AI Paper Atlas

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

Direction Sensitivity–Based Knowledge Distillation: Optimization-Aware Low-Rank Knowledge Transfer

Yongkai Liao, Xinxing Chen, Zhongzheng Fu, Haoyuan Wang, Jian Huang

AAAI 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Distillation with causal and adaptive guidance · 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
add DSKD direction-sensitive selectionBERT-of-Theseus · GLUE (average)average score: 77.7 -> 79.2Improved
add DSKDLGTM · GLUE (average)average score: 79.3 -> 79.8Improved
add DSKDDBKD · GLUE (average)average score: 72.5 -> 73.7Improved
add DSKDAD-KD · GLUE (average)average score: 79.8 -> 80.4Improved
add DSKDAD-KD · GLUE RTEaccuracy: 67.5 -> 66.9Hurt
use sensitivity-based directions instead of top singular values (SVD) (ablation)BERT-of-Theseus · GLUE RTEaccuracy: 65.7 -> 67.6Improved
select directions randomly instead of by sensitivity (ablation)BERT-of-Theseus · GLUE RTEaccuracy: 67.6 -> 63.5Hurt
combine first- and second-order sensitivity instead of first-order only (ablation)DSKD · GLUE RTEaccuracy: 67.2 -> 67.6Improved
combine first- and second-order sensitivity instead of second-order only (ablation)DSKD · GLUE RTEaccuracy: 66.8 -> 67.6Improved
use very large distillation weight lambda=10 (ablation)DSKD · GLUE RTEaccuracy: 67.6 -> 64.7Hurt
tune lambda between 0 and 1 versus lambda=0 (ablation)DSKD · GLUE CoLAaccuracy: +5.8% on CoLAImproved
update directions every Ts=100 steps instead of Ts=1 (ablation)DSKD · GLUE RTEaccuracy: 64.7 -> 67.6Improved

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