Topics

  • Beam search strikes a compromise between the efficiency of greedy search and the optimality of exhaustive search, at the expense of introducing one hyper-parameter , the number of paths (beam_size) we keep while running a beam search
  • Beam search will always find an output sequence with higher probability than greedy search, but is not guaranteed to find the most likely output
  • When , the beam search becomes the greedy search. When  is not limited, it becomes an exhaustive search