Machine Learning in Elixir: Plot of given `profits` function does not match the expected plot (B1.0, Chapter 3, Page 65)

Hey @seanmor5,

The function given in the book

defn profits(trees) do
  trees
  |> Nx.pow(4)
  |> Nx.negate()
  |> Nx.add(Nx.multiply(2, Nx.pow(trees, 3)))
  |> Nx.add(Nx.pow(trees, 2))
end

Results in the following plot which doesn’t match the expected plot in the book:

The accompanying Livebooks provides the following code to give the expected plot:

defn profits(trees) do
  -((trees - 1) ** 4) + trees ** 3 + trees ** 2
end

So my guess is that there is an error in the conversion to Nx functions.

Instead, the correct translation of the Livebook function should be as follows:

  defn profits(trees) do
    trees
    |> Nx.subtract(1)
    |> Nx.pow(4)
    |> Nx.negate()
    |> Nx.add(Nx.pow(trees, 3))
    |> Nx.add(Nx.pow(trees, 2))
  end
  • Adds the missing Nx.subtract(1) that the book doesn’t have
  • Removes the unnecessary doubling of the cube in Nx.add(Nx.multiply(2, Nx.pow(trees,3)))

This gives the correct expected plot

Thank you for pointing this out! I’ve fixed the issue in the next beta release!

1 Like