Given:
$$g(x)=e^{-1/f(x)}, \textbf{dom} g=\{x|f(x)<0\}$$
where $f$ is convex.
How can I prove that $g(x)$ is convex, concave or neither?
I have tried to find a counter example and failed. So based on those examples I've tried, I think $g$ is convex but not sure how to prove it.
Tried using the definition of convexity where $ 0 \le \theta \le 1$ and I know that: $$f(x_0) < 0, f(x_1) < 0$$ $$f(\theta x_0 + (1-\theta) x_1) \le \theta f(x_0) + (1-\theta) f(x_1)$$
and need to show that: $$g(\theta x_0 + (1-\theta) x_1) \le \theta g(x_0) + (1-\theta) g(x_1)$$ $$e^{-1/f(\theta x_0 + (1-\theta) x_1)} \le \theta e^{-1/f(x_0)} + (1-\theta) e^{-1/f(x_1)}$$
But I am stuck here. Taking the log of both sides didn't get me very far.