Author name cluster

Luca Putelli

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

AAMAS Conference 2025 Conference Paper

Towards Efficient Online Goal Recognition through Deep Learning

Lorenzo Serina
Mattia Chiari
Alfonso Emilio Gerevini
Luca Putelli
Ivan Serina

Online goal recognition (OGR) is the task of understanding the intention of an agent as it executes a plan, recognizing its goal every time it performs a new action. This task is important in different contexts, such as applications of cyber-security and human-robot collaboration. An effective OGR system should (i) compute the correct goal of the agent as early as possible with respect to the agent’s performed actions, and (ii) perform fast, given that, while the OGR system infers the agent’s goal, the agent keeps executing the plan. In this paper, we propose a deep-learning approach to OGR based on Recurrent Neural Networks. The approach is implemented in a new system that learns to predict the goal of an agent acting in a given planning domain using a training dataset for the domain. We propose a method that exploits planning-related knowledge for designing a training dataset that is effective for the OGR and for improving the system performance. An experimental evaluation of our system on several benchmark domains shows that it performs generally better than the state-of-the-art in terms of accuracy and execution time, considering both the requirements (i) and (ii).

PDF

NeSy Conference 2024 Conference Paper

Enhancing GPT-Based Planning Policies by Model-Based Plan Validation

Nicholas Rossetti
Massimiliano Tummolo
Alfonso Emilio Gerevini
Matteo Olivato
Luca Putelli
Ivan Serina

Abstract Despite Large Language Models (LLMs) have revolutionised Natural Language Processing (NLP), their capability of performing logical reasoning and automated planning is still debated. In this context, the state of the art is PlanGPT, a GPT-2 model specifically trained for planning tasks. This recent approach provides GPT-based planning policies with remarkable performance, but it can generate invalid plans containing violated action preconditions or unsatisfied goals. To address this limitation, we propose an extension of PlanGPT that integrates a plan validator into the generation process. The validator is exploited to prune invalid plan prefixes during the GPT token generation, obtaining a more robust and powerful solution to planning via GPT. We empirically evaluate the effectiveness of our approach and demonstrate its potential in various planning domains.

Details

AAMAS Conference 2024 Conference Paper

Fast and Slow Goal Recognition

Mattia Chiari
Alfonso Emilio Gerevini
Andrea Loreggia
Luca Putelli
Ivan Serina

Goal recognition is a crucial aspect of understanding the intentions and objectives of agents by observing some of their actions. The most prominent approaches to goal recognition can be divided into two main categories: (1) trustworthy systems, which exploit automated reasoning for computing plans compatible with the observed actions, and (2) swifter systems, which try to quickly infer goals, often overlooking complex cognitive processes, and have no formal guarantees of their results. This paper introduces a novel approach inspired by the dual process theory, which integrates these two techniques. A dual-process model is proposed, leveraging fast, experience-based recognition for immediate goal identification, and slow, deliberate analysis for deeper understanding. Machine learning techniques and classical planning techniques are employed to obtain this dual-process system. Experimental evaluations demonstrate the effectiveness of the approach, reducing the amount of resources required to compute a solution (e. g. , time to find a goal), while at the same time enhancing accuracy and robustness, especially in more complex scenarios.

PDF

ICAPS Conference 2024 Conference Paper

Learning General Policies for Planning through GPT Models

Nicholas Rossetti
Massimiliano Tummolo
Alfonso Emilio Gerevini
Luca Putelli
Ivan Serina
Mattia Chiari
Matteo Olivato

Transformer-based architectures, such as T5, BERT and GPT, have demonstrated revolutionary capabilities in Natural Language Processing. Several studies showed that deep learning models using these architectures not only possess remarkable linguistic knowledge, but they also exhibit forms of factual knowledge, common sense, and even programming skills. However, the scientific community still debates about their reasoning capabilities, which have been recently tested in the context of automated AI planning; the literature presents mixed results, and the prevailing view is that current transformer-based models may not be adequate for planning. In this paper, we address this challenge differently. We introduce a GPT-based model customised for planning (PLANGPT) to learn a general policy for classical planning by training the model from scratch with a dataset of solved planning instances. Once PLANGPT has been trained for a domain, it can be used to generate a solution plan for an input problem instance in that domain. Our training procedure exploits automated planning knowledge to enhance the performance of the trained model. We build and evaluate our GPT model with several planning domains, and we compare its performance w. r. t. other recent deep learning techniques for generalised planning, demonstrating the effectiveness of the proposed approach.

Details

ICAPS Conference 2023 Conference Paper

Goal Recognition as a Deep Learning Task: The GRNet Approach

Mattia Chiari
Alfonso Emilio Gerevini
Francesco Percassi
Luca Putelli
Ivan Serina
Matteo Olivato

Recognising the goal of an agent from a trace of observations is an important task with many applications. The state-of-the-art approach to goal recognition (GR) relies on the application of automated planning techniques. We study an alternative approach, called GRNet, where GR is formulated as a classification task addressed by machine learning. GRNet is primarily aimed at solving GR instances more accurately and more quickly by learning how to solve them in a given domain, which is specified by a set of propositions and a set of action names. The goal classification instances in the domain are solved by a Recurrent Neural Network (RNN). The only information required as input of the trained RNN is a trace of action labels, each one indicating just the name of an observed action. A run of the RNN processes a trace of observed actions to compute how likely it is that each domain proposition is part of the agent

Details

JAIR Journal 2023 Journal Article

Maintenance of Plan Libraries for Case-Based Planning: Offline and Online Policies

Alfonso Emilio Gerevini
Alessandro Saetti
Ivan Serina
Andrea Loreggia
Luca Putelli
Anna Roubickova

Case-based planning is an approach to planning where previous planning experience provides guidance to solving new problems. Such a guidance can be extremely useful, or even necessary, when the new problem is very hard to solve, or the stored previous experience is highly valuable, because, e.g., it was provided or validated by human experts, and the system should try to reuse it as much as possible. To do so, a case-based planning system stores in a library previous planning experience in the form of already encountered problems and their solutions. The quality of such a plan library critically influences the performance of the planner, and therefore it needs to be carefully designed and created. For this reason, it is also important to update the library during the lifetime of the system, as the type of problems being addressed may evolve or differ from the ones the library was originally designed for. Moreover, like in general case-based reasoning, the library needs to be maintained at a manageable size, otherwise the computational cost of querying it grows excessively, making the entire approach ineffective. In this paper, we formally define the problem of maintaining a library of cases, discuss which criteria should drive the maintenance, study the computational complexity of the maintenance problem, and propose offline techniques to reduce an oversized library that optimize different criteria. Moreover, we introduce a complementary online approach that attempts to limit the growth of the library, and we consider the combination of offline and online techniques to ensure the best performance of the case-based planner. Finally, we experimentally show the practical effectiveness of the offline and online methods for reducing the library.

PDF Details DOI