A Multi-Objective Approach to Mitigate Negative Side Effects

Authors: Sandhya Saisubramanian, Ece Kamar, Shlomo Zilberstein

IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical evaluation of our approach shows that the proposed framework can successfully mitigate NSE and that different feedback mechanisms introduce different biases, which influence the identification of NSE.
Researcher Affiliation Collaboration 1University of Massachusetts, Amherst, MA 2Microsoft Research, Redmond, WA
Pseudocode Yes Algorithm 1 Slack Estimation ( M, Ω)
Open Source Code No The paper does not mention providing open-source code for the methodology described.
Open Datasets No The paper describes custom domains ('Boxpushing', 'Driving') with generated instances rather than referencing publicly available datasets with access information.
Dataset Splits No The paper mentions 'three-fold cross validation is performed' for hyperparameter tuning but does not specify a training/validation/test dataset split for the main experiments.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU, GPU models, or memory specifications) used for running experiments.
Software Dependencies No Random forest regression from sklearn Python package is used for model learning. (The version number for the scikit-learn package is not specified).
Experiment Setup Yes The slack is computed using Algorithm 1 and γ = 0.95. Values averaged over 100 trials of planning and execution, along with their standard errors, are reported for the following domains.