Publications

Tagged as research
Written on

Peer-Reviewed Papers (In Chronological Order)

  1. Asai, M: 2019. Unsupervised Grounding of Plannable First-Order Logic Representation from Images. In ICAPS-2019 . on Arxiv.
  2. Asai, M.; Kajino, H: 2019. Towards Stable Symbol Grounding with Zero-Suppressed State AutoEncoder. In ICAPS-2019 . on Arxiv.
  3. Asai, M.; Fukunaga, A: 2017. Classical Planning in Deep Latent Space: Bridging the Subsymbolic-Symbolic Boundary. In AAAI-2018 . Extended manuscript on Arxiv.
  4. Masataro Asai, Akihiro Kishimoto, Adi Botea, Radu Marinescu, Elizabeth Daly M, and Spyros Kotoulas. 2017. Efficient Optimal Search under Expensive Edge Cost Computation. In IJCAI-2017. pdf
  5. Asai, M.; Fukunaga, A: 2017. "Exploration Among and Within Plateaus in Greedy Best-First Search". In ICAPS2017. pdf
  6. Asai, M.; Fukunaga, A: 2017. "Tie-Breaking Strategies for Cost-Optimal Best First Search." Journal of Artificial Intelligence Research 58 (2017): 67-121. pdf
  7. Asai, M.; Fukunaga, A: 2016. Tiebreaking Strategies for A* Search: How to Explore the Final Frontier. In AAAI-2016. pdf presentation poster. (Also submitted to Japanese Society for Artificial Intelligence, received JSAI Annual Conference Student Incentive Award)
  8. Asai, M.; Fukunaga, A: 2015. Solving Large-Scale Planning Problems by Decomposition and Macro Generation. In ICAPS2015. pdf . CAP planner is available at https://github.com/guicho271828/CAP .
  9. Asai, M.; Fukunaga, A: 2014. Fully Automated Cyclic Planning for Large-Scale Manufacturing Domains. In ICAPS2014. pdf poster presentation

Presentation can be moved forward/backward with N/P key. For the further help and usage, click here.

Workshop papers

  1. Asai, M.; Fukunaga, A: 2017. Classical Planning in Deep Latent Space: From Unlabeled Images to PDDL (and back). Knowledge Engineering for Planning and Scheduling (KEPS) Workshop (ICAPS2017).
  2. Asai, M.; Fukunaga, A: 2017. Tie-Breaking in A* as Satisficing Search. Heuristics and Search for Domain-independent Planning (HSDIP) Workshop (ICAPS2017).
  3. Asai, M.: 2017. Exploiting Search Space Structure in Classical Planning: Analyses and Algorithms (Dissertation Abstract, 2017 version). (ICAPS2017 Doctorial Consortium). pdf
  4. Asai, M.: 2016. Exploiting Search Space Structure in Classical Planning: Analyses and Algorithms (Dissertation Abstract). (ICAPS2016 Doctorial Consortium). pdf
  5. Endo, S; Asai, M.; Fukunaga, A: 2014. Evaluation of a Simple, Window-based, Replanning Approach to Plan Optimization. Heuristics and Search for Domain-independent Planning (HSDIP) Workshop (ICAPS2016). pdf presentation
  6. Asai, M.; Fukunaga, A: 2014. Applying Problem Decomposition to Extremely Large Planning Domains. Knowledge Engineering for Planning and Scheduling (KEPS) Workshop (ICAPS2014). pdf (submitted version -> final version) poster presentation

Software

See my github repo for the latest activity!

  • CAP – Component Abstraction Planner, which decompose the given problem, solve each subproblem, make the subplans into macros and then plans in an enhanced problem with those macros
    • 1.5 coverage in large domains!
    • Higer coverage in ipc2011 learning track, without learning time!
  • PDDL – A Common Lisp library to read/write/analyse PDDL files. It has
    • a PDDL reader / parser
    • CLOS-based object oriented interface to analyse each objects
      • various useful accessors for objects e.g. predicate/propositions/action/types
      • methods like (ground-action action objects)
    • a pretty formatter
    • a simulator (STRIPS and action-costs are supported)
  • CELL-ASSEMBLY – The PDDL files and the explanation of a CELL-ASSEMBLY manufacturing domain, which appears at [Asai, Fukunaga ICAPS2014]
  • aaai-template – For org-mode lovers and reserchers of artificial intelligence. A set of scripts and templates for faster publishing of papers with AAAI style.
    • Included Makefile automates
      • converting SVG images to grayscale pngs (I recommend Inkscape as an editor)
      • exporting .org files to .tex files
      • typeset the TeX files with pdflatex
      • checks for any Overfull hbox
      • checks for paper limit
    • Combination of simple scripts make-periodically would be useful. It waits for the changes, and if a change is detected, it runs make, then notify the result via inotify (pop up inteface available in gtk).