Cost-Effective Incentive Allocation via Structured Counterfactual Inference

Authors: Romain Lopez, Chenchen Li, Xiang Yan, Junwu Xiong, Michael Jordan, Yuan Qi, Le Song4997-5004

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

Reproducibility Variable Result LLM Response
Research Type Experimental We also establish theoretical error bounds for our estimation procedure and we empirically show that the approach leads to significant improvement on both synthetic and real datasets. 6 Experiments In our experiments, we consider only the case of a discrete action space. We compare our method CCPOv SIRE to a simple modification of Bandit Net...
Researcher Affiliation Collaboration 1Department of Electrical Engineering and Computer Sciences, University of California, Berkeley {romain lopez, jordan}@cs.berkeley.edu 2AI Department, Ant Financial Service Group {junwu.xjw, yuan.qi, le.song}@antfin.com 3Department of Computer Science, Shanghai Jiao Tong University lcc1992@sjtu.edu.cn, xyansjtu@163.com 4College of Computing, Georgia Institute of Technology
Pseudocode No The paper describes its algorithm components and steps but does not include structured pseudocode or an algorithm block.
Open Source Code No The paper does not include an unambiguous statement or a direct link to a source-code repository for the methodology described.
Open Datasets Yes We construct a concrete example from the Image Net dataset (Deng et al. 2009)
Dataset Splits Yes The dataset has 4608 samples in total, randomly split into training, validation and testing with ratio 0.6: 0.2: 0.2.
Hardware Specification Yes We ran our experiments on a machine with a Intel Xeon E5-2697 CPU and a NVIDIA Ge Force GTX TITAN X GPU.
Software Dependencies No The paper does not provide specific software dependencies with version numbers for replication.
Experiment Setup Yes To train CCPOv SIRE, we used stochastic gradient descent as a first-order stochastic optimizer with a learning rate of 0.01 and a three-layer neural network with 512 neurons for each hidden layer.