Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
MIMAMO Net: Integrating Micro- and Macro-Motion for Video Emotion Recognition
Authors: Didan Deng, Zhaokang Chen, Yuqian Zhou, Bertram Shi2621-2628
AAAI 2020 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our proposed network achieves state of the art performance on two video emotion datasets, the OMG emotion dataset and the Aff-Wild dataset. |
| Researcher Affiliation | Academia | Didan Deng,1 Zhaokang Chen,1 Yuqian Zhou,2 Bertram Shi1 1Neuromorphic Interactive System Laboratory, Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Kowloon, HK 2Image Formation and Processing Group, University of Illinois at Urbana-Champaign , Champaign, IL, USA EMAIL, EMAIL |
| Pseudocode | No | The paper describes the proposed model and method in text and with a diagram, but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Source code is available at https://github.com/wtomin/MIMAMO-Net. |
| Open Datasets | Yes | Our proposed network achieves state of the art performance on two video emotion datasets, the OMG emotion dataset and the Aff-Wild dataset. |
| Dataset Splits | Yes | The number of utterances in the training, validation and test sets are 2441, 617 and 2229 respectively. |
| Hardware Specification | No | Batch size varied among [16, 32, 64, 128, 256, 512]. Because different experiments had different GPU consumption, we chose the largest batch size that fits in our GPU memory (11GB). This indicates GPU memory capacity but does not specify the GPU model or CPU. |
| Software Dependencies | No | During training, we used stochastic gradient descent optimizer implemented in Py Torch, with momentum (0.9) and weight decay (5e 4). It mentions 'Py Torch' but does not provide a specific version number. |
| Experiment Setup | Yes | The number of epochs was set to be 25. Early stopping (5 epochs) was used to prevent overfitting. Batch size varied among [16, 32, 64, 128, 256, 512]. During training, we used stochastic gradient descent optimizer implemented in Py Torch, with momentum (0.9) and weight decay (5e 4). |