Workshop Program Overview

Sunday July 9

T1

T2

N2

N4

N5

N6

N7

Morning *

9:00 AM - 12:40 PM (CET)

-
-

Complexity

(Tutorial)

-

Afternoon *

2:00 PM - 5:00 PM (CET)

Unified Planning

(Tutorial)

Temporal Networks

(Tutorial)

* Check the program of Sunday, July 9 for the detailed schedule of the workshops

Monday July 10

T1

T2

N2

N4

N5

N6

N7

Morning **

9:00 AM - 12:40 PM (CET)

-
-

MDPS with RDDL

(Tutorial)

MACQ

(Tutorial)

Afternoon **

2:00 PM - 5:00 PM (CET)

Planning and LLMS

(Tutorial)

** Check the program of Monday, July 10 for the detailed schedule of the workshops


List of Workshops

Parisa Zehtabi, Alberto Pozanco, William Yeoh, Biplav Srivastava

Planning and scheduling are mature fields in terms of base techniques and algorithms to solve goal-oriented tasks. Planning approaches have been successfully applied to many domains including classical domains (e.g., logistics and Mars rovers) and, more recently, in oil and gas as well as mining industries. Similarly, scheduling approaches have also been applied to many industrial applications. However, very little work has been done in relation to the problems in the finance industry, which spans a diverse range of activities in financial markets, corporate finance, insurance, banking, and accounting. Recently, some large financial corporations have started AI research labs and researchers at those teams have found that there are plenty of open planning and scheduling problems to be tackled by the ICAPS community. For example, these include trading markets, workflow learning, generation and execution, transactions flow understanding, fraud detection, and customer journeys. In addition, planning problems tackled in other settings like dialog management and network penetration; and richer problem formations involving planning along with learning and scheduling, would also be relevant here.

See more here: FinPlan

July 10, 9:00 AM - 5:30 PM (CET) - T2

Rebecca Eifler, Benjamin Krarup, Alan Lindsay, Lindsay Sanneman, Sarath Sreedharan, Silvia Tulli, Stylianos Loukas Vasileiou

As artificial intelligence (AI) is increasingly being adopted into application solutions, the challenge of supporting effective interactions with humans is becoming more apparent. Partly this is to support integrated working styles, in which humans and intelligent systems cooperate in problem-solving, but also it is a necessary step in the process of building and calibrating trust as humans migrate greater competence and responsibility to such systems. The International Workshop on Human-Aware and Explainable Planning (HAXP), formerly known as the Explainable AI Planning (XAIP) workshop, brings together the latest and best in human-AI interaction and explainability, in the context of planning, scheduling, RL and other forms of sequential decision-making process. The workshop is collocated with ICAPS, the premier conference on automated planning and scheduling. Learn more: HAXP

See more here: HAXP

July 10, 2:00 PM - 5:50 PM (CET) - N6

Emilio M. Sanfilippo, Alessandro Umbrico

Automated Planning and Ontology are two well-established fields of Artificial Intelligence (AI). The former investigates techniques to formally model and reason about the effects of actions, and decide the combinations of actions that allow an agent to achieve goals. The latter investigates techniques to formally define knowledge (by formally describing domain entities and their interrelations), allowing agents to process information about objects, and events, and incrementally build and verify beliefs. Both Automated Planning and Ontology generally rely on logic to model knowledge and organize reasoning mechanisms. They support the development of cognitive capabilities that autonomous agents need to effectively act in the real world. In this context, the PLanning And onTology wOrkshop (PLATO) aims at bringing together researchers in these two fields of AI to address new research challenges, share their experiences, and learn from each other. The workshop aims at investigating the synergetic contributions of technologies and methods from these two fields. There are examples in the literature that have investigated the use of Ontology to generate planning models, find effective plans, and contextualize plans and action execution to domain features.

See more here: PLATO

Cameron Allen, Timo P. Gros, Michael Katz, Harsha Kokel, Hector Palacios, Sarath Sreedharan

While AI Planning and Reinforcement Learning communities focus on similar sequential decision-making problems, these communities remain somewhat unaware of each other on specific problems, techniques, methodologies, and evaluations. This workshop aims to encourage discussion and collaboration between researchers in the fields of AI planning and reinforcement learning. We aim to bridge the gap between the two communities, facilitate the discussion of differences and similarities in existing techniques, and encourage collaboration across the fields. We solicit interest from AI researchers that work in the intersection of planning and reinforcement learning, in particular, those that focus on intelligent decision-making. This is the fifth edition of the PRL workshop series that started at ICAPS 2020.

See more here: PRL

Sara Bernardini, Jesse Davis, Alan Fern, Daniel Höller, Joerg Hoffmann, Michael Katz, Michele Lombardi, Scott Sanner, Marcel Steinmetz, Sylvie Thiebaux, Eyal Weiss

Data-driven AI 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. The first are planning models that are (partially) learned from data (e.g., a weather forecast in a model of flight actions). The second are action-decision components learned from data, in particular, action policies or planning-control knowledge for making decisions in dynamic environments (e.g., manufacturing processes under resource-availability and job-length fluctuations). Given the nature of such data-driven artifacts, reliability is a key concern, prominently including safety, robustness, and fairness in various forms, but possibly other concerns as well. Arguably, this is one of the grand challenges in AI for the foreseeable future.

See more here: RDDPS

July 10, 2:00 PM - 5:00 PM (CET) - N5

Pascal Bercher, Daniel Höller, Julia Wichlacz, Ron Alford

The motivation for using hierarchical planning formalisms is manifold. It ranges from an explicit and predefined guidance of the plan generation process and the ability to represent complex problem solving and behavior patterns to the option of having different abstraction layers when communicating with a human user or when planning co-operatively. The best-known formalism in the field is Hierarchical Task Network (HTN) planning. In addition, there are several other hierarchical planning formalisms, e.g., hybrid planning (incorporating aspects from POCL planning), Hierarchical Goal Network (HGN) planning (incorporating a hierarchy on goals), or formalisms that combine task hierarchies with timeline planning (e.g. ANML). Hierarchies induce fundamental differences from classical planning, creating distinct computational properties and requiring separate algorithms from non-hierarchical planners. Many of these aspects of hierarchical planning are still unexplored. Thus, we encourage any contribution, independent of the underlying hierarchical planning formalism, and want to provide a forum for researchers to discuss the various aspects of hierarchical planning.

See more here: HPlan

July 9, 9:00 AM - 5:00 PM (CET) - N2

Riccardo De Benedictis, Marco Roveri, Shirin Sohrabi Araghi, Sabine Storandt

This workshop series aims to provide a stable forum on relevant topics connected to application-focused research and the deployment of P&S systems. The immediate legacy began in 2007 with the ICAPS'07 Workshop on `Moving Planning and Scheduling Systems into the Real World’, and continued in 2008-2022 with successful yearly editions. 2023 is the 16th edition of SPARK. The websites of the previous editions of SPARK are available here. These workshops presented a stimulating environment where researchers could discuss the opportunity and challenges in moving P&S developments into practice, and analyze domains and problem instances under study for, or closely inspired by, real industrial/commercial deployment of P&S techniques. The challenges and discussions that emerged in the last years’ editions set the baseline for this year’s SPARK workshop. A goal of the workshop series is the definition of a longer term set of challenges that could be of benefit for the research community as well as practitioners. SPARK is the ideal incubator to test, discuss, mature and improve potential papers for that main track with the feedback of an excellent audience, and a great place for the inception of new applications and challenges. Authors of accepted papers will be encouraged to share their domains and instances, or parts of them, towards a library of practical benchmarking problems that could also be useful for the community. Accepted papers will be presented in plenary or poster sessions during the workshop. Each presented paper will receive comments from a designated moderator, in order to start the discussion at the workshop.

See more here: SPARK

July 10, 9:00 AM - 5:50 PM (CET) - T1

Iman Awaad, Alberto Finzi, Andrea Orlandini

AI Planning & Scheduling (P&S) methods are key to enabling intelligent robots to perform autonomous, flexible, and interactive behaviours. Researchers in the P&S community have continued to develop approaches and produce planners, representations, as well as heuristics that robotics researchers can make use of. However, there remain numerous challenges complicating the uptake, use and successful integration of P&S technology in robotics, many of which have been addressed by robotics researchers with novel solutions. Strong collaboration and synergy between researchers in both communities is vital to the continued growth of the fields in a way that provide mutual benefits to the two communities. To foster this, the PlanRob workshop aims to provide a stable, long-term forum (having been held annually at ICAPS since 2013) where researchers from both the P&S and Robotics communities can openly discuss relevant issues, research and development progress, future directions and open challenges related to P&S when applied to Robotics. In addition to the usual paper submissions, the workshop?s format naturally lends itself to preliminary results, position papers as well as to work focused on challenges in using and integrating planners in robotics systems.

See more here: PlanRob

July 9, 2:00 PM - 5:05 PM (CET) - N4

Sunadita Patra, Wiktor Piotrowski, Mak Roberts, Tiago Stegun Vaquero

Automated planners are increasingly being integrated into online acting systems. The integration may, for example, embed a domain-independent temporal planner in a manufacturing system (e.g., the Xerox printer application) or autonomous vehicles (e.g., a planetary rover or anunderwater glider). The integration may resemble something more like an “acting and planning stack” where an automated planner produces an activity or task plan that is further refined by an actor before being executed by the execution platform of the actor, such as, a reactive controller (e.g., robotics). Or the integration may be a domain-specific policy that maps states to actions (e.g., reinforcement learning). Models for planning and execution can be same or different; the planning model can define context-dependent actions schema for online (re-)planning or can just specify flexibility to be handled separately at execution time. Online learning may or may not be involved, and may include adjusting or augmenting the model, determining when to repair versus replan, learning to switch policies, etc. A specific focus of these integrations involves online deliberation and managing the execution of actions, bringing to the foreground concerns over how much computational effort planning should invest over time.

See more here: IntEx

July 10, 9:00 AM - 5:30 PM (CET) - T2

Richard Magnotti, Simona Ondrckova, Denson George, Kristyna Pantuckova, Christopher Geib

Plan recognition, activity recognition, and intent recognition all involve making inferences about other actors from observations of their behavior, i.e., their interaction with the environment and with each other. The observed actors may be software agents, robots, or humans. This synergistic area of research combines and unifies techniques from user modeling, machine vision, intelligent user interfaces, human/computer interaction, autonomous and multi-agent systems, natural language understanding, and machine learning. It plays a crucial role in a wide variety of applications including: Assistive technology; Software assistants; Computer and network security; Behavior recognition; Coordination in robots and software agents; E-commerce and collaborative filtering. This wide-spread diversity of applications and disciplines, while producing a wealth of ideas and results, has unfortunately contributed to fragmentation in the field, as researchers publish relevant results in a wide spectrum of journals and conferences. Further, as there is no single common conference for this work, we believe the workshop we are proposing will provide a valuable place to discuss and improve preliminary work of this sub-field.

See more here: PAIR

Clemens Buchner, Daniel Gnad, Thorsten Klossner, Sofia Lemons

Heuristics and search algorithms are the two key components of heuristic search, one of the main approaches to many variations of domain-independent planning, including classical planning, temporal planning, planning under uncertainty and adversarial planning. This workshop seeks to understand the underlying principles of current heuristics and search methods, their limitations, ways for overcoming those limitations, as well as the synergy between heuristics and search. The HSDIP workshop has always been welcoming of multidisciplinary work, for example, drawing inspiration from operations research (like row and column generation algorithms), convex optimization (like gradient optimization for hybrid planning), constraint programming, or satisfiability. The workshop is meant to be an open and inclusive forum, and we encourage papers that report on work in progress or that do not fit the mold of a typical conference paper. Non-trivial negative results are welcome to the workshop, but we expect the authors to argue for the significance of the presented results.

See more here: HSDIP

Lukas Chrpa, Ron Petrick, Mauro Vallati, Tiago Vaquero

Despite the progress in automated planning and scheduling systems, these systems still need to be fed by carefully engineered domain and problem descriptions and they need to be fine-tuned for particular domains and problems. Knowledge engineering for AI planning and scheduling deals with the acquisition, design, validation and maintenance of domain models, and the selection and optimization of appropriate machinery to work on them. These processes impact directly on the success of real-world planning and scheduling applications. The importance of knowledge engineering techniques is clearly demonstrated by a performance gap between domain-independent planners and planners exploiting domain-dependent knowledge. The workshop will continue the tradition of several International Competitions on Knowledge Engineering for Planning and Scheduling (ICKEPS) and prior KEPS workshops. Rather than focusing only on software tools and domain encoding techniques –which are topics of ICKEPS– the workshop will cover all aspects of knowledge engineering for AI planning and scheduling.

See more here: KEPS