AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Generalized Relevance Learning Grassmann Quantization

Mohammad Mohammadi, Mohammad Babai, Michael H. F. Wilkinson

TPAMI 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory

Read the paper

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 GRLGQ subspace prototypes instead of GTLVQGLVQ classifiers · MNISTaccuracy: 94.50 -> 94.78Improved
use GRLGQ instead of GTLVQGLVQ classifiers · Yale Face Databaseaccuracy: 94.64 -> 96.50Improved
use GRLGQ instead of GGPLCRimage-set classifier · Ex. Yale Baccuracy: 88.57 -> 90.12Improved
use GRLGQ instead of GGPLCRimage-set classifier · ETH-80accuracy: 96.75 -> 94.5Hurt
use GRLGQ instead of AMLSLimage-set classifier · UCF Sportsaccuracy: 74.00 -> 80.00Improved
add relevance factors over principal angles (ablation)GRLGQ · ETH-80accuracy vs subspace dimension: other methods drop below 90% when d exceeds 3Improved
increase subspace dimensionality with relevance factors (ablation)GRLGQ · Ex. Yale BaccuracyNo clear effect
use prototype strategy instead of storing training subspacesGRLGQ vs G(G)PLCR · ETH-80 parameter countnumber of parameters: 60000 -> 16,005Improved
use GRLGQ instead of G(G)PLCRparameter count · Ex. Yale Bnumber of parameters: 134400 -> 280000Hurt

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