Figure 4.

The pipeline for testing the Hyperdimensional Inference Layer. Images are presented to the hashing network H, which are then hashed into binary vectors. As during training, these are projected into hyperdimensional vectors. The result is then XOR'd with the HIL. The XOR distributes across the terms in the HIL and creates noise for terms corresponding to incorrect classes. Only the correct class will deviate from the noise and will be detected as the best matching class by computing the Hamming Distance between the result of the XOR and every vector representation of the classes; the class with the smallest Hamming Distance is selected as the correct classification.