19 May 2020
This framework is composed of two networks, generator and discriminator which are trained adversarially. The generator networks competes with discriminator network to generate data which seems to be sampled from true data distribution and discriminator is supposed to differentiate the real data from generated data.
20 Jul 2019
Dijkstra’s Algorithm is the most basic search algorithm for a planning problem. For a given start state, goal state and a weighted graph, Dijkstra’sAlgorithm can find the sequence of states with the smallest sum of edge costs connecting these states. But unaware of the direction in which goal state actually lies, this algorithm finds the shortest path from the start state to every other state which is closer to start state than goal state. This seemingly uninteresting fact appeared as a revelation to me as I noticed the metaphorical resemblance it had with life.