This guide helps nonprofits and social enterprises learn how to apply artificial intelligence and machine learning to social, humanitarian and environmental challenges.
Whether you’re a novice or interested in upleveling your skill set, get started by exploring the resources in this guide. Here, we’ll help you understand the types of problems your organization can solve with ML, learn how to identify and prepare data sources, and review guidelines on how to develop and utilize ML responsibly.
Artificial intelligence (AI) is the general study of making intelligent machines. Machine learning (ML), a subset of AI, focuses on the ability of machines to receive data and learn for themselves without being programmed with rules. ML differs from traditional programming by allowing you to teach your program with examples rather than a list of instructions. Instead of writing instructions, or rules, while programming, machine learning enables you to "train" an algorithm so that it can learn on its own, and then adjust and improve as it learns more about the information it is processing.
In the content below, you can start building a basic understanding of AI and ML, learn about ML terminology, and explore interactive demos to see the things you can do with ML.
Learn what machine learning is and how it works.
Explore how machine learning works through hands-on experiments with pictures, drawings, language, music, and more. No coding required.
Find out how machine learning is beginning to improve computers, and many of the things we use them for.
From helping farmers in Japan sort cucumbers to helping doctors in India diagnose eye disease, machine learning is changing the way people use code to solve problems and improve lives.
This glossary defines general machine learning terms as well as terms specific to TensorFlow.
Take a peek into how researchers, organizations, and students around the world are using AI and ML to solve complex problems. AI and ML allow us to find patterns never seen before, make decisions, and help shape the world around us.
Google researchers are using AI to predict when and where floods will occur and share these early warnings via Google Public Alerts to keep people informed and safe.
Australian researchers applied AI to drone imagery to track and protect sea cows, an endangered species.
Tanzanian farmers partnered with Penn State researchers to use image recognition to detect plant disease in cassava crops in Tanzania.
Google is partnering with NOAA to apply machine learning to whale sounds to better monitor and conserve whale populations.
Global Fishing Watch uses ML to map fishing activity worldwide, help inform sustainable policy by identifying areas that are overfished.
High school students developed a TensorFlow model to predict wildfires in California.
Quill.org’s machine learning-based tool helps students improve their writing and grammar skills.
Learn about the types of problems you can solve with machine learning and understand the mindset you'll need to frame your problem as a machine learning problem. Fast-paced summaries and interactive exercises help you decide whether machine learning is right for your specific problem, whether your data is right for machine learning, and whether machine learning can give you the outcomes you're looking for.
Learn how to think about and solve problems using machine learning.
It is an oft-repeated maxim that "ML is only as good as the data that it uses." For this reason, it is important not only to know where to look for data, but also what questions to ask when you find it.
In the social sector, accessing high-quality datasets can be challenging. At the same time, not all ML datasets consist of hundreds of thousands of examples. In fact, you may be able to use a small dataset and leverage existing “knowledge” from a mature ML model by reusing parts of existing models (e.g., TensorFlow Hub or through transfer learning technology, like Cloud AutoML. In addition, organizations may have access to more useful data than they realize, whether proprietary (e.g., historical evaluation forms), or datasets in the public domain (more on that below). You may also be able to find collaborators who are willing to share data for social good applications.
Wherever you find your data, consider the following questions before you use the dataset to train your ML system:
After answering these questions, think about whether the data has the features that you need for training or if the data is diverse or complete enough to help you answer our ML questions. The Data Preparation and Feature Engineering course below covers those topics in more detail.
Explore the “Construct your dataset” portion of this course, which will walk you through how to approach collecting and developing a dataset.
Explore how human bias can influence datasets and learn strategies to identify biases and evaluate their effects.
The Google Cloud Public Datasets Program hosts copies of structured and unstructured data to make it easier for users to discover, access, and utilize public data in the cloud. These datasets are hosted for free.
Explore, execute, share, and comment on code for any open dataset with the in-browser analytics tool, Kaggle Kernels. You can also download datasets in an easy-to-read format.
Earth Engine's public data archive includes more than forty years of historical imagery and Earth science datasets, updated and expanded daily.
Google’s Dataset Search tool allows users to search across thousands of dataset repositories on the Web. Once you’ve found a relevant dataset, Dataset Search will direct you to the repository or provider where the data is hosted.
Google has given a lot of thought to the responsible development and use of AI. We have published a set of AI principles, which outline our commitment to develop technology responsibly. We have also developed a set of Responsible AI Practices, which outline general recommended practices for AI, as well as specific considerations for fairness, interpretability, privacy, and security. And our Machine Learning Crash Course module on fairness provides strategies to identify biases in training data and evaluate their effects on model outputs.
Google is committed to making progress in the responsible development of AI and to sharing knowledge, research, tools, datasets, and other resources with the larger community. This document outlines general best practices for AI as well as our current work and recommended practices in fairness, interpretability, privacy, and security.
We are optimistic about the potential for AI and other advanced technologies to empower and benefit people, but we recognize that these same technologies also raise important challenges that we need to address clearly, thoughtfully, and affirmatively. These principles set out our commitment to develop technology responsibly and establish specific areas we will not pursue.
Learn ways to keep fairness considerations top of mind when building, evaluating, and deploying machine learning models. Dive deeper by exploring datasets and classifiers with a few techniques in an interactive colaboratory exercise.
Visually investigate and explore the performance and fairness of a machine learning model. With no coding required, the What-If Tool gives you a simple, intuitive, and powerful way to play with a trained machine learning model on a set of data.
It’s not always the case that you need to have an expert on staff in order to use AI successfully. While AI was once the realm of PhDs and ML researchers, it is now accessible to anyone with a solid understanding of programming fundamentals. With some training, developers who can write C++, Python, JavaScript, or even SQL can apply machine learning to their organization’s challenges, thanks to the availability of open-sourced AI libraries, platforms, or APIs (such as Kubernetes, Cloud AutoML, BigqueryML, DialogFlow, TensorFlow, TensorFlow Hub, and more). See the next section for trainings to get started.
When organizations do not have the required skill set in-house, they can find support from external partners. You may consider reaching out to research, engineering, or data science experts associated with local universities or businesses. There may even be existing pro-bono engineering or data science networks in your community that focus on facilitating these types of partnerships.
Explore Google's recommended practices and trainings to help you get started designing and developing a machine learning system.
Google's fast-paced, practical introduction to machine learning. A self-study guide for aspiring machine learning practitioners.
Presents a style for machine learning, similar to the Google C++ Style Guide or other popular programming guides. Suitable for anyone with a basic knowledge of machine learning.
Google's self-study course shows you how to construct a dataset and transform the data correctly. Suitable for anyone who's taken the Machine Learning Crash Course.
Access additional ML educational resources from across the web, including self-study courses, exercises and information to help you expand your knowledge of machine learning.