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].
On Testing for Discrimination Using Causal Models
Authors: Hana Chockler, Joseph Y. Halpern5548-5555
AAAI 2022 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | We show (among other things) that the problem of deciding fairness as we have deο¬ned it is co NP-complete, but then argue that, despite that, in practice the problem should be manageable. To summarize, the main contribution of this paper lies in creating a framework that clearly delineates what a regulator will have to do in order to certify an AI system for fairness. We also examine the complexity of determining whether a system is fair, and show that it is co-NP-complete in the size (i.e., number of variables) of the system, but then argue that this should not be a problem in practice. We leave verifying this to future work. |
| Researcher Affiliation | Collaboration | Hana Chockler1, 2, Joseph Y. Halpern3 1 causa Lens 2 Department of Informatics, King s College London 3 Computer Science Department, Cornell University EMAIL, EMAIL |
| Pseudocode | No | The paper contains definitions, theorems, and proofs, but it does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper is theoretical and focuses on a framework and complexity analysis; it does not mention the release of any source code for its methodology. |
| Open Datasets | No | This is a theoretical paper that does not use datasets for empirical evaluation. While it mentions 'historical data that is used by the bank to train its AI system' within the conceptual framework, this does not refer to data used by the authors for their research. |
| Dataset Splits | No | This is a theoretical paper and does not describe experiments that would involve dataset splits for training, validation, or testing. |
| Hardware Specification | No | This is a theoretical paper and does not describe experiments that would require specific hardware specifications. |
| Software Dependencies | No | This is a theoretical paper and does not describe experiments that would require specific software dependencies with version numbers. |
| Experiment Setup | No | This is a theoretical paper and does not describe experiments that would require specific experimental setup details, hyperparameters, or training configurations. |