Introduces a topological regularization framework for NMF that uses persistent homology to enforce desired structures in basis functions within a unified optimization objective.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
RSD fits shared three-anchor charts S_t to GPT-2 hidden states for target words, derives co-membership readouts M_t, and audits against WiC same-sense labels, passing 16 of 53 words as diagnostic coverage.
GNRBMF extends NRBMF by adding graph Laplacian regularization on the coefficient matrix to encourage similar representations for nearby samples while retaining non-negativity in the reduced biquaternion domain.
citing papers explorer
-
Non-negative Matrix Factorisation with Topological Regularisation
Introduces a topological regularization framework for NMF that uses persistent homology to enforce desired structures in basis functions within a unified optimization objective.
-
RSD: Moving Local Triangular Charts for Auditing Language-Model Hidden States
RSD fits shared three-anchor charts S_t to GPT-2 hidden states for target words, derives co-membership readouts M_t, and audits against WiC same-sense labels, passing 16 of 53 words as diagnostic coverage.
-
Graph Regularized Non-negative Reduced Biquaternion Matrix Factorization for Color Image Recognition
GNRBMF extends NRBMF by adding graph Laplacian regularization on the coefficient matrix to encourage similar representations for nearby samples while retaining non-negativity in the reduced biquaternion domain.