# (2016. 10) Diverse Beam Search

* Submitted on 2016. 10
* Ashwin K Vijayakumar, Michael Cogswell, Ramprasath R. Selvaraju, Qing Sun, Stefan Lee, David Crandall and Dhruv Batra

## Simple Summary

> propose Diverse Beam Search(DBS), an alternative to BS that decodes a list of diverse outputs by optimizing for a diversity-augmented objective. We observe that our method finds better top-1 solutions by controlling for the exploration and exploitation of the search space -- implying that DBS is a better search algorithm. Moreover, these gains are achieved with minimal computational or memory overhead as compared to beam search.

![images](https://1712266326-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-LMrEcS7cR9bGTHSnCnB%2F-LRazIqPSIKca1ujTQk5%2F-LRazKz8ptgG9BtSB7-N%2Fdbs_1.png?generation=1542547411140190\&alt=media)

* optimize an objective that consists of two terms – the sequence likelihood under the model and a dissimilarity term that encourages beams across groups to differ.
* This diversity-augmented model score is optimized in a doubly greedy manner – greedily optimizing along both time (like BS) and groups (like DivMBest).

![images](https://1712266326-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-LMrEcS7cR9bGTHSnCnB%2F-LRazIqPSIKca1ujTQk5%2F-LRazKzAD81EPhXQ-Oft%2Fdbs_2.png?generation=1542547410990589\&alt=media)