Reward function values using learned weights (../walk-forecasting_learned_weights.txt)

You can use these files to confirm that you are computing the reward correcly from the features (../feature_maps/*)

Reward = sum_i weight(i) * feature_map(i)
