A Self-explaining Neural Architecture for Generalizable Concept Learning
Authors: Sanchit Sinha, Guangzhi Xiong, Aidong Zhang
IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirical results show that our method improves both concept fidelity measured through concept overlap and concept interoperability measured through domain adaptation performance. |
| Researcher Affiliation | Academia | Sanchit Sinha1 , Guangzhi Xiong1 and Aidong Zhang1 1University of Virginia Charlottesville, VA, USA {sanchit, hhu4zu, aidong}@virginia.edu |
| Pseudocode | No | The paper does not contain any sections or figures explicitly labeled as 'Pseudocode' or 'Algorithm', nor does it present structured steps in a code-like format. |
| Open Source Code | No | The paper mentions 'An appendix of the paper with more comprehensive results can also be viewed at https://arxiv.org/abs/2405.00349', which is a link to the paper itself, not specific source code. Footnote 1 links to a third-party library 'lightly' on GitHub, but this is not the authors' own source code for their methodology. |
| Open Datasets | Yes | Digits: This setting utilizes MNIST and USPS [Le Cun et al., 1998; Hull, 1994] with Hand-written images of digits and Street View House Number Dataset (SVHN) [Netzer et al., 2011] with cropped house number photos. Vis DA-2017 [Peng et al., 2017]: contains 12 classes of vehicles sampled from Real (R) and 3D domains. Domain Net [Venkateswara et al., 2017]: contains 126 classes of objects (clocks, bags, etc.) sampled from 4 domains Real (R), Clipart (C), Painting (P), and Sketch (S). Office-Home [Peng et al., 2019]: Office-Home contains 65 classes of office objects like calculators, staplers, etc. sampled from 4 different domains Art (A), Clipart (C), Product (P), and Real (R). |
| Dataset Splits | No | The paper mentions 'We train each dataset for 10000 iterations with early stopping', which implies the use of a validation set. However, it does not specify the exact percentages or counts for the training/validation splits, nor does it refer to predefined splits with specific details for reproducibility. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for the experiments, such as GPU models, CPU specifications, or memory. It only discusses the network architectures (LeNet, ResNet34) and software libraries used. |
| Software Dependencies | No | The paper mentions 'We utilize the lightly1 library for implementing Sim CLR transformations [Chen, 2020]' and provides a GitHub link for 'lightly' in a footnote. However, it does not specify a version number for the 'lightly' library or any other software dependencies, which is required for reproducibility. |
| Experiment Setup | Yes | RCE Framework: We utilize the Mean Square Error as the reconstruction loss and set sparsity regularizer λ to 1e-5 for all datasets. The weights ω1 = ω2 = 0.5 are utilized for digit, while they are set at ω1 = 0.8 and ω2 = 0.2 for object tasks. [...] Learning: We utilize the lightly1 library for implementing Sim CLR transformations [Chen, 2020]. We set the temperature parameter (τ) to 0.5 by default [Xu et al., 2019] for all datasets. The training objective is Contrastive Cross Entropy (NTXent) [Chen, 2020]. [...] For the training procedure, we utilize the SGD optimizer with momentum set to 0.9 and a cosine decay scheduler with an initial learning rate set to 0.01. We train each dataset for 10000 iterations with early stopping. The regularization parameters of λ1 and λ2 are set to 0.1 respectively. For Digits, β is set to 1 while it is set to 0.5 for objects. |