Indirect Trust is Simple to Establish
Authors: Elham Parhizkar, Mohammad Hossein Nikravan, Sandra Zilles
IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate existing state-of-the-art methods for computing indirect trust in numerous simulations, demonstrating that the best ones tend to be of prohibitively large complexity. We propose a new and easy to implement method for computing indirect trust, based on a simple prediction with expert advice strategy as is often used in online learning. This method either competes with or outperforms all tested systems in the vast majority of the settings we simulated, while scaling substantially better. Our results demonstrate that existing systems for computing indirect trust are overly complex; the problem can be solved much more efficiently than the literature suggests. |
| Researcher Affiliation | Academia | Elham Parhizkar , Mohammad Hossein Nikravan and Sandra Zilles Department of Computer Science, University of Regina, Canada e.parhizkarabyaneh@gmail.com, nikravam@uregina.ca, zilles@cs.uregina.ca |
| Pseudocode | Yes | Algorithm 1 The ITEA Algorithm |
| Open Source Code | No | The paper does not include any explicit statements about releasing source code, nor does it provide a link to a code repository. |
| Open Datasets | No | The paper describes a simulated environment and how data is generated within that simulation ("The trustworthiness value of each individual trustee is sampled uniformly at random... Each reported result is the average over 100 such sampled trustee combinations.") rather than using a pre-existing, publicly available dataset with a direct access link or formal citation. |
| Dataset Splits | No | The paper describes its simulation setup and evaluation metrics (e.g., "after S successful interactions, the ratio... was recorded") but does not specify traditional training, validation, and test dataset splits with percentages or sample counts. |
| Hardware Specification | Yes | We measured CPU time on a Mac Book Pro/2.3 GHz Intel Core i7 for various settings. |
| Software Dependencies | No | The paper mentions specific methods and models like "Beta Reputation System (BRS)" and "DP-Dirichlet model", and a "squared-error loss" function, but it does not specify software dependencies with version numbers (e.g., Python 3.x, PyTorch 1.x). |
| Experiment Setup | Yes | ITEA requires a loss function and a learning rate. As loss function, we chose squared-error loss. As mentioned before, the learning rate η = p 8 ln(K)/T yields good guarantees, where K is the number of advisers and T the number of rounds of interaction with the trustee whose trustworthiness is being estimated. In our simulations, K = 100 and T is not known in advance, so that we replaced T by the total number of interactions (when measuring MAE) or the target number of positive interactions (when measuring RFU). ... Initialization: wk,0 = 1 K , 1 k K |