Planning with Learned Binarized Neural Networks Benchmarks for MaxSAT Evaluation 2021

This document provides a brief introduction to learned automated planning problem where the state transition function is in the form of a binarized neural network (BNN), presents a general MaxSAT encoding for this problem, and describes the four domains, namely: Navigation, Inventory Control, System Administrator and Cellda, that are submitted as benchmarks for MaxSAT Evaluation 2021.

[1]  Peter J. Stuckey,et al.  Theoretical and Experimental Results for Planning with Learned Binarized Neural Network Transition Models , 2020, CP.

[2]  Blai Bonet,et al.  LP-Based Heuristics for Cost-Optimal Planning , 2014, ICAPS.

[3]  Peter J. Stuckey,et al.  Encoding Linear Constraints into SAT , 2014, CP.

[4]  Scott Sanner,et al.  Compact and efficient encodings for planning in factored state and action spaces with learned Binarized Neural Network transition models , 2020, Artif. Intell..

[5]  Buser Say.  Optimal Planning with Learned Neural Network Transition Models , 2020 .

[6]  Buser Say,et al.  A Unified Framework for Planning with Learned Neural Network Transition Models , 2021, AAAI.

[7]  Scott Sanner,et al.  Scalable Planning with Deep Neural Network Learned Transition Models , 2020, J. Artif. Intell. Res..

[8]  Scott Sanner,et al.  Planning in Factored State and Action Spaces with Learned Binarized Neural Network Transition Models , 2018, IJCAI.

[9]  Paolo Traverso,et al.  Automated Planning: Theory & Practice , 2004 .

[10]  Scott Sanner,et al.  Reward Potentials for Planning with Learned Neural Network Transition Models , 2019, CP.

[11]  Scott Sherwood Benson,et al.  Learning action models for reactive autonomous agents , 1996 .

[12]  Herbert A. Simon,et al.  Rule Creation and Rule Learning Through Environmental Exploration , 1989, IJCAI.

[13]  Peter J. Stuckey,et al.  Sequencing Operator Counts , 2015, ICAPS.

[14]  Corbeil-Essonnes.  The Legend of Zelda , 2011 .

[15]  Ran El-Yaniv,et al.  Binarized Neural Networks , 2016, ArXiv.

[16]  Scott Sanner,et al.  Nonlinear Hybrid Planning with Deep Net Learned Transition Models and Mixed-Integer Linear Programming , 2017, IJCAI.

[17]  Carlos Guestrin,et al.  Max-norm Projections for Factored MDPs , 2001, IJCAI.

[18]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[19]  Malte Helmert,et al.  The Fast Downward Planning System , 2006, J. Artif. Intell. Res..

[20]  Bernhard Nebel,et al.  The FF Planning System: Fast Plan Generation Through Heuristic Search , 2011, J. Artif. Intell. Res..

[21]  Shie Mannor,et al.  Scaling Up Approximate Value Iteration with Options: Better Policies with Fewer Iterations , 2014, ICML.

[22]  Scott W. Bennett,et al.  Real-world robotics: Learning to plan for robust execution , 1996, Machine Learning.

[23]  Yolanda Gil,et al.  Acquiring domain knowledge for planning by experimentation , 1992 .

[24]  Bart Selman,et al.  Planning as Satisfiability , 1992, ECAI.