Recommendation-system A/B tests can systematically underestimate a new algorithm's true deployment effect because a small treatment group cannot trigger the full ecosystem's feedback loops, a bias the paper formalizes and illustrates.
Gaps in the Thue--Morse word
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
The Thue--Morse sequence is a prototypical automatic sequence found in diverse areas of mathematics, and in computer science. We study occurrences of factors $w$ within this sequence, more precisely, the sequence of gaps between consecutive occurrences. This gap sequence is morphic; we prove that it is not automatic as soon as the length of $w$ is at least two, thereby answering a question by J.~Shallit in the affirmative. We give an explicit method to compute the \emph{discrepancy} of the number of occurrences of the block $\mathtt{01}$ in the Thue--Morse sequence. We prove that the sequence of discrepancies is the sequence of output sums of a certain base-$2$ transducer.
citation-role summary
background 1
citation-polarity summary
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Algorithm Adaptation Bias in Recommendation System Online Experiments
Recommendation-system A/B tests can systematically underestimate a new algorithm's true deployment effect because a small treatment group cannot trigger the full ecosystem's feedback loops, a bias the paper formalizes and illustrates.