paper-with-me

홈 › Papers

Soft-TransFormers for Continual Learning

2024-11-25 · Haeyong Kang, Chang D. Yoo

Inspired by Well-initialized Lottery Ticket Hypothesis (WLTH), which provides suboptimal fine-tuning solutions, we propose a novel fully fine-tuned continual learning (CL) method referred to as Soft-TransFormers (Soft-TF). Soft-TF sequentially learns and selects an optimal soft-network or subnetwork for each task. During sequential training in CL, Soft-TF jointly optimizes the weights of sparse layers to obtain task-adaptive soft (real-valued) networks or subnetworks (binary masks), while keeping the well-pre-trained layer parameters frozen. In inference, the identified task-adaptive network of Soft-TF masks the parameters of the pre-trained network, mapping to an optimal solution for each task and minimizing Catastrophic Forgetting (CF) - the soft-masking preserves the knowledge of the pre-trained network. Extensive experiments on Vision Transformer (ViT) and CLIP demonstrate the effectiveness of Soft-TF, achieving state-of-the-art performance across various CL scenarios, including Class-Incremental Learning (CIL) and Task-Incremental Learning (TIL), supported by convergence theory.

📄 PDF Abstract BibTeX arXiv:2411.16073

Code (1)

ihaeyong/soft-tf 공식 구현 tf

Tasks

class-incremental learningClass Incremental LearningContinual LearningIncremental Learning

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Adam 설명 없음
Residual Connection 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Similar Papers 제목 키워드 기반

Lifelong In-Context Learning with Transformers Requires Parametric Forms of Attention

2026-06-24 · Luke McDermott, Robert W. Heath, Rahul Parhi arxiv

Lifelong continual learning remains an obstacle on the path to human-like intelligence. Modern transformers show sparks of intelligence with in-context learning. The quadratic nature of attention, however, prohibits tran…

Continual Learning

Parameter Efficient Continual Learning for Sparse Event-Based Transformers

2026-08-27 · Vaishnavi Nagabhushana, Kartikay Agrawal, Ayon Borthakur arxiv

Robotic and edge intelligence systems operate in dynamic environments where data arrives continuously, requiring models to adapt while preserving previously learned knowledge under strict memory and energy constraints. W…

parameter-efficient fine-tuningclass-incremental learningContinual LearningEvent-based vision

Continual Transformers: Redundancy-Free Attention for Online Inference

2022-01-17 · Lukas Hedegaard, Arian Bakhtiarnia, Alexandros Iosifidis

Transformers in their common form are inherently limited to operate on whole token sequences rather than on one token at a time. Consequently, their use during online inference on time-series data entails considerable re…

Action DetectionAudio ClassificationClassificationOnline Action Detection+2

Learning Mamba as a Continual Learner

2024-12-01 · Chongyang Zhao, Dong Gong

Continual learning (CL) aims to efficiently learn and accumulate knowledge from a data stream with different distributions. By formulating CL as a sequence prediction task, meta-continual learning (MCL) enables to meta-l…

Continual LearningMambaState Space Models

Technical Report for ICCV 2021 Challenge SSLAD-Track3B: Transformers Are Better Continual Learners

2022-01-13 · Duo Li, Guimei Cao, Yunlu Xu, Zhanzhan Cheng 외

In the SSLAD-Track 3B challenge on continual learning, we propose the method of COntinual Learning with Transformer (COLT). We find that transformers suffer less from catastrophic forgetting compared to convolutional neu…

Continual LearningKnowledge Distillationobject-detectionObject Detection