Selected work

Representative Papers

When does a network's training history predict its future learning better than its current state? Evidence from a response probe and a forecasting screen

Oct 07, 2026

This study investigates when the training history of neural networks better predicts future learning than their current state, a dimension frequently overlooked in research on plasticity loss and critical periods. Employing short-horizon response probing, synthetic regression-based predictor screening, and statistical consistency measures such as the intraclass correlation coefficient (ICC), this work systematically compares the predictive capacity of training history versus current states for small multilayer perceptrons. It provides the first quantification of the predictive gain offered by training history relative to current states, delineating its temporal boundaries. The findings reveal that historical information confers a significant predictive advantage only when the current state lacks informativeness, such as during early training stages, with no discernible difference observed later. Consequently, this work establishes the novel insight that training history becomes valuable precisely when the current state is uninformative.

0 citationsRead paper

Force without transmission: a depth-induced rank collapse that no loss on the representation reopens

Oct 07, 2026

This study addresses the issue of rank collapse during Transformer training, which causes learning stagnation that conventional loss terms fail to remedy. Through experiments on small-scale Transformers and gradient analysis, this work proposes a "gradient reachability" theory, revealing that the disruption of gradient propagation pathways is the primary cause of such repair failures. Accordingly, a dynamic skip connection regulation mechanism is designed to restart gradient flow. The authors demonstrate that relying solely on loss terms is ineffective, whereas restoring skip connections enables unsupervised rank recovery. By distinguishing the critical factors of network reparability along the dimensions of force magnitude and transmission pathways, this research offers a novel perspective for understanding and reversing rank collapse, even though post-recovery performance remains inferior to that of healthy models.

0 citationsRead paper

Criteria-first, semantics-later: reproducible structure discovery in image-based sciences

Feb 17, 2026

This work addresses the vulnerability of existing image analysis methods to label drift under semantic-priority paradigms, which compromises their reliability in open science, cross-sensor/cross-site comparability, and long-term monitoring. To overcome this limitation, the authors propose a novel “standard-first, semantics-later” paradigm that decouples structural discovery from semantic mapping. By leveraging cybernetics, the principle that observation entails distinction, and information theory, they formulate explicit optimization criteria to extract stable, semantics-agnostic structures—such as partitions, structural fields, or hierarchies—prior to aligning them with domain-specific ontologies. This framework ensures that structural outputs remain independent of labeling schemes, enabling multiple interpretations and long-term interoperability. Validated across diverse domains, the approach demonstrates broad applicability in scenarios where labels are non-scalable, thereby advancing structural findings as FAIR, AI-ready digital objects suitable for digital twins and continuous monitoring.

0 citationsRead paper

Recent publications

Latest Papers

When does a network's training history predict its future learning better than its current state? Evidence from a response probe and a forecasting screen

Oct 07, 2026

This study investigates when the training history of neural networks better predicts future learning than their current state, a dimension frequently overlooked in research on plasticity loss and critical periods. Employing short-horizon response probing, synthetic regression-based predictor screening, and statistical consistency measures such as the intraclass correlation coefficient (ICC), this work systematically compares the predictive capacity of training history versus current states for small multilayer perceptrons. It provides the first quantification of the predictive gain offered by training history relative to current states, delineating its temporal boundaries. The findings reveal that historical information confers a significant predictive advantage only when the current state lacks informativeness, such as during early training stages, with no discernible difference observed later. Consequently, this work establishes the novel insight that training history becomes valuable precisely when the current state is uninformative.

0 citationsRead paper

Force without transmission: a depth-induced rank collapse that no loss on the representation reopens

Oct 07, 2026

This study addresses the issue of rank collapse during Transformer training, which causes learning stagnation that conventional loss terms fail to remedy. Through experiments on small-scale Transformers and gradient analysis, this work proposes a "gradient reachability" theory, revealing that the disruption of gradient propagation pathways is the primary cause of such repair failures. Accordingly, a dynamic skip connection regulation mechanism is designed to restart gradient flow. The authors demonstrate that relying solely on loss terms is ineffective, whereas restoring skip connections enables unsupervised rank recovery. By distinguishing the critical factors of network reparability along the dimensions of force magnitude and transmission pathways, this research offers a novel perspective for understanding and reversing rank collapse, even though post-recovery performance remains inferior to that of healthy models.

0 citationsRead paper

Criteria-first, semantics-later: reproducible structure discovery in image-based sciences

Feb 17, 2026

This work addresses the vulnerability of existing image analysis methods to label drift under semantic-priority paradigms, which compromises their reliability in open science, cross-sensor/cross-site comparability, and long-term monitoring. To overcome this limitation, the authors propose a novel “standard-first, semantics-later” paradigm that decouples structural discovery from semantic mapping. By leveraging cybernetics, the principle that observation entails distinction, and information theory, they formulate explicit optimization criteria to extract stable, semantics-agnostic structures—such as partitions, structural fields, or hierarchies—prior to aligning them with domain-specific ontologies. This framework ensures that structural outputs remain independent of labeling schemes, enabling multiple interpretations and long-term interoperability. Validated across diverse domains, the approach demonstrates broad applicability in scenarios where labels are non-scalable, thereby advancing structural findings as FAIR, AI-ready digital objects suitable for digital twins and continuous monitoring.

0 citationsRead paper