CMAP uses text-space task routing with frozen CLIP prototypes, multi-prototype cross-modal confidence, and symmetric gating to achieve new state-of-the-art results on the MTIL benchmark with 2.5M parameters.
Con- tinual learning with pre-trained models: A survey
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
HC-SOINN with STAR captures topological manifold structure in class features and aligns it to non-linear drift, improving over point-wise NCM when integrated into existing CIL methods.
Collaborative Parameter Learning freezes 50-75% of parameters whose updates cause forgetting and updates only the 25-50% that mitigate it, allowing LLMs to learn 20-48% more new questions with negligible forgetting and lower compute cost.
iGSP uses implicit gradient subspace projection in two phases to enable efficient continual adaptation of vision-language models, claiming SOTA accuracy with 42.7% fewer trainable parameters and 86.9% less total parameter growth.
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.
C-Flat Turbo accelerates continual learning by skipping redundant flatness gradients via direction-invariance observations and linear adaptive scheduling, delivering 1-1.25x speedup with comparable accuracy.
Token Factory maps dense, sparse, and sequence features into learned soft tokens, cutting LRM prompt length by ~70% while keeping ranking quality on par and improving retrieval of fresh videos.
Proposes a frugal functional representation for star-shaped image objects to analyze contours and textures, illustrated on supervised classification.
SoTU merges sparse orthogonal delta parameters learned across streaming tasks to fuse knowledge and mitigate forgetting in pre-trained model continual learning.
citing papers explorer
-
CMAP: Cross-Modal Adaptive Prompting for Multi-Domain Task-Incremental Learning
CMAP uses text-space task routing with frozen CLIP prototypes, multi-prototype cross-modal confidence, and symmetric gating to achieve new state-of-the-art results on the MTIL benchmark with 2.5M parameters.
-
Beyond Point-wise Neural Collapse: A Topology-Aware Hierarchical Classifier for Class-Incremental Learning
HC-SOINN with STAR captures topological manifold structure in class features and aligns it to non-linear drift, improving over point-wise NCM when integrated into existing CIL methods.
-
Collaborative Parameter Learning: Mitigating Forgetting via Parameter-Level Gradient Analysis
Collaborative Parameter Learning freezes 50-75% of parameters whose updates cause forgetting and updates only the 25-50% that mitigate it, allowing LLMs to learn 20-48% more new questions with negligible forgetting and lower compute cost.
-
iGSP:Implicit Gradient Subspace Projection for Efficient Continual Learning of Vision-Language Models
iGSP uses implicit gradient subspace projection in two phases to enable efficient continual adaptation of vision-language models, claiming SOTA accuracy with 42.7% fewer trainable parameters and 86.9% less total parameter growth.
-
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.
-
A Faster Path to Continual Learning
C-Flat Turbo accelerates continual learning by skipping redundant flatness gradients via direction-invariance observations and linear adaptive scheduling, delivering 1-1.25x speedup with comparable accuracy.
-
Token Factory: Efficiently Integrating Diverse Signals into Large Recommendation Models
Token Factory maps dense, sparse, and sequence features into learned soft tokens, cutting LRM prompt length by ~70% while keeping ranking quality on par and improving retrieval of fresh videos.
-
A Functional Data Framework For Analyzing Shapes and Textures in Images
Proposes a frugal functional representation for star-shaped image objects to analyze contours and textures, illustrated on supervised classification.
-
Sparse Orthogonal Parameters Tuning for Continual Learning
SoTU merges sparse orthogonal delta parameters learned across streaming tasks to fuse knowledge and mitigate forgetting in pre-trained model continual learning.