Workshop on Reliability In Planning and Learning (RIPL)

ICAPS'25 Workshop on Reliability In Planning and Learning (RIPL) Melbourne, Victoria, Australia

Tuesday, November 11, 2025 from 8:30 to 15:00 in Room 3

Aim and Scope of the Workshop

Learning is the dominating trend in AI at this time. From a planning and scheduling perspective – and for sequential decision making in general – this is manifested in two major kinds of technical artifacts that are rapidly gaining importance. First, planning models generated by large language models, or otherwise learned or partially learned from data (such as a weather forecast in a model of flight actions). Second, planning/search information learned from data, in particular action policies or planning-control knowledge for making decisions in dynamic environments (reinforcement learning or per-domain generalizing knowledge in PDDL). Reliability is a key concern in such artefacts, prominently including safety, robustness, and fairness in various forms, but possibly other concerns as well. Arguably, this is indeed one of the grand challenges in AI for the foreseeable future.

RIPL is an evolution of the Reliable Data-Driven Planning and Scheduling (RDDPS) workshop that ran ICAPS 2022-2024. The workshop scope was edited to be more broadly inclusive, covering any aspect relevant to reliability in planning and learning, in particular LLM-generated planning models. The title was changed to reflect this scope in a more direct manner.

RIPL aims to coordinate with the workshops on Planning and Reinforcement Learning (PRL) as well as Language Models for Planning (LM4Plan), with RIPL covering the reliability-related aspects of these areas. Joint sessions across workshops are a possibility we will evaluate depending on submissions and workshop timing.

Schedule

RIPL will take place on Tuesday, November 11, 2025 in Room 3

8:30 - 10:00 – Session 1 (Learning and Planning):
- Welcome and Open Remarks
- Effective Data Generation and Feature Selection in Learning for Planning
  Mingyu Hao, Dillon Chen, Felipe Trevizan and Sylvie Thiebaux
- Learning Lifted Action Models from Unsupervised Visual Traces
  Kai Xi, Stephen Gould and Sylvie Thiebaux
- Using Action-Policy Testing in RL to Reduce the Number of Bugs
  Hasan Ferit Eniser, Songtuan Lin, Nicola Müller, Anastasia Isychev, Valentin Wüstholz, Isabel Valera, Jörg Hoffmann and Maria Christakis
- Option Invention for Continual Hierarchical Reinforcement Learning and Planning
  Rashmeet Kaur Nayyar and Siddharth Srivastava
10:00 – 10:30 – Coffee Break
10:30 – 12:00 – Session 2 (Classical Planning and Action Policy Safety Analysis):
- Is This a Good Decision? Action Optimality Checking in Classical Planning
  Jan Eisenhut, Daniel Fišer, Wheeler Ruml and Jörg Hoffmann
- Probabilistic Safety Verification of Neural Policies via Predicate Abstraction
  Marcel Vinzent, Holger Hermanns and Joerg Hoffmann
- Combined talk:
  - Policy Safety Testing in Non-Deterministic Planning: Fuzzing, Test Oracles, Fault Analysis
    Chaahat Jain, Daniel Sherbakov, Marcel Vinzent, Marcel Steinmetz, Jesse Davis and Joerg Hoffmann
  - Safety Debugging of Tree-Ensemble Action Policies in AI Planning: From Fault Detection to Fault Fixing
    Lorenzo Cascioli, Chaahat Jain, Marcel Steinmetz, Jesse Davis and Joerg Hoffmann
12:00 – 13:30 – Lunch Break
13:30 – 15:00 – Session 3 (GPT for Planning and Numerical Planning):
- Enhancing GPT-based Planning Policies by Model-based Plan Validation
  Nicholas Rossetti, Massimiliano Tummolo, Alfonso Emilio Gerevini, Matteo Olivato, Luca Putelli and Ivan Serina
- Integrating Classical Planners with GPT-based Planning Policies
  Massimiliano Tummolo, Nicholas Rossetti, Alfonso Emilio Gerevini, Matteo Olivato, Luca Putelli and Ivan Serina
- Adapting to Novelties in Numeric Planning
  Nir Aharoni, Yarin Benyamin, Shiwali Mohan and Roni Stern
- Open discussion

Topics of Interest

The workshop welcomes contributions to any topic that roughly falls into the following problem space:

Data-driven artifacts: Learned or ML-generated planning and scheduling models (e.g., action models, transition probabilities and environment prediction); learned action-decisions (e.g., action policies, components thereof and previous plans); learned search guidance (e.g., heuristics and state rankings); and combinations thereof.
Objectives: Reliability in whatever form, including risk, safety, robustness, fairness, error bounds, etc., alongside possibly other concerns such as scalability and data efficiency, system design/engineering principles and challenges, and the interactions of these with reliability.
Methodologies include any issue relating to robustness in: learning or generating artefacts as per (1); planning and scheduling algorithms in the presence of such learned artifacts; analyzing such artifacts (e.g., reasoning, verification, testing, etc.); making such analyses amenable to human users (e.g., visualization, interaction); and potentially others as relevant to the workshop objectives.

Submission Details

All papers must be formatted like at the main conference (ICAPS author kit). Submitted papers should be anonymous for double-blind reviewing. Paper submission is via EasyChair.

We call for two kinds of submissions:

Technical papers, of length up to 8 pages plus unlimited references and appendices. The workshop is meant to be an open and inclusive forum, and we encourage papers that report on work in progress.
Position papers, of length up to 4 pages plus unlimited references and appendices. Given that reliability of data-driven planning and scheduling is rather new at ICAPS, we encourage authors to submit positions on what they believe are important challenges, questions to be considered, approaches that may be promising. We will include any position relevant to discussing the workshop topic. We expect to group position paper presentations into a dedicated session, followed by an open discussion.

Every submission will be reviewed by members of the program committee according to the usual criteria such as relevance to the workshop, significance of the contribution, and technical quality.

Policy on Previously Published Materials

Please do not submit papers that are already accepted for the ICAPS main conference. All other submissions, e.g., papers under review for AAAI'25, are welcome. Authors submitting papers rejected from the ICAPS main conference, please ensure you do your utmost to address the comments given by ICAPS reviewers. Also, it is your responsibility to ensure that other venues your work is submitted to allow for papers to be already published in “informal” ways (e.g., on proceedings or websites without associated ISSN/ISBN).

Organizing Committee

Felipe Trevizan, Australian National University, Australia
Charles Gretton, Australian National University, Australia
Daniel Höller, Saarland University, Germany
Marcel Steinmetz, University of Toulouse, France
Marcel Vinzent, Saarland University, Germany
Jörg Hoffmann, Saarland University, Germany
Sylvie Thiebaux, University of Toulouse, France, and Australian National University, Australia