A POMDP Formulation of Proactive Learning

Authors: Kyle Wray, Shlomo Zilberstein

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our method is compared with the original three algorithms proposed by Donmez and Carbonell and a simple baseline. We demonstrate that our approach matches or improves upon the original approach within five different oracle scenarios, each on two datasets.
Researcher Affiliation Academia Kyle Hollins Wray and Shlomo Zilberstein College of Information and Computer Sciences University of Massachusetts Amherst, MA 01003, USA {wray,shlomo}@cs.umass.edu
Pseudocode Yes Algorithm 1 Proactive Learning POMDP: Initially unknown oracles; thus, it requires clustering.
Open Source Code No Finally, we will provide our source code so that others may more easily build upon this work to design and implement high caliber proactive learners.
Open Datasets Yes We create similar experiments to those of Donmez and Carbonell for two of the well-known UCI datasets they used: Adult and Spambase (Lichman 2013). ... Lichman, M. 2013. UCI machine learning repository. URL: http://archive.ics.uci.edu/ml.
Dataset Splits No The paper mentions using a "test set" and
Hardware Specification Yes Our implementation uses Python 3.4.3 with scikit-learn 0.16.1, Num Py 1.9.2, and Sci Py 0.15.1, run on an Intel(R) Core(TM) i7-4702HQ CPU at 2.20GHz, 8GB of RAM, and a Nvidia(R) Ge Force GTX 870M.
Software Dependencies Yes Our implementation uses Python 3.4.3 with scikit-learn 0.16.1, Num Py 1.9.2, and Sci Py 0.15.1, run on an Intel(R) Core(TM) i7-4702HQ CPU at 2.20GHz, 8GB of RAM, and a Nvidia(R) Ge Force GTX 870M. We leverage a high-performing GPU-based implementation of PBVI using CUDA(C) 6.5 (Wray and Zilberstein 2015a; 2015b).
Experiment Setup Yes PBVI uses horizon 2du (increasing it only improves performances) and discount factor 0.9.