Continual Compositional Zero-Shot Learning

Authors: Yang Zhang, Songhe Feng, Jiazheng Yuan

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

Reproducibility Variable Result LLM Response
Research Type Experimental We design the CCZSL evaluation protocol and conduct extensive experiments on widely used benchmarks, demonstrating the superiority of our method compared to the state-of-the-art CZSL methods.
Researcher Affiliation Academia 1School of Computer Science and Technology, Beijing Jiaotong University, Beijing, China 2Tangshan Research Institute, Beijing Jiaotong University, Beijing, China 3College of Science and Technology, Beijing Open University, Beijing, China {23111124, shfeng}@bjtu.edu.cn, jzyuan@139.com
Pseudocode Yes Algorithm 1 The optimization procedure of CCZSL.
Open Source Code No The paper does not contain an explicit statement about releasing source code or a link to a code repository.
Open Datasets Yes We conduct experiments on two widely adopted datasets in CZSL, which are UT-Zappos [Yu and Grauman, 2014] and C-GQA [Naeem et al., 2021].
Dataset Splits Yes Specifically, We split UT-Zappos into 3 sessions and C-GQA into 6 sessions. Details of splits are present in Table 2, including number of new attributes, new objects, new training compositions, new validation compositions and new testing compositions in each session.
Hardware Specification No The paper mentions using a ResNet-18 pre-trained on ImageNet but does not specify any hardware used for training or inference (e.g., GPU model, CPU type).
Software Dependencies No The paper mentions using the Adam optimizer, but it does not specify any software libraries or their version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes The number of super-primitives K is set to 4 for UT-Zappos and 20 for C-GQA. The original primitive embeddings are learned from scratch. For all datasets, we train the model using Adam optimizer with a learning rate of 5e-5. The temperature factor is 0.05 for all datasets.