AI 360Cornell Certificate Program

Artificial Intelligence (AI) is no longer a futuristic concept; it’s a reality that is transforming our world. This course uncovers the true essence of AI, cutting through the hype to reveal its practical applications and real-world value. You'll explore AI's role in creating significant business opportunities and learn to apply a value-driven framework to identify these opportunities. Through interactive exercises and hands-on data analytics, you'll gain the skills to characterize the business potential of AI products.

The course will cover the spectrum of AI technologies, from traditional rule-based systems to cutting-edge neural networks and generative AI. You'll focus on practical applications, learning not only how to leverage AI's power but also how to address its risks and ethical implications. By the end of this course, you'll be able to distinguish between hype and reality, formulate AI product ideas with actionable value, and effectively utilize AI algorithms in product development to drive impactful results.

Artificial Intelligence (AI) has advanced to simulate human intelligence, enabling computers to perform tasks such as understanding natural language, learning from experience, problem-solving, and making decisions. This course introduces you to both traditional and modern AI approaches, starting with Good Old-Fashioned AI (GOFAI) which relies on symbolic logic. You'll learn how to teach computers to make predictions and decisions using machine learning techniques, focusing on practical applications that solve real-world problems and create business value.

Through hands-on exercises, you will design and refine machine learning models, including logistic regression and decision trees. You'll develop the skills to build AI systems that can predict outcomes and improve over time. By the end of this course, you will be able to create AI solutions that effectively leverage data to meet specific business objectives.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Creating Business Value with AI

In the world of Artificial Intelligence (AI), data is the lifeblood of effective models. The adage "garbage in, garbage out" highlights the necessity of using high-quality data to train AI systems. This course is designed to equip you with the skills to define data requirements and acquire necessary data through web scraping techniques. You will learn to categorize and analyze data for relevance and insights, ensuring its quality through meticulous cleaning and preprocessing.

We will examine various aspects of data preparation, including handling missing values, identifying and addressing outliers, and ensuring data consistency. The course also covers the critical issues of data bias, privacy, and ethical considerations, providing strategies to mitigate these challenges. You will explore how to build resilient AI models and understand the influence of data on different features of a business model. By the end of this course, you will be able to leverage data for strategic competitive advantage and create AI models that drive meaningful business outcomes.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Creating Business Value with AI
• Exploring Good Old-Fashioned AI (GOFAI)

Good Old-Fashioned AI (GOFAI) is effective for many tasks but has limitations when dealing with complex data patterns. Neural networks (NN), inspired by the human brain, offer a powerful alternative by learning from data patterns and relationships. This course will introduce you to the foundations and architectures of neural networks, enabling you to train, evaluate, and optimize these models to improve performance. You will explore the impact of data volume and model complexity on predictions, ensuring you select the most suitable NN models for various business scenarios. Additionally, the course addresses ethical considerations, such as model bias and data privacy, to ensure responsible AI implementation.

By examining real-world applications and engaging in hands-on exercises, you will develop practical skills in configuring neural networks and evaluating their performance. This course will equip you with the knowledge to apply these techniques effectively and create an action plan to address ethical concerns, ensuring your AI projects are both effective and responsible.

You are required to have completed the following courses or have equivalent experience before taking this course:

• • Creating Business Value with AI
  • Exploring Good Old-Fashioned AI (GOFAI)
  • Leveraging Data for AI Solutions

Generative AI is revolutionizing how we create and innovate, offering new possibilities for product development and user engagement. This course delves into the current uses of generative AI and explores future innovations. You will learn to leverage generative AI to test and refine your value-creation ideas, considering the sociocultural implications of your product concepts. The course will guide you through the process of using generative AI techniques to bring a specific AI product idea to life, with a strong emphasis on ethical considerations, such as model bias and data privacy.

Through hands-on exercises and real-world examples, you will develop practical skills in crafting business ideas using generative AI. You will learn to enhance AI-driven user interfaces and experiences, create effective prompting strategies for AI text generation, and develop strategic plans for ethical AI implementation. By the end of this course, you will be equipped to use generative AI to drive innovation and create value in a responsible and impactful way.

You are required to have completed the following courses or have equivalent experience before taking this course:

• • • Creating Business Value with AI
    • Exploring Good Old-Fashioned AI (GOFAI)
    • Leveraging Data for AI Solutions
    • Expanding AI Power and Value Through Neural Networks

By some estimates, 90% of the data that has ever existed has been created in the last two years. This is a staggering figure and has given rise to new challenges and opportunities in almost every industry: what kind of data do you need to collect to compete, and how can you make sense of it once you have collected it? As technology evolves and the volume of data increases, how can you make the best use of all this information? How can you use the data to help drive your decision-making? How can you make data work for you? How can you ensure your data accurately reflects the population in which you're interested?

In this course, you will determine the types of engineering and business questions you can answer, the kinds of problems you can solve, and the decisions you can make, all through using data analytics. You will explore best practices for collecting information so that you can make informed predictions, develop insights, and better inform organizational decision-making. You will see real-world examples that demonstrate how those tools work. Additionally, you will have a chance to apply some of the concepts to your own work. You will explore best practices for sampling and examine how different types of sampling are each suited for different situations. Finally, you will see real-world examples that demonstrate how those tools work and have a chance to practice sampling techniques in some case study scenarios.

Visualization is one of the most simple and effective ways to find patterns in data. These patterns include: What is the general range and shape of the data set? Are there any clusters of observations? Which variables correlate with each other? Are there any obvious outliers?

As your data set grows in terms of the number of data points and variables, however, it becomes increasingly difficult to visualize all this information at once. At most, you can plot data points on a three-dimensional axis and add further distinctions of size, color, shape, and so on. Yet this can easily become too busy and difficult to read. How, then, do we find patterns in really big data sets?

In this course, you will explore several powerful and commonly utilized techniques for distilling patterns from data. You will implement each of these techniques using the free and open-source statistical programming language R with real-world data sets. The focus will be on making these methods accessible for you in your own work.

You are required to have completed the following course or have equivalent experience before taking this course:

• Understanding Data Analytics

When you have large groups of objects, it is often helpful to split them into meaningful groups or clusters. One example of this would be to identify different types of customers so that a company can more efficiently route their calls to a helpline. As a second example, suppose an automobile manufacturer wanted to segment their market to target the ads more carefully. One approach might be to take a database of recent car sales, including the social demographics associated with each customer, and segment the population purchasing each type of automobile into meaningful groups.

Specialized approaches exist if your data contains information that relates to time and geography. You can use this additional information to identify geographical and temporal hotspots. Hotspots are regions of high activity or a high value of a particular variable. These results can help you focus your attention on a particular region where a problem is occurring more than usual, such as the incidence of asthma in a large city. In both cluster and hotspot analysis, the results can help you discover new and interesting features, problems, and red flags regarding the data being analyzed.

In this course, you will explore several powerful and commonly utilized techniques for performing both cluster and hotspot analysis. You will implement these techniques using the free and open-source statistical programming language R with real-world data sets. The focus will be on making these methods accessible and applicable to your work.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Understanding Data Analytics
• Finding Patterns in Data Using Association Rules, PCA, and Factor Analysis

A story can play an important role in understanding data. It can help distill complex information into something manageable- something we can think about easily, relate to, and use to make decisions. For many problems that we encounter globally, however, a story that describes what already happened is not enough precision for the job we want to perform. Often, we would like to use available data to make numerically accurate predictions about what might happen in the future. This task requires the construction of mathematical models that are well suited to our real-world problems.

In this course, you will explore several types of statistical models used with data to make predictions. These models bring with them a whole batch of important concerns, such as estimation and validation, that make the entire process into both an art and a science. You will implement each of these techniques using the free and open-source statistical programming language R with real-world data sets. The focus will be on making these methods accessible for you in your own work.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Understanding Data Analytics
• Finding Patterns in Data Using Association Rules, PCA, and Factor Analysis
• Finding Patterns in Data Using Cluster and Hotspot Analysis

Supervised learning is a general term for any machine learning technique that attempts to discover the relationship between a data set and some associated labels for prediction. In regression, the labels are continuous numbers. This course will focus on classification, where the labels are taken from a finite set of numbers or characters. The prototypical and perhaps most well-known example of classification is image recognition. The goal is to take an image (represented by its pixel values) and determine what objects are in the image. Is it a dog? A grapefruit? A stop sign?

There are many practical classification tasks, such as determining whether an individual's financial history makes them high risk for a loan, whether there is a defect in a material based on some sensor readings, or whether a new email is spam or not. These problems share the same basic form and can be solved with many different types of mathematical, statistical, and probabilistic models developed by the machine learning community.

In this course, you will explore several powerful and commonly utilized techniques for supervised learning. You will implement each of these techniques using the free and open-source statistical programming language R with real-world data sets. The focus will be on making these methods accessible for you in your own work.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Understanding Data Analytics
• Finding Patterns in Data Using Association Rules, PCA, and Factor Analysis
• Finding Patterns in Data Using Cluster and Hotspot Analysis
• Regression Analysis and Discrete Choice Models

Neural networks, a nonlinear supervised learning modeling tool, have become hugely popular within the last two decades because they have been successfully applied to a wide range of problems, including automatic language processing, image classification, object detection, speech recognition, and pattern recognition. They are mathematical models that are loosely built up based on an analogy to the interconnected neuron in the brain. They take in a vector or matrix of input data and output either a classification value or an approximation to a functional value. The beauty is that the relationships between the inputs and outputs can be highly non-linear and complex.

In this course, you will explore the mechanics of neural networks and the intricacies involved in fitting them to data for prediction. Using packages in the free and open-source statistical programming language R with real-world data sets, you will implement these techniques. The focus will be on making these methods accessible for you in your own work.

You are required to have completed the following courses or have equivalent experience before taking this course:

• Understanding Data Analytics
• Finding Patterns in Data Using Association Rules, PCA, and Factor Analysis
• Finding Patterns in Data Using Cluster and Hotspot Analysis
• Regression Analysis and Discrete Choice Models
• Supervised Learning Techniques

In this course, you will use the Maximum Likelihood Estimate (MLE) to approximate distributions from data. Using the Bayes Optimal Classifier, you will learn how the assumptions you make will impact your estimations. You will then learn to apply the Naive Bayes Assumption to estimate probabilities for problems that contain a high number of dimensions. Ultimately, you will apply this understanding to implement the Naive Bayes Classifier in order to build a name classification system.

The following course is required to be completed before taking this course:

• Problem-Solving with Machine Learning

In this course, you are introduced to and implement the Perceptron algorithm, a linear classifier that was developed at Cornell in 1957. Through the exploration of linear and logistic regression, you will learn to estimate probabilities that remain true to the problem settings. By using gradient descent, we minimize loss functions. Ultimately, you will apply these skills to build a email spam classifier.

The following courses are required to be completed before taking this course:

• Problem-Solving with Machine Learning
• Estimating Probability Distributions

In this course, you will be introduced to the classification and regression trees (CART) algorithm. By implementing CART, you will build decision trees for a supervised classification problem. Next, you will explore how the hyperparameters of an algorithm can be adjusted and what impact they have on the accuracy of a predictive model. Through this exploration, you will practice selecting an appropriate model for a problem and dataset. You will then load a live dataset, select a model, and train a classifier to make predictions on that data.

The following courses are required to be completed before taking this course:

• Problem-Solving with Machine Learning
• Estimating Probability Distributions
• Learning with Linear Classifiers

In this course, you will investigate the underlying mechanics of a machine learning algorithm's prediction accuracy by exploring the bias variance trade-off. You will identify the causes of prediction error by recognizing high bias and variance while learning techniques to reduce the negative impacts these errors have on learning models. Working with ensemble methods, you will implement techniques that improve the results of your predictive models, creating more reliable and efficient algorithms.

These courses are required to be completed prior to starting this course:

• Problem-Solving with Machine Learning
• Estimating Probability Distributions
• Learning with Linear Classifiers
• Decision Trees and Model Selection

In this course, you will explore support-vector machines and use them to find a maximum margin classifier. You will then construct a mental model for how loss functions and regularizers are used to minimize risk and improve generalization of a learning model. Through the use of feature expansion, you will extend the capabilities of linear classifiers to find non-linear classification boundaries. Finally, you will employ kernel machines to train algorithms that can learn in infinite dimensional feature spaces.

These courses are required to be completed prior to starting this course:

• Problem-Solving with Machine Learning
• Estimating Probability Distributions
• Learning with Linear Classifiers
• Decision Trees and Model Selection
• Debugging and Improving Machine Learning Models

In this course, you will investigate the fundamental components of machine learning that are used to build a neural network. You will then construct a neural network and train it on a simple data set to make predictions on new data. We then look at how a neural network can be adapted for image data by exploring convolutional networks. You will have the opportunity to explore a simple implementation of a convolutional neural network written in PyTorch, a deep learning platform. Finally, you will yet again adapt neural networks, this time for sequential data. Using a deep averaging network, you will implement a neural sequence model that analyzes product reviews to determine consumer sentiment.

These courses are required to be completed prior to starting this course:

• Problem-Solving with Machine Learning
• Estimating Probability Distributions
• Learning with Linear Classifiers
• Decision Trees and Model Selection
• Debugging and Improving Machine Learning Models
• Learning with Kernel Machines

To perform basic computations in the Machine Learning certificate program, you need the ability to solve elementary linear algebra problems in two dimensions. In this course, you will execute mathematical computations on vectors and measure the distance from a vector to a line. This course will provide you with the theory and activities to start building the linear algebra foundation needed to be successful in your Machine Learning courses.

This optional self-paced course supports the required linear algebra in the Machine Learning certificate. If you are already comfortable with the computations from the pretest, we recommend that you take the final assessment to confirm your readiness.

This Machine Learning certificate program requires you to think and solve problems in multiple dimensions. In this course, you will learn to solve linear algebra problems in three or more dimensions and perform computations with matrices.You will perform computations that focus on solving problems in high dimension; that is, multiple dimensions. This course will provide you with the theory and activities to solidify the linear algebra foundation needed to be successful in your Machine Learning courses.

This optional self-paced course supports the required linear algebra in the Machine Learning certificate. If you are already comfortable with the computations from the pretest, we recommend that you take the final assessment to confirm your readiness.

Symposium sessions feature three days of live, highly interactive virtual Zoom sessions that will explore today’s most pressing topics. The AI Symposium offers you a unique opportunity to engage in real-time conversations with peers and experts from the Cornell community and beyond. Using the context of your own experiences, you will take part in reflections and small-group discussions to build on the skills and knowledge you have gained from your courses.

Join us for the next Symposium, in which we’ll share experiences from across the industry, inspiring real-time conversations about best practices, innovation, and the future of AI. You will support your coursework by applying your knowledge and experiences to some of the most pressing topics and trends in the field. By participating in relevant and engaging discussions, you will discover a variety of perspectives and build connections with your fellow participants from across a variety of industries.

All sessions are held on Zoom.

Future dates are subject to change. You may participate in as many sessions as you wish. Attending Symposium sessions is not required to successfully complete any certificate program. Once enrolled in your courses, you will receive information about upcoming events. Accessibility accommodations will be available upon request.