: CODAP

In this self-paced activity designed for anyone from high school students to adult learners, start by exploring a dataset about mammals using CODAP. What will you discover? Next, dive deep into a dataset from the National Health and Nutrition Examination Survey. Make graphs and learn to tell your own data story. Unlike other approaches to data science, you don’t need to use programming languages!

Conversation Starters:

What is data science?
What fields use data science?

Activity

DOWNLOAD STUDENT HANDOUT (above)

The inception of the periodic table of elements can be traced back to the early 19th century when chemists like Dmitri Mendeleev and Lothar Meyer independently recognized recurring patterns in the properties of elements. In 1869, Mendeleev famously organized the elements into a table based on their atomic weights, grouping them by similarities in chemical properties and leaving gaps for yet-to-be-discovered elements. This groundbreaking arrangement paved the way for the prediction of the properties of elements not yet observed, demonstrating the remarkable power of systematic organization in the field of chemistry.

This analysis of the known elements and their properties would today fall into the field of data science. Data science is an interdisciplinary field that uses statistics, computing, and scientific methods to extract knowledge and insights from noisy data. Data science has ushered in a new era of innovation in chemistry. Chemists are increasingly relying on data-driven approaches to accelerate research and discovery. Data science techniques such as machine learning and deep learning are employed to analyze vast datasets of chemical information, predict molecular properties, and optimize chemical reactions.

One AI tool data scientists are expanding the use of is Generative Adversarial Networks (GANs). GANs are a class of artificial neural networks consisting of two interconnected components: a generator and a discriminator, engaged in a competitive learning process. The generator creates synthetic data, such as images, audio, or text, with the aim of mimicking real data distributions. Simultaneously, the discriminator evaluates the generated data’s authenticity by distinguishing between real and synthetic examples.

While GANs have many uses in science when it comes to synthetic data, it’s more commonly known for its applications with images and videos, specifically with deepfakes. This AI-generated content manipulates existing content to convincingly depict individuals saying or doing things they never did. For example, replacing a president’s face with that of a famous actor. The U.S. Defense Advanced Research Projects Agency (DARPA) has created counter-AI programs to spot deepfakes. But the project continues as a key problem is that machine learning systems can be trained to outmaneuver these tools.

This activity utilizes the Concord Consortium’s Common Online Data Analysis Platform (CODAP) with a custom file. This file contains all of the alien element data and hides attributes that aren’t accessed at the beginning of the activity and can be accessed via a browser window on a computer or tablet.

This activity also uses a python script to act as the discriminator in the GAN. This file is designed to be opened by Python software (freely available) on a laptop or desktop computer. The file will download as a .txt file and must be resaved with the .py extension for it to work with Python.

Post-Laboratory Extension

Post-Laboratory Extension

GANs Extension

How to Detect Deepfakes and Avoid Disinformation

Use this PBS lesson plan to dive deeper into deepfakes and learn some of the ways to identify one.

Chemistry Extension

Chemistry Name Game

As students play this game, they will learn by compounds form as they do. They will also learn how to correctly name chemical compounds and write chemical formulas.

Additional Resources

On the position of helium and neon in the Periodic Table of Elements

Viewpoints on whether Helium should be in group 2 or 18 in the Periodic Table of Elements

Voices of DARPA Podcast, Episode 69: Demystifying Deepfakes

Dr. Wil Corvey, program manager for DARPA’s Semantic Forensics (SemaFor), discusses how the program goes beyond detection to delve deeper into understanding the intent behind manipulated media and how their team is creating tools available for today’s analysts.

Keep Going! Your Students May Also Enjoy...

AI Decision Trees

Explore how AI systems classify information for real-world applications like spam filtering and product recommendations.

Artificial Intelligence

Students will learn what it means for a computer to be intelligent and how AI software uses data to identify different things.

Making Medicine

Students will perform protein purification using column chromatography to gain a greater understanding of the biomanufacturing process of going from a cell to a protein to a product.