AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

A Spectral Approach to Item Response Theory

Duc Nguyen, Anderson Ye Zhang

NeurIPS 2022 · Learning theory and algorithms › Prediction and decision losses · approach: Ranking models from comparisons · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use CMLE instead of the spectral/MMLE/JMLE estimators on large rating data with missing entriesCMLE · BXheld-out AUC: 0.57 -> 0.503Hurt
replace MMLE with the spectral method on BXSpectral · BXheld-out AUC: 0.577 -> 0.546Hurt
use spectral method instead of MMLE for top-K ranking on ML-1MSpectral · ML-1Mtop-10 accuracy: 0.6 -> 0.8Improved
use spectral method instead of JMLE for top-K ranking on Book-GenomeSpectral · Book-Genometop-10 accuracy: 0.2 -> 0.6Improved
spectral method versus MMLE/JMLE on large rating data, AUC and log-likelihoodSpectral · Book-Genomeheld-out AUC: 0.665 vs 0.658No clear effect
use spectral method instead of MMLE for inference time on ML-20MSpectral · ML-20Mtotal inference time (s): 5.36k -> 1.58kImproved
use spectral method instead of CMLE for inference time on BXSpectral · BXtotal inference time (s): 156k -> 205Improved
use spectral method on small, complete education datasets (LSAT, UCI, 3 GRADES)Spectral · LSATheld-out AUC: 0.707 vs 0.707No clear effect
misspecify MMLE prior (sigma=2 instead of 1) in synthetic dataMMLE · Synthetic Rasch datal2 error of betaHurt
small number of items m relative to nJMLE · Synthetic Rasch datal2 error of betaHurt
moderately sized dataset with missing dataCMLE · Synthetic Rasch datal2 error of betaHurt
use accelerated (modified Markov chain) version instead of original spectral algorithm (ablation)Spectral · Real-life datasetsconvergence speed / numerical issuesImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy