Safe Exploration for Interactive Machine Learning

Authors: Matteo Turchetta, Felix Berkenkamp, Andreas Krause

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this paper, we apply our framework to safe Bayesian optimization and to safe exploration in deterministic Markov Decision Processes (MDP), which have been analyzed separately before. Our method outperforms other algorithms empirically. Section 4: Applications and Experiments.
Researcher Affiliation Academia Matteo Turchetta Dept. of Computer Science ETH Zurich matteotu@inf.ethz.ch Felix Berkenkamp Dept. of Computer Science ETH Zurich befelix@inf.ethz.ch Andreas Krause Dept. of Computer Science ETH Zurich krausea@ethz.ch
Pseudocode Yes Algorithm 1 GOOSE, Algorithm 2 Safe Expansion (SE)
Open Source Code No The paper does not contain an explicit statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We use Digital Terrain Models of Mars available from Mc Ewen et al. (2007).
Dataset Splits No The paper describes the generation of synthetic data and the use of pre-existing models (Digital Terrain Models of Mars) but does not provide specific details on how these datasets were split into training, validation, and test sets with percentages or sample counts.
Hardware Specification No The paper does not provide specific details about the hardware used for running the experiments (e.g., CPU, GPU models, memory specifications).
Software Dependencies No The paper mentions 'Sci Py (Virtanen et al., 2019)' but does not provide a specific version number for SciPy or any other software dependencies crucial for replication.
Experiment Setup Yes In our experiments, we set βt = 3 for all t ≥ 1 as suggested by Turchetta et al. (2016). We optimize samples from a GP with zero mean and Radial Basis Function (RBF) kernel with variance 1.0 and lengthscale 0.1 and 0.4 for a one-dimensional and two-dimensional, respectively. The observations are perturbed by i.i.d Gaussian noise with σ = 0.01. For simplicity, we set the objective and the constraint to be the same, f = q.