Online Action Recognition
Authors: Alejandro Suárez-Hernández, Javier Segovia-Aguas, Carme Torras, Guillem Alenyà11981-11989
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We report results on benchmarks from the International Planning Competition and PDDLGym, where OARU recognizes actions accurately with respect to expert knowledge, and shows real-time performance. We have implemented and evaluated OARU in a benchmark of 9 domains (Muise 2016; Silver and Chitnis 2020). Table 1: Results with (a) full and (b) partial observability. T is in milliseconds, M in MB, and Prec. and Rec. in %. |
| Researcher Affiliation | Academia | 1IRI Institut de Rob otica i Inform atica Industrial, CSIC-UPC 2Universitat Pompeu Fabra |
| Pseudocode | Yes | Algorithm 1 OARU |
| Open Source Code | Yes | 3https://github.com/sprkrd/sat strips learn |
| Open Datasets | Yes | We have implemented3 and evaluated4 OARU in a benchmark of 9 domains (Muise 2016; Silver and Chitnis 2020). We report results on benchmarks from the International Planning Competition (IPC) (Muise 2016) and PDDLGym (Silver and Chitnis 2020) |
| Dataset Splits | No | The paper does not explicitly provide training/test/validation dataset splits, specific percentages, or sample counts, as its learning approach is incremental rather than based on fixed splits. |
| Hardware Specification | Yes | Machine specifications: AMD Ryzen 7 3700X @ 3.6GHz, 32GB of DDR4 RAM CL15 @ 3200MHz. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers) needed to replicate the experiment. |
| Experiment Setup | Yes | To perform AU given α1 and α2, we encode as a WPMS the problem of finding an injective partial function τ : Dα1 Dα2 such that τ(o1) = o2 implies that o1 and o2 will map to the same object in αu. Wbig must be large enough so that preserving predicates has priority over avoiding the introduction of parameters. Wbig min(|Dα1|, |Dα2|) + 1 accomplishes this. We use the quadratic encoding. |