As Artificial Intelligence (AI) expands throughout every crevice of society, the opportunities for investing in this nascent industry become more accessible and diversified. But this rapid proliferation also comes with a steep learning curve and a lot of noise to sift through. We’ve compiled a list of key AI terms to know so investors can make more informed decisions. After all, we humans, like AI models, are only as powerful as the information we’re given.
Generative AI1 is perhaps the most well-known type of AI after ChatGPT made waves in late 2022. This subset of AI generates original content, like text, images, videos, and music and is fed by massive amounts of data combed from the internet. Large Language Models (LLMs)1 form the backbone of these programs by reacting to human language with appropriate responses, whether original or pre-programmed.
Similarly, Natural Language Processing (NLP)1 is a subset of AI that uses algorithms and computational models to sort through human and computer language for meaning and context. NLP is used to assess brand sentiment, inform predictive search engines, and power voice assistants like Siri and Alexa.
Machine learning1 can be similar to Generative AI, but does not create original content. ML is fed data in much the same way with the goal of making recommendations or predictions using algorithms. ML relies heavily on humans optimizing the models and teaching them what to do.
A Neural Network2 is a more autonomous version of machine learning whereby a program modeled after the human brain learns from itself instead of from humans in a supervised environment. Deep Learning3 is a sophisticated type of neural network with three or more layers and is used in many voice-enabled technologies as well as self-driving cars.
Self-driving cars also rely on Computer Vision1, which is a computer’s ability to assess images, objects, and videos and then offer recommendations. Computer vision requires a lot of data on which humans train it how to respond to different visual cues.
Digital Twins1 are virtual replications of real-world objects and spaces in which AI/ML are deployed to make decisions or predictions. These are used to model car manufacturing facilities, mimic an in-person retail experience, and test products.
All of the aforementioned forms of AI rely on two major components: computing power and data. Compute2 refers to the computational resources required for AI, such as how fast and how much data can be processed. Compute has doubled every 3.4 months since 2012 and may soon face limitations depending on innovations in hardware.
Training data1 is the data used to train the AI model to do specific tasks, whereas test data1 is data that was not used to train the AI model and therefore tests how accurate the model is on its own once training is complete. Data is arguably the most important component of AI. The type of data, how it is sourced, how it is labeled, who has permission to use it, and what biases are inherent within it all have significant effects on the end result.
The TrueShares Technology, AI, and Deep Learning ETF (LRNZ) provides exposure to all kinds of AI by focusing not on singular applications of AI, but instead on the supporting technologies that make every type of AI possible. From Nvidia’s specialized GPUs and world-class software (Jetson) to Snowflake’s cloud-based data storage and analytics to Crowdstrike’s cybersecurity technology, LRNZ is focused on a concentrated portfolio of 20-30 companies. Each company is thoughtfully selected and monitored by our portfolio manager with expertise in the technology sector. For investors who want in on AI, it pays to be informed.
For a full list of holdings, please visit www.true-shares.com/lrnz.
The TrueShares AI & Deep Learning ETF (AI ETF) is also subject to the following risks: Artificial Intelligence, Machine Learning and Deep Learning Investment Risk – the extent of such technologies’ versatility has not yet been fully explored. There is no guarantee that these products or services will be successful and the securities of such companies, especially smaller, start-up companies, are typically more volatile than those of companies that do not rely heavily on technology.