A Big-Picture Look at our Craft: Thought-Provoking Videos About Machine Learning and AI
Take a well-deserved break while your models are training to reflect on the macro-level impact of machine learning with our choice of videos from leaders in the field.
We often get swept up in the details of our work, so deep in focus on the intricacies of statistics and Python code. Once in a while, we can benefit by taking a short break to relax and embrace an open-minded and big-picture assessment of our profession and its foreseen impact on society. Here we present a handful of popular videos we found helpful and insightful when contemplating the impact of our work from more lofty heights. The topics are diverse, demonstrating the positive and negative intricacies of acquiring big data, training algorithms, and the future use and regulation of various AI technologies. We found these videos to be inspirational that also provide a jolt of motivation and a fun opportunity to refine our professional goals. Check them out and inform us of what value they bring to you as a data scientist or machine learning pundit.
AI economics: How cheap predictions will change the world
Ajay Agrawal has a cool and unique job title: artificial intelligence economist. In this video, he describes the benefits of reducing the mysteriousness and vastness of AI to the simple question, AI reduces the cost of what, exactly? His analysis in this video stems from his popular book co-authored with Avi Goldfarb and Joshua Gans entitled Prediction Machines.
The answer to his broad question is surprisingly simple: machine learning reduces the costs of making predictions. When predictions become ever-more cheaper with advancements in AI capacities, predictions will become used more often in novel areas and applied to new problems, which opens the possibility for new technological and scientific advancements. Assessing the future costs of making predictions in a specific context appears to be a reliable guide towards foreseeing where the popularity of AI will boom. Frame your current machine learning projects within this cost perspective. Do you gain any new insight into its long-term application?
Members of Zetane had the opportunity to attend a few lectures in Montreal by Agrawal’s co-author, Goldfarb, and found the seminars to be top-notch and of particular interest to biz-dev members of an AI company.
Art in the age of machine intelligence
“What does it mean to be an AI in the 21st century?”
Media artist Refik Anadol is revolutionizing the application of machine learning in art by centering his career around this question. He uses diverse datasets — from cultural archival artifacts to images of city spaces — to reimagine traditional depictions of nature and architecture. He then transforms this data into immersive, dream-like worlds. Some of these synthetic hallucinations serve simultaneously as an homage to the ever-changing nature of the world around us and as a time capsule for disappearing memories. Prepare to dive head-first into futuristic worlds made with AI where reality and the digital realm are indistinguishable.
Let’s scan the whole planet with LiDAR
Hey there lovers of big data, this talk is for you. Speaker Chris Fisher describes how LiDAR technology can be used to disrupt the field of archeology and expand opportunities for cultural activism. His project, the Earth Archive, uses LiDAR, Light Detection and Ranging scans to create a non-invasive virtual model of the Earth’s surface. The project is ambitious yet necessary in the preservation of cultural and ecological heritage for future generations. Fisher warns that we are running out of time to develop a baseline record of Earth as it was before the devastating impacts of climate change. A “before” picture of our planet that includes a scan of everything from major cities to the rainforests in Honduras is vital for us to comparatively measure and combat unwanted change. The Earth Archive aims to provide open-source access to these dataset scans of the Earth for all fields of study.
Also of interest here is how applying LiDAR data techniques within archaeology provides a more detailed and cost- and time-efficient method of surveying the land under our feet.
The value of your humanity in an automated future
Want to know how to beat the system and protect your job against AI automation? According to Kevin Roose and his book “Futureproof”, you need to focus on what makes you unique: being human. In this capturing video, Roose explains that the power of neural networks can be used to train robots to replace humans in any industry and that the best way to prepare for this automated future is to master your soft skills acquired from understanding the arts and humanities; here is where you will find the understated value of empathizing with the human experience that robots cannot replicate. This video will hopefully make you feel irreplaceable simply by being you, a human.
Depression is a silent killer. AI is turning up the volume
Don’t have much time before heading back to work? This enlightening short video is just the thing for you.
Author Eric Topol describes the topic of his recent book about how we can use AI to analyze the human brain and diagnose mental illnesses such as depression. By “digitiz[ing] the state of mind”, we acquire data to train AI models to understand elements of nonverbal communication — e.g., speech patterns, speed of talking, and tone of voice — to identify symptoms of mental illnesses. Also of interest for treating mental health issues are observations that some people feel more comfortable disclosing their thoughts and mental health status to a digital confidante rather than a mental health practitioner. This research is of great value: having an AI digital assistant to track and support people seeking mental health services would help counter current problems with overwhelmed and inadequate mental health services that are particularly strained during the pandemic.
If the topic of using AI to treat depression is of interest to a reader, we recommend that you check out the work being done by the folks at Alfred Health.
Can a robot pass a university entrance exam?
Noriko Arai wanted to know if AI technology today can pass an entrance exam for a top college. The answer may surprise you!
The Todai Robot Project of the University of Tokyo studies the performance of AI compared to humans, “especially on skills and expertise which are believed to be acquired only by humans and only through education”. In this video, Arai describes our desire to invent a “fully automatic, mass-solving machine [that] has been a dream since the invention of the term artificial intelligence.” Today we understand that we need to first attend to the fact that there is a remarkable difference between knowledge and understanding. This video is an overall call for reform, urging a shift in the education system away from memorization — an area where AI will supersede us with minimal effort — to learning with an emphasis on meaning.
Interactive experiences with perceptual computing
Technology is a race forward with no end in sight. This video provides a prime example with a glimpse into what’s happening at Intel as the company advances its many innovation initiatives. Back in 2013, Achin Bhowmik hosted a TED Talk about Intel’s development of “natural, intuitive and immersive” human-computer interactions known as perceptual computing. Perceptual computing involves adding senses and perceptions using 3D cameras and object detection and tracking technologies. These foundational studies open opportunities for computers to perceive and communicate with humans, the world, and each other. Here we get a taste of the rapid development of computer “brains” and how they soon may be on par with those of humans. The boundaries between nature and technology are ever more blurred.
Many thanks to Jason Behrmann of Zetane Systems for helpful edits and contributions to drafts of this article.
