Need to build a neural net from scratch? Wondering about Elon Musk’s latest views on AI? Looking for a quick way to calculate eigenvalues? The Internet can offer a wealth of information for artificial intelligence and machine learning practitioners—sometimes too much information.
That’s why we’ve created this curated list of the best resources for AI/ML pros. Whether you’re looking for answers to burning questions, advice for building projects, or just trying to stay abreast of the latest developments in your discipline, you’re likely to find the answers in these podcasts, newsletters, communities, and YouTube channels.
3Blue1Brown, which has 4.51 million YouTube subscribers, is the brainchild of Grant Sanderson, who says he started the channel as a side project while he was wrapping up his time studying math and computer science at Stanford. Videos in the channel are a combination of mathematics and entertainment. Sanderson uses animation to make complex concepts accessible to non-math geeks and to share his enthusiasm for math as a useful tool and an art form. Among the topics touched on by the channel are linear algebra, calculus, neural networks, the Riemann hypothesis, Fourier transform, quaternions, and topology.
The center aims to create a new field—the science and engineering of intelligence—by bringing together computer scientists, cognitive scientists, and neuroscientists to work in close collaboration. This new field is dedicated to developing a computationally based understanding of human intelligence and establishing an engineering practice based on that understanding. It is supported by the National Science Foundation, under a Science and Technology Centers Integrative Partnerships award.
A sister channel of Numberphile, Computerphile is one of the finest channels on YouTube, as its 2.12 million subscribers can attest. Each video is engaging from every beginning. These short videos are packed full of concepts communicated through wit and wisdom. The host plays the role of the audience and asks questions that trouble amateurs, making it feel like an interactive experience. The videos focus primarily on the science behind how computers work. This knowledge is crucial for learning concepts such as computer vision and the computational powers of GPU.
Deeplizard is a training company focused on deep learning. Training videos in its YouTube channel cover subjects such as deep learning fundamentals, TensorFlow, PyTorch, reinforcement learning, and generative adversarial networks.
This is one of the fastest-growing YouTube channels that cover AI and ML. Fridman is a deep learning researcher from MIT famous for his tutorials on computer vision models. In his videos, which have attracted 1.68 million subscribers, Fridman interviews high-profile figures in the industry, including Qualcomm CEO Cristiano Amon, Mark Zuckerberg, and Elon Musk, giving his followers a rich source of content to help understand the world of AI.
This channel makes videos about new papers from research labs at companies such as Google and Facebook. It also covers popular ideas in deep learning and AI and topics including computer vision, natural language processing, graph embeddings, generative adversarial networks, and reinforcement learning, with a coding video sprinkled in from time to time.
Kilcher is an ML researcher and engineer. His YouTube Channel focuses on explaining ML papers, programming, and issues in the AI community, as well as the broader impact of AI in society. In a typical video, Kilcher introduces a paper and describes the novelty behind it. His analyses are detailed, but he makes the papers easy to understand by breaking them up into digestible chunks. In addition to explanations with a bit of snark from time to time, Kilcher also covers topics such as deep learning architecture, natural language processing, and reinforcement learning.
Facebook's parent company, Meta, is heavily involved with AI, and its YouTube channel is a way to keep abreast of Meta’s fundamental and applied research. The channel covers a wide range of topics, from getting AI models to understanding language as a human understands it to taking rapid action against new and evolving harmful online content.
This channel by a British AI researcher focuses on safety research—what humanity can do to anticipate the problems AI might pose in the future and work out ways to ensure that AI developments are safe and beneficial. Subjects touched on by Miles include AI alignment, mesa optimizers, and utility maximizers.
OpenAI is an AI laboratory made up of a for-profit corporation, OpenAI LP, and its parent, the nonprofit OpenAI Inc., whose mission is to promote and develop AI in a way that benefits humanity. Its Codex model, which creates code by parsing natural language, is the subject of many of the videos in this YouTube channel. These include using Codex to create original art from a simple sentence, convert Python to Ruby, and create a space game.
IBM Data Science Specialist Nicholas Renotte says the goal of his YouTube channel is to make it easier to get started by breaking through all the fancy jargon and complex math in the data science field and getting his audience up and running with some “sick” ML projects. Some of those projects include building a comment toxicity model using deep learning, creating deepfakes, and building a deep facial recognition app.
If you’re looking for great interactive tutorials about data science topics, this is the channel to tune in to. Rohrer, a staff ML engineer and data scientist at LinkedIn, clearly knows what he’s talking about, but he cuts down on the jargon surrounding the subject and conveys the concepts with intuitive real-world examples. If you want to learn how large companies engage in data science on massive user datasets, turn to someone who does it every day.
Channel host Harrison Kinsley conducts interactive coding sessions and walks his audience of 1.12 million subscribers through every line of code in his projects, which include one that involves building a neural network from scratch. The videos are a rich source of information for anyone who is serious about a career in data science. Sentdex brings back the fun to coding in a great way.
The Simons Institute of the Theory of Computing is a new venue for collaborative research in theoretical computer science. Housed on the UC Berkeley campus, the institute aims to bring together the world’s leading researchers in theoretical computer science and related fields, as well as the next generation of outstanding young scholars. The channel also explores deep unsolved problems about the nature and limits of computation. Some sample topics covered in the videos in this channel include quantum computing, Markov persuasion, equilibrium computation and ML, and the invisible hand of prediction.
A new research paper on AI gets released every other day. It can describe be a small adjustment in the gradient descent method that improved accuracy by 0.5% or something as big as GANs. Whatever the number, Two Minute Papers, with 1.22 million YouTube subscribers, ensures that its audience does not miss out on the most significant advancements of interest to the AI community.
W&B is an AI vendor that produces a lot of content for ML engineers. Some of it involves the vendor, such as “Fine-Tuning GPT-3 with the OpenAI API and Weights & Biases to Generate Doctor Who Episode Synopses.” There are also interviews with leading industry experts—data scientist Andrada Olteanu, Nvidia CEO Jensen Huang, and Kaggle legend Chris Deotte—as well as how-to videos on subjects, such as fastAI and blur, Keras, and JAX.
Eight years ago, podcast host and Emerj CEO Daniel Faggella was a martial arts instructor in a small town. Today, he’s helping global enterprises and thousands of vendors see a return on investment from their AI investments. Faggella uses his investment savvy each week as he interviews top AI and ML executives, investors, and researchers to unearth the facts and trends that matter to business leaders.
This online community offers podcasts, blogs, and events where accomplished data professionals share stories, knowledge, and advice. Some recent topics include advice for Ph.D. students in investment research, using data for marketing automation, the missing piece for autonomous driving, basics of data-driven transformation for business, and the top five high-growth AI/ML companies.
This podcast makes no bones about being unabashedly technical and noncommercial. Annoying pitches, corporate speak, and MBA waffle are banned, according to its managers, Tim Scarfe, Yannic Kilcher, and Keith Duggar. Most of the podcasts are interviews with leading thinkers in the AI space.
UC Berkeley professor, AI researcher, and entrepreneur Pieter Abbeel hosts this podcast. In each episode, he interviews high-profile figures in the robotics field from all over the world and explores how far humanity has come in its mission to create conscious computers, mindful machines, and rational robots.
Industry analyst, speaker, commentator, and thought leader Sam Charrington interviews some of the top minds in AI and ML for this podcast. Technologies covered include deep learning, natural language processing, neural networks, analytics, computer science, and data science.
This is part of VentureBeat’s stable of newsletters. Published on Fridays, it has a nice set of editorial thoughts up front and roughly a dozen picks of news articles from VentureBeat and elsewhere. It’s a good source for learning about the major stories in any given week.
This newsletter is one of 64 published by Essentials. With a circulation of 400,000, the publication delivers a good-sized roundup of AI articles that are nicely formatted. In addition to the newsletter, Essentials offers specialized subtopics in ethics, applied use cases, robotics, research, marketing, healthcare, cybersecurity, AI for good, and the business of AI.
This newsletter and website covers breaking news about AI on a daily basis. Stories are concisely summarized with bullet points and linked to full articles. A great source to satisfy information junkies with a jones for the latest AI news.
This newsletter is a great way to catch up on a week’s worth of AI news. Curated articles focus on nontechnical coverage about AI advancements, business, and societal impacts. In addition to links with descriptions, the newsletter also has a “top news” section with a brief summary of each article’s contents. The publication is one of the most comprehensive in the news field.
If you’re reading this article you’re probably already a member of AI Exchange, a community that brings together more than 20,000 AI researchers, practitioners, and leaders from across the world. AI Exchange was launched in October 2021 as part of the TransformX conference. Its mission is to bring together AI/ML researchers and practitioners to share content and advice, participate in engaging online and in-person events, and connect with fellow members.
Community members join Exchange to stay up to date on the latest AI research and trends from peers and leaders who are pushing the boundaries of what’s possible, whether in academia, research, industry, or government. AI Exchange features Tech Talks, live conference sessions, articles, and videos by and for AI practitioners.
More than 30,000 developers are part of the Made with ML community, which can be a useful resource for ML engineers looking for ideas for projects to build and for those looking to share and showcase projects that they’ve already built.
This online community aims to meet the need of ML operations engineers by sharing best practices and information needed to solve the unique problems MLOps engineers face every day when building production AI/ML pipelines.