Long-term autonomy requires autonomous systems to adapt as their capabilities no longer perform as expected. To achieve this, a system must first be capable of detecting such changes. Creating and maintaining a system ontology is a comprehensive solution for this; an agent-maintained formal selfmodel will take the role of this system ontology. It would act as a repository of information about all the processes and functionality of the autonomous system, forming a systematic approach for detecting action failures. Our work will focus on Belief-Desire-Intention (BDI) [25] programming languages as they are well known for their use in developing intelligent agents [1, 6, 16, 21].

AOP (Agent-Oriented Programming) [30] is a programming paradigm that aims at providing effective languages and tools to develop agent-based software systems (e.g., [16]). The most challenging parts of the development of complex agent-based software systems are expected to be leveraged by AOP languages and tools. AOP languages and tools allow programmers to reason on high-level views of multi-agent systems instead of concentrating on fine-grained details that tend to distract attention from targeted problems. Examples of such challenging parts of the development of complex agent-based software systems are the deployment of agents to network hosts, the routing of messages across the network, and the handling of sensor information. The abstractions that AOP languages normally support and that make such languages suitable for agent-based software development are discussed, for example, in the specifications from FIPA (Foundation for Intelligent Physical Agents), now IEEE FIPA Standards Committee (http://www.fipa.org),

Autonomous systems developed with the Belief-Desire-Intention (BDI) architecture are usually mostly implemented in simulated environments. In this project we sought to build a BDI agent for use in the real world for campus mail delivery in the tunnel system at Carleton University. Ideally, the robot should receive a delivery order via a mobile application, pick up the mail at a station, navigate the tunnels to the destination station, and notify the recipient. We linked the Robot Operating System (ROS) with a BDI reasoning system to achieve a subset of the required use cases. ROS handles the low-level sensing and actuation, while the BDI reasoning system handles the high-level reasoning and decision making. Sensory data is orchestrated and sent from ROS to the reasoning system as perceptions. These perceptions are then deliberated upon, and an action string is sent back to ROS for interpretation and driving of the necessary actuator for the action to be performed. In this paper we present our current implementation, which closes the loop on the hardware-software integration, and implements a subset of the use cases required for the full system.

Risk management resulting from the actions and states of the different elements making up a operating room is a major concern during a surgical procedure. Agent-based simulation shows an interest through its interaction concepts, interactivity and autonomy of different simulator entities. We want in our study to implement a generator of alerts to listen the evolution of different settings applied to the simulator of agents (human fatigue, material efficiency, infection rate ...). This article presents our model, its implementation and the first results obtained. It should be noted that this study also made it possible to identify several scientific obstacles, such as the integration of different levels of abstraction, the coupling of species, the coexistence of several scales in the same environment and the deduction of unpredictable alerts. Case-based reasoning (CBR) is a beginning of response relative to the last lock mentioned and will be discussed in this paper.

Artificial intelligence is en route to changing all industries and the robotics industry is not an exception. Presently, the innovative combination of AI and robotics has created a number of futuristic possibilities, in all the industry domains. While most of us will agree that most robots will be humanoids in 10 years from now; in many environments, robots are designed to emulate a range of behaviors and physical abilities will reflect a best fit for those characteristics. An exception will likely be robots that provide medical or other care or companionship for humans, and perhaps service robots that are meant to establish a more personal and'humanized' relationship. Though related, some would argue that the correct term is machine vision or robot vision rather than computer vision, because "robots seeing" involves more than just computer algorithms; engineers and roboticists also have to account for camera hardware that allow robots to process physical data.

The ability to jointly understand the geometry of objects and plan actions for manipulating them is crucial for intelligent agents. We refer to this ability as geometric planning. Recently, many interactive environments have been proposed to evaluate intelligent agents on various skills, however, none of them cater to the needs of geometric planning. We present PackIt, a virtual environment to evaluate and potentially learn the ability to do geometric planning, where an agent needs to take a sequence of actions to pack a set of objects into a box with limited space. We also construct a set of challenging packing tasks using an evolutionary algorithm. Further, we study various baselines for the task that include model-free learning-based and heuristic-based methods, as well as search-based optimization methods that assume access to the model of the environment. Code and data are available at https://github.com/princeton-vl/PackIt.

When a company relies on distributed users' data to train a machine learning model, federated learning [37, 54, 25] promotes the idea that users/customers' data should be kept local, and only the locally held/learned hypothesis will be shared/contributed from each user. While federated learning has observed success in keyboard recognition [21] and in language modeling [8], existing works have made an implicit assumption that participating users will be willing to contribute their local hypotheses to help the central entity to refine the model. Nonetheless, without proper incentives, agents can choose to opt out of the participation, to contribute either uninformative or outdated information, or to even contribute malicious model information. Though being an important question for federated learning [54, 34, 55, 56], this capability of providing adequate incentives for user participation has largely been overlooked. In this paper we ask the questions that: Can a machine learning hypothesis be incentivized/elicited by a certain form of scoring rules from self-interested agents? The availability of a scoring rule will help us design a payment for the elicited hypothesis properly to motivate the reporting of high-quality ones. The corresponding solutions complement the literature of federated learning by offering a generic template for incentivizing users' participation.

Deep counterfactual value networks combined with continual resolving provide a way to conduct depth-limited search in imperfect-information games. However, since their introduction in the DeepStack poker AI, deep counterfactual value networks have not seen widespread adoption. In this paper we introduce several improvements to deep counterfactual value networks, as well as counterfactual regret minimization, and analyze the effects of each change. We combined these improvements to create the poker AI Supremus. We show that while a reimplementation of DeepStack loses head-to-head against the strong benchmark agent Slumbot, Supremus successfully beats Slumbot by an extremely large margin and also achieves a lower exploitability than DeepStack against a local best response. Together, these results show that with our key improvements, deep counterfactual value networks can achieve state-of-the-art performance.

We present the Battlesnake Challenge, a framework for multi-agent reinforcement learning with Human-In-the-Loop Learning (HILL). It is developed upon Battlesnake, a multiplayer extension of the traditional Snake game in which 2 or more snakes compete for the final survival. The Battlesnake Challenge consists of an offline module for model training and an online module for live competitions. We develop a simulated game environment for the offline multi-agent model training and identify a set of baseline heuristics that can be instilled to improve learning. Our framework is agent-agnostic and heuristics-agnostic such that researchers can design their own algorithms, train their models, and demonstrate in the online Battlesnake competition. We validate the framework and baseline heuristics with our preliminary experiments. Our results show that agents with the proposed HILL methods consistently outperform agents without HILL. Besides, heuristics of reward manipulation had the best performance in the online competition. We open source our framework at https://github.com/awslabs/sagemaker-battlesnake-ai.

Agent unified modeling languages (AUML) are agent-oriented approaches that supports the specification, design, visualization and documentation of an agent-based system. This paper presents the use of Prometheus AUML approach for the modeling of a Pre-assessment System of five interactive agents. The Pre-assessment System, as previously reported, is a multi-agent based e-learning system that is developed to support the assessment of prior learning skills in students so as to classify their skills and make recommendation for their learning. This paper discusses the detailed design approach of the system in a step-by-step manner; and domain knowledge abstraction and organization in the system. In addition, the analysis of the data collated and models of prediction for future pre-assessment results are also presented.