Dynamically Anchored Prompting for Task-Imbalanced Continual Learning

Authors: Chenxing Hong, Yan Jin, Zhiqi Kang, Yizhou Chen, Mengke Li, Yang Lu, Hanzi Wang

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate that the proposed DAP results in 4.5% to 15% absolute improvements over state-of-the-art methods on benchmarks under task-imbalanced settings.
Researcher Affiliation Collaboration 1Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University, Xiamen, China 2Fujian Key Laboratory of Sensing and Computing for Smart City, School of Informatics, Xiamen University, Xiamen, China 3Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ), Shenzhen, China 4Univ. Grenoble Alpes, Inria, CNRS, Grenoble INP, LJK, 38000 Grenoble, France.
Pseudocode No The paper describes the method and its mathematical formulations but does not include a structured pseudocode or algorithm block.
Open Source Code Yes Our code is available at https://github.com/chenxing6666/DAP.
Open Datasets Yes Specifically, the CIFAR-100 dataset [Krizhevsky et al., 2009] includes 100 classes of natural images... The Image Net-R dataset [Wang et al., 2022d] contains 200 classes of images...
Dataset Splits No The paper mentions training and testing datasets but does not provide specific details about a validation dataset split or how it was used.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions using 'Adam' as an optimizer but does not list specific software dependencies with version numbers (e.g., Python, PyTorch, CUDA versions).
Experiment Setup Yes We train DAP using Adam with β1, β2 of 0.9, a learning rate of 0.01, and a batch size of 64. We resize the input images to a 224 224 resolution and normalize them between 0 and 1. To ensure models converge, we train TICL-CIFAR-100 for 5 epochs per task, TICL-Image Net-R for 50 epochs each task.