nucleation

LIT-1546: Learning author-topic models from text corpora

1 min read

Properties5
StatusActive — read 2026-10-26 (NOTE-1180) in full, from the typeset article on Griffiths's lab page at Princeton; worth reading for the author–topic model as the plain-text case of members' overlapping participation in themes, for its factorization of the document–word matrix as ΦΘA (THEORY-972), and for the measured trade between pooling by author and per-document flexibility (THEORY-1003). Read it knowing that authorship is observed, never inferred, that the number of topics and the hyperparameters are fixed, that an author's topics are proportions summing to one, that identifiability is claimed only up to permutation when the likelihood allows a continuum, and that its readings of Enron employees' roles are by inspection.
Version2
Tagsprobabilistic-modeling, information-retrieval
Cited as “paper” byNOTE-1180 — Rosen-Zvi, Chemudugunta, Griffiths, Smyth and Steyvers — Learning author-topic models from text corpora
Cited as “source” byTHEORY-1003 — On NIPS papers (1,557 training and 183 test documents, 100 topics, fixed hyperparameters), the author–topic model gave test documents lower perplexity than LDA while sixteen or fewer of their words had been observed and higher perplexity beyond that; one fictitious author per document improved on the author–topic model throughout and lost to LDA after about 64 words, and an author model without topics did worst, THEORY-972 — In the author–topic model the likelihood of a corpus depends on the topics Φ and the authors' topic distributions Θ only through the document–word matrix ΦΘA, A being fixed by the observed authorship, so LDA (each document its own author) and the topic-free author model (each author one topic) are special cases; and the likelihood leaves Φ and Θ unidentified not only up to the permutations of topics the paper names but along a continuum of changes of topic basis whenever every topic gives every word positive probability

Graph View

Rosen-Zvi, M., Chemudugunta, C., Griffiths, T., Smyth, P. & Steyvers, M. (2010). Learning author-topic models from text corpora. ACM Transactions on Information Systems 28(1): 1–38. doi:10.1145/1658377.1658381.

Read in full on 2026-10-26 from the typeset copy on Griffiths’s lab page (NOTE-1180). The summary is checked against the reading.

Why this is here

Cited at T21 for the second component of the reply’s combined model: which employees take part in which interpretive themes, with overlapping memberships. The record held the author–topic model only as the baseline of NOTE-008 (LIT-013), in its nonparametric form. It bears on QUESTION-276, as the plain-text case of inferring members’ overlapping participation from what they write.

Backlinks