Fast and Slow Planning

Fabiano, Francesco; Pallagani, Vishal; Ganapini, Marianna Bergamaschi; Horesh, Lior; Loreggia, Andrea; Murugesan, Keerthiram; Rossi, Francesca; Srivastava, Biplav

doi:10.48550/arxiv.2303.04283

Cited by 3 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They can be seen as akin to meta-heuristic approaches, capable of accelerating plan generation in a variety of settings. As such, their application, governed by cognitive-inspired frameworks like SOFAI (Fabiano et al 2023), could delineate when and where their use is most advantageous.…”

Section: Discussionmentioning

confidence: 99%

“…Efforts to generate multimodal, text, and image-based goalconditioned plans are exemplified by (Lu et al 2023b). Additionally, a subset of studies in this survey investigates the fine-tuning of seq2seq, code-based language models (Pallagani et al 2022(Pallagani et al , 2023b, which are noted for their advanced Application of LLMs in Planning Language Translation ( 23) Xie et al 2023;Guan et al 2023;Chalvatzaki et al 2023;Yang, Ishay, and Lee 2023;Wong et al 2023;Kelly et al 2023;Lin et al 2023c;Sakib and Sun 2023;Yang et al 2023b;Parakh et al 2023;Dai et al 2023;Yang et al 2023a;Shirai et al 2023;Ding et al 2023b;Zelikman et al 2023;Pan et al 2023;Xu et al 2023b;Brohan et al 2023;Yang, Gaglione, and Topcu 2022;Chen et al 2023a;You et al 2023) Plan Generation (53) (Sermanet et al 2023;Li et al 2023b;Pallagani et al 2022;Silver et al 2023;Pallagani et al 2023b;Arora and Kambhampati 2023;Fabiano et al 2023;Chalvatzaki et al 2023;Gu et al 2023;Silver et al 2022;Hao et al 2023a;Lin et al 2023b;Yuan et al 2023b;Gandhi, Sadigh, and Goodman 2023;…”

Section: Plan Generationmentioning

confidence: 99%

“…On the other hand, seq2seq LLMs can generate valid plans but struggle with generalization across diverse domains. This limitation highlights an opportunity for a synergistic approach: integrating even imperfect LLM outputs with symbolic planners can expedite heuristic searches, thereby enhancing efficiency and reducing search times (Fabiano et al 2023).…”

Section: Plan Generationmentioning

confidence: 99%

“…While LLMs attempt to mimic symbolic solvers through in-context learning for brain-inspired modules, this approach lacks adaptability and is a superficial understanding of complex cognitive processes. To overcome these issues, developing systems where neural and symbolic components are intrinsically intertwined is critical as it would accurately mirror human cognitive capabilities in planning (Fabiano et al 2023).…”

Section: Brain-inspired Planningmentioning

confidence: 99%

See 3 more Smart Citations

On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS)

Pallagani,

Muppasani,

Roy

et al. 2024

ICAPS

View full text Add to dashboard Cite

Automated Planning and Scheduling is among the growing areas in Artificial Intelligence (AI) where mention of LLMs has gained popularity. Based on a comprehensive review of 126 papers, this paper investigates eight categories based on the unique applications of LLMs in addressing various aspects of planning problems: language translation, plan generation, model construction, multi-agent planning, interactive planning, heuristics optimization, tool integration, and brain-inspired planning. For each category, we articulate the issues considered and existing gaps. A critical insight resulting from our review is that the true potential of LLMs unfolds when they are integrated with traditional symbolic planners, pointing towards a promising neuro-symbolic approach. This approach effectively combines the generative aspects of LLMs with the precision of classical planning methods. By synthesizing insights from existing literature, we underline the potential of this integration to address complex planning challenges. Our goal is to encourage the ICAPS community to recognize the complementary strengths of LLMs and symbolic planners, advocating for a direction in automated planning that leverages these synergistic capabilities to develop more advanced and intelligent planning systems. We aim to keep the categorization of papers updated on https://ai4society.github.io/LLM-Planning-Viz/, a collaborative resource that allows researchers to contribute and add new literature to the categorization.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Plan Generationmentioning

confidence: 99%

Section: Plan Generationmentioning

confidence: 99%

Section: Brain-inspired Planningmentioning

confidence: 99%

See 2 more Smart Citations

On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS)

Pallagani,

Muppasani,

Roy

et al. 2024

ICAPS

View full text Add to dashboard Cite

show abstract

“…This capacity is consolidated through repeated interactions with their environments. Using such a strategy, the SOFAI [207], [208] cognitive architecture employs a fast learner that continually improves by updating itself through the problem-solving experience of a slow reasoner. SOFAI uses a subsidiary meta-cognitive module to supervise the cognitive activities of the two main cognitive systems.…”

Section: Cognitive Architectures For Adversarial Robustnessmentioning

confidence: 99%

Robust appearance modeling for object detection and tracking: a survey of deep learning approaches

Mumuni

2022

Prog Artif Intell

View full text Add to dashboard Cite

We review current and emerging knowledge-informed and brain-inspired cognitive systems for realizing adversarial defenses, eXplainable Artificial Intelligence (XAI), and zero-shot or few-shot learning. Data-driven deep learning models have achieved remarkable performance and demonstrated capabilities surpassing human experts in many applications. Yet, their inability to exploit domain knowledge leads to serious performance limitations in practical applications. In particular, deep learning systems are exposed to adversarial attacks, which can trick them into making glaringly incorrect decisions. Moreover, complex data-driven models typically lack interpretability or explainability, i.e., their decisions cannot be understood by human subjects. Furthermore, models are usually trained on standard datasets with a closed-world assumption. Hence, they struggle to generalize to unseen cases during inference in practical open-world environments, thus, raising the zero-or few-shot generalization problem. Although many conventional solutions exist, explicit domain knowledge, brain-inspired neural network and cognitive architectures offer powerful new dimensions towards alleviating these problems. Prior knowledge is represented in appropriate forms and incorporated in deep learning frameworks to improve performance. Brain-inspired cognition methods use computational models that mimic the human mind to enhance intelligent behavior in artificial agents and autonomous robots. Ultimately, these models achieve better explainability, higher adversarial robustness and data-efficient learning, and can, in turn, provide insights for cognitive science and neurosciencethat is, to deepen human understanding on how the brain works in general, and how it handles these problems.Index Terms-Domain knowledge, cognitive architecture, brain-inspired neural network, explainable AI, adversarial attack, zero-shot generalization. ✦ INTRODUCTION State of deep learningMachine learning is an artificial intelligence (AI) concept that enables computing systems to learn useful relationships from data and then use this information to identify learned patterns and make predictions on new inputs. Deep learning (DL) is a machine learning method that uses a multi-layered arrangement of computational units to learn relevant patterns from data. Many deep learning constructs have been proposed for various use-cases. Some of the most popular ones are deep convolutional neural networks (DCNN) [1], [2] recurrent neural networks (RNN) [3], generative adversarial networks (GANs) [4], deep reinforcement learning [5] and vision transformers [6], [7]). DCNNs are a class of deep neural networks (DNNs) specially designed to handle image data.Recently, deep learning methods have accomplished remarkable milestones in many critical applications. In medical domains, some state-of-the-art models (e.g., as reported

show abstract

In Defence of Good Old-Fashioned Artificial Intelligence Approaches in Intelligent Transportation Systems

Vallati,

Chrpa

2023

2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC)

View full text Add to dashboard Cite

In recent years, Artificial Intelligence (AI) has been increasingly used in traffic management and control, particularly in the smart city context. However, the vast majority of recent AI-based approaches rely on data-driven black-box models that hinder the ability to understand the behaviour and dynamics that lead to a given output. On the contrary, Good Old-Fashioned Artificial Intelligence approaches that are based on symbolic models, such as automated planning, can provide the transparency and explainability needed in realworld applications.This paper focuses on the benefits of using automated planning techniques in Intelligent Transportation Systems (ITS), with a focus on explainability. A case study is presented to demonstrate how the components of an automated planning system can support explainability, the types of explanations that can be obtained, and the way in which such explanations can be generated.

show abstract

Fast and Slow Planning

Cited by 3 publications

References 23 publications

On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS)

On the Prospects of Incorporating Large Language Models (LLMs) in Automated Planning and Scheduling (APS)

Robust appearance modeling for object detection and tracking: a survey of deep learning approaches

In Defence of Good Old-Fashioned Artificial Intelligence Approaches in Intelligent Transportation Systems

Contact Info

Product

Resources

About