AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

RD-Suite: A Benchmark for Ranking Distillation

Zhen Qin, Rolf Jagerman, Rama Kumar Pasumarthi, Honglei Zhuang, He Zhang, Aijun Bai, Kai Hui, Le Yan, Xuanhui Wang

NeurIPS 2023 · Label-efficient and robust learning › Knowledge distillation · approach: Self and data-free distillation · 1 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
turn softmax teacher-score transform off for MSE (ablation)MSE distillation · NQMRR@10: 54.87 -> 58.49Improved
turn softmax teacher-score transform off for MSE (ablation)MSE distillation · MSMARCOMRR@10: 40.03 -> 42.06Improved
turn softmax teacher-score transform off for PairMSE (ablation)PairMSE · NQMRR@10: 55.61 -> 58.34Improved
turn softmax transform on for Softmax loss (ablation)Softmax listwise distillation · NQMRR@10: 52.67 -> 59.08Improved
turn softmax transform on for GumbelNDCG (ablation)GumbelNDCG · NQMRR@10: 53.06 -> 57.90Improved
turn softmax transform on for LambdaLoss (ablation)LambdaLoss · NQMRR@10: 56.52 -> 58.36Improved
add RankDistil distillation to relevance-only trainingBERT ranker · MSMARCOMRR@10: 40.03 -> 41.29Improved
add RD distillation to relevance-only trainingBERT ranker · MSMARCOMRR@10: 40.03 -> 40.25No clear effect
add Softmax listwise distillation to relevance-only trainingBERT ranker · NQMRR@10: 52.67 -> 59.08Improved
distill from a weak MSMARCO teacher into NQSoftmax distillation · NQ transferMRR@10: 52.67 -> 56.72Improved
add RD distillation on tabular Web30KMLP ranker · Web30KMRR@10: 27.81 -> 27.44No clear effect
add Softmax distillation on Web30KMLP ranker · Web30KMRR@10: 27.81 -> 29.54Improved

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