This piece in Quartz focused on a lesser-known US college –St. John’s– where things are done quite differently. Away from current trends towards offering students an increasingly wide array of subject choices, this contrarian institution has reaffirmed itself in the needed return to the roots of the liberal arts curriculum.
What is so different about St. John’s? Well, a number of things. It seems like a prejudice against fickle novelty is one. Works that are introduced in the curriculum need to have withstood the test of time. Starting with the Greeks, ‘recent’ science reading put to the students often date from the 1950’s and 1960’s.
Rather than ad hoc variations, all students get the same set list: four years of literature, language, philosophy, political science and economy, and math. Three years of laboratory science, and two of music. That’s it, in an effort to build robust critical thinking capabilities for their small yearly cohorts of graduates.
Another distinct feature is the importance given to context in learning. The sequence in which readings are addressed is considered key; also, not to place texts in their political, social or historic circumstance for discussion. Ideology need not creep into judging the merits of different intellectual pieces of work.
At St. John’s they believe that focus should be put on the texts themselves. The curriculum is carefully designed not only to build knowledge, but also to understand how knowledge is ultimately created.
At Iris.ai we share much of this view. We continuously teach our machine by carefully focusing its learning on texts themselves, de-emphasizing the often-biased metadata around them (like citations).
In fact, a few weeks ago we were honoured –and terribly excited!– to be featured in Science magazine’s latest report (see below) covering the impact of Artificial Intelligence in automating scientific research discovery. How about that for general validation?
Follow our young AI’s journey from tackling concept extraction and topic modeling into, amongst other goals, addressing the identification and formulation of existing and new scientific hypotheses. I promise it will be an exciting one!