McKinsey: Notes from the A.I. Frontier

According to McKinsey's research deep learning techniques such as feed forward neural networks, recurrent neural networks, and convolutional neural networks account for about 40% of the annual value potentially created by all analytics techniques. They estimate that these specific pieces of technology alone can create between $3.5 trillion and $5.8 trillion in value annually.

Read More
mike gervits
National Academy of Sciences: Lectured Students 1.5x More Likely to Fail

In this study, students being lectured are 1.5x more likely to fail than students in classes with more stimulating, active learning methods. Conventional classrooms, however, continue to be driven by the lecture method because one-to-one instruction has been too expensive. Instructors do not have the time or resources to personalize every lesson or drill deep into topics where each student is struggling.


Read More
mike gervits
Bloom's Landmark Study on the Benefits of One-to-One Instruction

Benjamin Bloom empirically proved the achievement benefits of one-to-one instruction over the conventional classroom in 1984. 90% of students receiving one-to-one instruction tested in the 20% of student who learned in conventional classroom settings. Calling it the "Two Sigma Problem," he lamented that "most students do have the ability to reach this high level of learning" but that individualized instruction is "too costly for most societies to bear."

Read More
mike gervits