ECMRNet is a continual-learning restoration network that decomposes features into isolated groups, expands new groups for novel degradations, prunes via structural entropy, and mines historical components for compound degradations in open-world TIR imaging.
IEEE transactions on pattern analysis and machine intelligence , volume=
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
years
2026 6representative citing papers
MixSD mixes tokens from the base model's expert and naive conditionals to create distribution-aligned supervision for knowledge injection, yielding better memorization-retention trade-offs than SFT across scales and benchmarks.
Annotation-anchored training reduces semantic diversity collapse in post-trained language models by a factor of six compared to standard supervised fine-tuning while preserving instruction-following and improving with scale.
Different valid temporal partitions of the same streaming dataset can produce materially different rankings and performance numbers for continual learning methods.
STAR-IOD applies scale-decoupled topology alignment and K-Means-based pseudo-label refinement to reduce catastrophic forgetting in remote sensing incremental object detection, reporting 1.7% and 2.1% mAP gains on new DIOR-IOD and DOTA-IOD datasets.
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
citing papers explorer
-
Expandable, Compressible, Mineable: Open-World Thermal Image Restoration
ECMRNet is a continual-learning restoration network that decomposes features into isolated groups, expands new groups for novel degradations, prunes via structural entropy, and mines historical components for compound degradations in open-world TIR imaging.
-
MixSD: Mixed Contextual Self-Distillation for Knowledge Injection
MixSD mixes tokens from the base model's expert and naive conditionals to create distribution-aligned supervision for knowledge injection, yielding better memorization-retention trade-offs than SFT across scales and benchmarks.
-
Annotations Mitigate Post-Training Mode Collapse
Annotation-anchored training reduces semantic diversity collapse in post-trained language models by a factor of six compared to standard supervised fine-tuning while preserving instruction-following and improving with scale.
-
Temporal Taskification in Streaming Continual Learning: A Source of Evaluation Instability
Different valid temporal partitions of the same streaming dataset can produce materially different rankings and performance numbers for continual learning methods.
-
STAR-IOD: Scale-decoupled Topology Alignment with Pseudo-label Refinement for Remote Sensing Incremental Object Detection
STAR-IOD applies scale-decoupled topology alignment and K-Means-based pseudo-label refinement to reduce catastrophic forgetting in remote sensing incremental object detection, reporting 1.7% and 2.1% mAP gains on new DIOR-IOD and DOTA-IOD datasets.
-
HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.