Continual Knowledge Consolidation LORA for Domain Incremental Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Paeedeh, Naeem, Pratama, Mahardhika, Ding, Weiping, Cao, Jimmy, Mayer, Wolfgang, Kowalczyk, Ryszard
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914100825554944
author Paeedeh, Naeem
Pratama, Mahardhika
Ding, Weiping
Cao, Jimmy
Mayer, Wolfgang
Kowalczyk, Ryszard
author_facet Paeedeh, Naeem
Pratama, Mahardhika
Ding, Weiping
Cao, Jimmy
Mayer, Wolfgang
Kowalczyk, Ryszard
contents Domain Incremental Learning (DIL) is a continual learning sub-branch that aims to address never-ending arrivals of new domains without catastrophic forgetting problems. Despite the advent of parameter-efficient fine-tuning (PEFT) approaches, existing works create task-specific LoRAs overlooking shared knowledge across tasks. Inaccurate selection of task-specific LORAs during inference results in significant drops in accuracy, while existing works rely on linear or prototype-based classifiers, which have suboptimal generalization powers. Our paper proposes continual knowledge consolidation low rank adaptation (CONEC-LoRA) addressing the DIL problems. CONEC-LoRA is developed from consolidations between task-shared LORA to extract common knowledge and task-specific LORA to embrace domain-specific knowledge. Unlike existing approaches, CONEC-LoRA integrates the concept of a stochastic classifier whose parameters are sampled from a distribution, thus enhancing the likelihood of correct classifications. Last but not least, an auxiliary network is deployed to optimally predict the task-specific LoRAs for inferences and implements the concept of a different-depth network structure in which every layer is connected with a local classifier to take advantage of intermediate representations. This module integrates the ball-generator loss and transformation module to address the synthetic sample bias problem. Our rigorous experiments demonstrate the advantage of CONEC-LoRA over prior arts in 4 popular benchmark problems with over 5% margins.
format Preprint
id arxiv_https___arxiv_org_abs_2510_16077
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Continual Knowledge Consolidation LORA for Domain Incremental Learning
Paeedeh, Naeem
Pratama, Mahardhika
Ding, Weiping
Cao, Jimmy
Mayer, Wolfgang
Kowalczyk, Ryszard
Machine Learning
Artificial Intelligence
Domain Incremental Learning (DIL) is a continual learning sub-branch that aims to address never-ending arrivals of new domains without catastrophic forgetting problems. Despite the advent of parameter-efficient fine-tuning (PEFT) approaches, existing works create task-specific LoRAs overlooking shared knowledge across tasks. Inaccurate selection of task-specific LORAs during inference results in significant drops in accuracy, while existing works rely on linear or prototype-based classifiers, which have suboptimal generalization powers. Our paper proposes continual knowledge consolidation low rank adaptation (CONEC-LoRA) addressing the DIL problems. CONEC-LoRA is developed from consolidations between task-shared LORA to extract common knowledge and task-specific LORA to embrace domain-specific knowledge. Unlike existing approaches, CONEC-LoRA integrates the concept of a stochastic classifier whose parameters are sampled from a distribution, thus enhancing the likelihood of correct classifications. Last but not least, an auxiliary network is deployed to optimally predict the task-specific LoRAs for inferences and implements the concept of a different-depth network structure in which every layer is connected with a local classifier to take advantage of intermediate representations. This module integrates the ball-generator loss and transformation module to address the synthetic sample bias problem. Our rigorous experiments demonstrate the advantage of CONEC-LoRA over prior arts in 4 popular benchmark problems with over 5% margins.
title Continual Knowledge Consolidation LORA for Domain Incremental Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2510.16077