Weighted Clock Logic Point Process
Authors: Ruixuan Yan, Yunshi Wen, Debarun Bhattacharjya, Ronny Luss, Tengfei Ma, Achille Fokoue, Anak Agung Julius
ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on synthetic datasets manifest our model s ability to recover the ground-truth rules and improve computational efficiency. In addition, experiments on real-world datasets show that our models perform competitively when compared with state-of-the-art models. |
| Researcher Affiliation | Collaboration | 1Rensselaer Polytechnic Institute 2IBM T.J. Watson Research Center |
| Pseudocode | Yes | Algorithm 1 Simulation of a homogeneous Poisson process with intensity rate λ. |
| Open Source Code | Yes | Our code is available at https://ICLR-CLNN. |
| Open Datasets | Yes | Linked In [(Xu et al., 2017)]. Mimic II [(Saeed et al., 2011)]. Stack Overflow [(Grant & Betts, 2013)]. |
| Dataset Splits | Yes | Each synthetic dataset contains 1, 000 event streams partitioned into three sets: training (70%), validation (15%), and test (15%). |
| Hardware Specification | Yes | The experiments are run using the Adam W optimizer in Pytorch (1.10.2) on a Windows 10 system desktop with a 16-core CPU (i7, 3.60GHz) and 32 GB RAM. |
| Software Dependencies | Yes | The experiments are run using the Adam W optimizer in Pytorch (1.10.2) on a Windows 10 system desktop with a 16-core CPU (i7, 3.60GHz) and 32 GB RAM. |
| Experiment Setup | Yes | The truth value threshold is set as α = 0.5, and the clock signal for representing an event not occurring in H t is set as Z = 1.5Tmax, where Tmax is the maximal ending time among all the event streams. During the training process, we initialize the parameters using four approaches (see Appendix C.5 for more details) and report the best one... For simplicity, ulilj are set as 0 to study the ordering representations. The truth value threshold is α = 0.5, and Z = 1.5Tmax, same as the setting for the synthetic datasets, and the number of subformulas is n = 5, and the parameters are initialized as random numbers from a uniform distribution on [0, 1). |