Is Data Science Making Our Brains Less Mathematical?
Written on
Chapter 1: The Impact of Data Science on Mathematical Skills
In recent discussions, I've been pondering whether the plethora of programming tools available today is causing a decline in our ability to solve mathematical problems using traditional methods. This idea stems from my deep interest in Data Science and statistical programming, as many of you may know. I wonder if our eagerness to code solutions is making us less adept at tackling problems on paper, especially when a straightforward approach could yield quicker answers.
Back in the late 1990s, during my Ph.D. studies, our options for mathematical programming were quite limited. We typically resorted to coding only when we struggled to derive solutions manually. In contrast, the current landscape offers a vast array of programming options, which sometimes seems to be the preferred method for problem-solving.
To reignite my grasp of essential mathematical concepts, I've been reading Martin Liebeck’s A Concise Introduction to Pure Mathematics. This book presents intriguing problems that should be manageable for any competent high school student. To test this, I posed three of these problems to a group of data scientists. Here’s what unfolded.
Problem 1: The Restaurant Bill
Consider this scenario:
At a restaurant, the total bill amounts to x dollars and y cents, where both x and y are integers greater than 1, satisfying the equation ( y^{4/3} = x^{5/6} ). What is the total bill?
I requested three data scientists to solve this. They all turned to code, calculating ( y^{4/3} ) for values of y from 2 to 99 and then checking if the result raised to ( 3/4 ) was an integer. This process, although effective, was time-consuming.
However, it can be resolved simply with paper and pencil. By converting the indices to a common denominator of 6, we conclude that ( (y^?)^{1/6} = (x^?)^{1/6} ), leading to ( y? = x? ). Since y must be below 100, we can use powers of 2, resulting in a bill of $256.32.
Problem 2: Finding n
Next, consider the equation:
Given that n and ( sqrt{n-2} + sqrt{n+2} ) are both integers, determine n.
Two data scientists opted to simulate values of n. They quickly found a solution, with one using a simple guess-and-check method. However, this approach only partially addresses the problem, as it doesn't prove that n has a unique solution.
To confirm the solution and demonstrate its uniqueness, we note that ( n-2 ) must be positive, implying ( n geq 2 ). We also find that ( (sqrt{n-2} + sqrt{n+2})^2 ) must be an integer. By expanding, we derive that ( 2(n + sqrt{n^2-4}) ) must also be an integer. It follows that ( sqrt{n^2-4} ) must be an integer as well. Consequently, for n less than or equal to 2, we conclude that n must equal 2.
Problem 3: The Salamander Conundrum
Here’s the final, and arguably the most challenging, problem:
On a deserted island, 30 salamanders exist—15 red, 7 blue, and 8 green. When two salamanders of different colors meet, they both transform into the third color. When two of the same color meet, they each change into one of the other two colors. Is it possible for all salamanders to eventually turn red?
Once again, the data scientists turned to simulations, with some running random trials up to a million times without success. Another attempted to program a loop that would terminate when all salamanders turned red, but this process seemed endless.
This illustrates the distinction between statistics and pure mathematics. A statistician might conclude that it’s improbable for all salamanders to turn red based on their simulations, as they reduce probabilities to a point where the null hypothesis is deemed "unlikely." However, this does not constitute proof.
To establish proof, we can analyze the remainder of each color when divided by three: red is 0, blue is 1, and green is 2. Notably, when any two salamanders meet, the remainder structure remains unchanged, preserving the balance of 0, 1, and 2. Therefore, if all 30 salamanders were to turn red, the remainder would be 0, 0, 0, which is impossible under the established rules.
In conclusion, how would you tackle these problems? Do you believe that the rise of data science is leading to a decline in our mathematical thinking? I welcome your thoughts in the comments.
The first video, featuring Andrew Huberman, delves into the relationship between laziness, discipline, exercise, stress, and journaling. He discusses how these factors interconnect and affect our productivity and mental health.
The second video provides insights on how to rewire your brain to overcome laziness, offering practical strategies that can be implemented today.