A causal view of compositional zero-shot recognition

Authors: Yuval Atzmon, Felix Kreuk, Uri Shalit, Gal Chechik

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate this approach on two datasets for predicting new combinations of attribute-object pairs: A well-controlled synthesized images dataset and a real-world dataset which consists of fine-grained types of shoes. We show improvements compared to strong baselines. Code and data are provided in https://github.com/nv-research-israel/causal_comp
Researcher Affiliation Collaboration 1NVIDIA Research, Tel Aviv, Israel 2Bar-Ilan University, Ramat Gan, Israel 3Technion Israel Institute of Technology
Pseudocode No The paper describes the model and its components, including loss functions, but does not provide any formal pseudocode or algorithm blocks.
Open Source Code Yes Code and data are provided in https://github.com/nv-research-israel/causal_comp
Open Datasets Yes We evaluate our approach on the Zappos dataset, which consists of fine-grained types of shoes... We use the split of [50] and the provided ResNet18 pretrained features... To evaluate compositional methods on a well-controlled clean dataset, we generated a synthetic-images dataset containing images of easy Attribute-Object categories. We used the CLEVr framework [25], hence we name the dataset AO-CLEVr.
Dataset Splits Yes It uses 23K images for training of 83 seen pairs, a validation set with 3K images from 15 seen and 15 unseen pairs, and a test set with 3K images from 18 seen and 18 unseen pairs. All the metrics we report for our approach and compared baselines are averaged over 5 random initializations of the model. For cross-validation, we used two types of splits. The first uses the same unseen pairs for validation and test. Importantly, we vary the ratio of unseen:seen pairs on a range of (2:8, 3:7, ...,7:3), and for each ratio we draw 3 random seen-unseen splits.
Hardware Specification Yes All experiments were performed on a cluster of DGX-V100 machines.
Software Dependencies No The paper mentions using MLPs and the Adam optimizer [30], but it does not specify versions for any programming languages or libraries (e.g., Python, PyTorch, TensorFlow).
Experiment Setup Yes We aim to learn the parameters of these mappings such that the (approximated) negative log-likelihood of Eq. (3) is minimized. In addition, we also include in the objective several regularization terms designed to encourage properties that we want to induce on these mappings. Specifically, the model is trained with a linear combination of three losses. L = Ldata + λindep Lindep + λinvert Linvert, where λindep 0 and λinvert 0 are hyperparameters.