ICAR: Image-Based Complementary Auto Reasoning

Authors: Xijun Wang, Anqi Liang, Junbang Liang, Ming Lin, Yu Lou, Shan Yang

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

Reproducibility Variable Result LLM Response
Research Type Experimental Compared with the SOTA methods, this approach achieves up to 5.3% and 9.6% in FITB score and 22.3% and 31.8% SFID improvement on fashion and furniture, respectively.
Researcher Affiliation Collaboration Xijun Wang1,2, Anqi Liang2, Junbang Liang2, Ming Lin1,2, Yu Lou2, Shan Yang2 1 University of Maryland, College Park, USA 2 Amazon, USA
Pseudocode No The paper does not contain structured pseudocode or an algorithm block.
Open Source Code No The paper does not provide an explicit statement or link to open-source code for the described methodology.
Open Datasets Yes Benchmark Datasets: In the following experiments, we evaluate our proposed ICAR using four datasets. Deep Rooms (Gadde, Feng, and Martinez 2021)... STL-Home (Kang et al. 2019)... STL-Fashion (Kang et al. 2019)... And Exact Street2Shop (Hadi Kiapour et al. 2015)...
Dataset Splits No The paper mentions models are 'trained for 500 epochs' but does not specify the size, percentage, or specific creation method for training, validation, or test splits. It mentions 'test-split' for some datasets but no details for validation.
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models used for running the experiments.
Software Dependencies No The paper mentions using 'Adam W' and 'CNN-based (Res Net50)' but does not list specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes Models are trained for 500 epochs with a batch size of 256. We choose 1 negative sample in the triplet loss. And we use 1.0, 1.0, and 0.05 as the weights for crossentropy loss, triplet loss, and regularizer loss respectively.