Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
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.
Joint Attribute and Model Generalization Learning for Privacy-Preserving Action Recognition
Authors: Duo Peng, Li Xu, Qiuhong Ke, Ping Hu, Jun Liu
NeurIPS 2023 | Venue PDF | LLM Run Details | Input Tokens: 19,992 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 5,433 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments demonstrate the effectiveness and generalization of the proposed framework compared to state-of-the-arts. |
| Researcher Affiliation | Academia | Duo Peng SUTD Singapore EMAIL Li Xu SUTD Singapore EMAIL Qiuhong Ke Monash University Australia EMAIL Ping Hu UESTC China EMAIL Jun Liu SUTD Singapore EMAIL |
| Pseudocode | Yes | Algorithm 1: Overall Training Scheme |
| Open Source Code | No | The paper does not provide an explicit statement or link for the open-sourcing of the described methodology's code. |
| Open Datasets | Yes | We conduct experiments using two benchmarks. The first benchmark, HMDB51-VISPR, is comprised of HMDB51 [31] dataset and VISPR [30] dataset. The second benchmark, UCF101-VISPR, consists of UCF101 [29] dataset and VISPR [30] dataset. |
| Dataset Splits | Yes | Specifically, we first construct a support set for virtual training, and a query set for virtual testing. ... On each benchmark, we construct the support set with the videos containing 60% of the privacy attributes in the training data Xtrain, and use the remaining training data to construct the query set. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions various models used (e.g., Image Transformation model, C3D, Mobile Net-V2, UNet, R3D-18, Res Net-50) but does not specify software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | We set γ (in Eq. 1) as 0.4, the learning rate α for virtual training (in Eq. 5) as 5e 4, and the learning rate β for meta-optimization (in Eq. 8) as 1e 4. |