From 5c4f2c883f648e8b886e9292b1c2598d8a574c07 Mon Sep 17 00:00:00 2001
From: Bas Nijholt <basnijholt@gmail.com>
Date: Wed, 25 Sep 2019 10:55:19 +0200
Subject: [PATCH] clarify caption

---
 paper.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/paper.md b/paper.md
index 2ddf429..2aa83c3 100755
--- a/paper.md
+++ b/paper.md
@@ -39,7 +39,7 @@ One of the most significant complications here is to parallelize this algorithm,
 ![Visualization of a 1-D point choosing algorithm for a black box function (grey).
 We start by calculating the two boundary points.
 Two consecutive existing data points (black) $\{x_i, y_i\}$ define an interval.
-Each interval has a loss associated with it that can be calculated from the points inside the interval $L_{i,i+1}(x_i, x_{i+1}, y_i, y_{i+1})$.
+Each interval has a loss $L_{i,i+1}$ associated with it that can be calculated from the points inside the interval $L_{i,i+1}(x_i, x_{i+1}, y_i, y_{i+1})$ and optionally of $N$ next nearest neighbouring intervals.
 At each iteration the interval with the largest loss is indicated (red), with its corresponding candidate point (green) picked in the middle of the interval.
 The loss function in this example is the curvature loss.
 ](figures/algo.pdf){#fig:algo}
-- 
GitLab