An AI Planning Solution to Scenario Generation for Enterprise Risk Management
Authors: Shirin Sohrabi, Anton Riabov, Michael Katz, Octavian Udrea
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 7 Experimental Evaluation In this section, we evaluate the performance of the planner, quality of the clusters measured by the size of the cluster, and how informative each cluster is, measured by number of predicates and business implications. In the next section, we provide details on the pilot deployment of the Scenario Planning Adviser (SPA) tool, feedback and the lessons learned in interacting with the domain experts as well as the business users. |
| Researcher Affiliation | Industry | Shirin Sohrabi, Anton V. Riabov, Michael Katz, Octavian Udrea IBM T.J. Watson Research Center 1101 Kitchawan Rd, Yorktown Heights, NY 10598, USA |
| Pseudocode | No | The paper describes algorithms and methods but does not present them in a clearly labeled "Pseudocode" or "Algorithm" block with structured steps. |
| Open Source Code | No | The paper does not provide any link or explicit statement about making the source code for their methodology publicly available. |
| Open Datasets | No | The paper describes using internal data ( |
| Dataset Splits | No | The paper mentions testing with different numbers of observations and discusses clustering, but does not provide specific training/validation/test splits of a dataset for reproducibility. |
| Hardware Specification | Yes | All our experiments were run on a 16-core 2.93 GHz Intel(R) Xeon(R) ES-2680 processor with 264 GB RAM. |
| Software Dependencies | No | The paper mentions using a 'top-k planner (Riabov, Sohrabi, and Udrea 2014)' and implementing 'the LM-cut heuristic (Helmert and Domshlak 2009)' but does not provide specific version numbers for these or any other software dependencies. |
| Experiment Setup | Yes | The timeout was set to 30 minutes. The penalty is relative to the cost of the other actions in the domain. Note, a high discard cost may cause a planner to consider many long and unlikely paths, while a low discard may cause a planner to discard observations without trying to explain them. Hence, we pick a middleground, a penalty that is five times the cost of the next-med action. |