Most large topic models are approximately separable. Ding, W., Ishwar, P., & Saligrama, V. In 2015 Information Theory and Applications Workshop, ITA 2015, San Diego, CA, USA, February 1-6, 2015, pages 199–203, 2015.
Most large topic models are approximately separable [link]Paper  doi  bibtex   
@inproceedings{DBLP:conf/ita/DingIS15,
  author    = {Weicong Ding and
               Prakash Ishwar and
               Venkatesh Saligrama},
  title     = {Most large topic models are approximately separable},
  booktitle = {2015 Information Theory and Applications Workshop, {ITA} 2015, San
               Diego, CA, USA, February 1-6, 2015},
  pages     = {199--203},
  year      = {2015},
  crossref  = {DBLP:conf/ita/2015},
  url       = {https://doi.org/10.1109/ITA.2015.7308989},
  doi       = {10.1109/ITA.2015.7308989},
  timestamp = {Sun, 25 Oct 2020 01:00:00 +0200},
  biburl    = {https://dblp.org/rec/conf/ita/DingIS15.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

Downloads: 0