Probabilistic Latent Maximal Marginal Relevance
Shengbo Guo and Scott Sanner
In: SIGIR 2010, 19-23 July 2010, Geneva, Switzerland.
Diversity has been heavily motivated in the information re- trieval literature as an objective criterion for result sets in search and recommender systems. Perhaps one of the most well-known and most used algorithms for result set diver- sification is that of Maximal Marginal Relevance (MMR). In this paper, we show that while MMR is somewhat ad- hoc and motivated from a purely pragmatic perspective, we can derive a more principled variant via probabilistic infer- ence in a latent variable graphical model. This novel deriva- tion presents a formal probabilistic latent view of MMR (PLMMR) that (a) removes the need to manually balance relevance and diversity parameters, (b) shows that specific definitions of relevance and diversity metrics appropriate to MMR emerge naturally, and (c) formally derives variants of latent semantic indexing (LSI) similarity metrics for use in PLMMR. Empirically, PLMMR outperforms MMR with standard term frequency based similarity and diversity met- rics since PLMMR maximizes latent diversity in the results.