A Comparative Approach to Understanding General Intelligence: Predicting Cognitive Performance in an Open-ended Dynamic Task

This article explores the characteristics and scope of artificial intelligence (AI) competitions in medical imaging. A retrospective evaluation of AI competitions related to medical imaging was conducted between 2017 and 2023. Relevant terms associated with AI and competitions were searched using the PubMed database and the grand-challenge website, and applicable studies were included in the review. The 26 AI competitions included in the review covered a wide range of topics, from brain imaging to extremities and from stroke detection to bone age estimation, with many organized through international collaborations between engineering and medical professionals. Various national screening and teleradiology databases, as well as university databases, were used. Teams from different regions worldwide participated in these competitions. These initiatives contribute to the global adoption of AI technologies in healthcare. Moreover, they help raise awareness among high school students, medical students, radiology trainees, and young radiologists of the intersection between AI and medical imaging. AI competitions play a crucial role in fostering collaboration between the medical field and AI, driving innovation, and increasing societal awareness of AI applications in healthcare.

This study proposes a distributed formation control framework for multi-agent systems by integrating deep reinforcement learning (DRL) with interdisciplinary methodologies from control theory, robotics, and artificial intelligence. Aimed at addressing collaborative formation challenges in dynamic environments, the approach uses graph-theoretic modeling for formation-keeping and obstacle avoidance and dynamic Attention-based Advantage Actor-Critic (DA-A2C) algorithm to solve collaborative formation problems in dynamic environments like autonomous drone swarms and robotic fleets. To balance formation consistency and obstacle repulsion, this strategy optimizes neighbor perception dynamic weight allocation using the dynamic Ω-zone mechanism. Experimental validations in simulated environments with static/moving obstacles show superior performance over SA2C, MATD3, and DDPG baselines, covering formation control and obstacle avoidance in complex situations and outperforming other algorithms in adaptability and robustness. This research can improve cross-domain applications in intelligent transportation, industrial automation, and emergency response systems. The method works in multi-robot formation systems, and future study will focus on 3D environment generalization and efficient communication protocols.

Conventional planning strategies often fail due to the implementation of top-down, fixed, and rigid designs in environments that are, themselves, constantly evolving. This research provides an alternative theoretical approach to understand and address problems in dynamic environments using Complex Adaptive Systems (CAS) approach. In current complexity research, spatial planners typically engage complexity principles through computer models in the form of agent-based and cellular automata models. However, to date, there has been a lack of explicit attention towards activating CAS dynamics beyond computer models to instead examine how CAS dynamics might play out, in real-time, within the physical environment. With the increasing prevalence of digital signals and the ‘internet of things,’ spatial environment artifacts can now be conceptualized as computational devices capable of learning through sensing. This research aims to demonstrate the potential operationalization of CAS dynamics on the ground, and their ability to address issues in the built environment, by developing a global coastline waste collection system: Autonomous Trash Skimmers. By conceptualizing these Skimmers as ‘smart’ agents capable of self-organization through the means of digital signals, the research provides a novel application of CAS dynamics - operationalizing these in the real world and moving beyond computer models.

This paper presents the evolutionary programming with an ensemble of memories to deal with optimization problems in dynamic environments. The proposed algorithm modifies a recent version of evolutionary programming by introducing a simulated-annealing-like dynamic strategy parameter as well as applying local search towards the most improving directions. Diversity of the population is enhanced by an ensemble of external archives that serve as short-term and long-term memories. The archive members also act as the basic solutions when environmental changes occur. The algorithm is tested on a set of 6 multimodal problems with a total 49 change instances provided by CEC 2009 Competition on Evolutionary Computation in Dynamic and Uncertain Environments and the results are presented.

The increasing volume of traffic has led to severe challenges, including traffic congestion, heightened energy consumption, increased air pollution, and prolonged travel times. Addressing these issues requires innovative approaches for optimizing road network utilization. While Deep Reinforcement Learning (DRL)-based methods have shown remarkable effectiveness in dynamic scenarios like traffic management, their primary focus has been on single-agent setups, limiting their applicability to real-world multi-agent systems. Managing agents and fostering collaboration in a multi-agent reinforcement learning scenario remains a challenging task. This paper introduces a cooperative multi-agent federated reinforcement learning algorithm named FLDQN to address the challenge of agent cooperation by solving travel time minimization challenges in dynamic multi-agent reinforcement learning (MARL) scenarios. FLDQN leverages federated learning to facilitate collaboration and knowledge sharing among intelligent agents, optimizing vehicle routing and reducing congestion in dynamic traffic environments. Using the SUMO simulator, multiple agents equipped with deep Q-learning models interact with their local environments, share model updates via a federated server, and collectively enhance their policies using unique local observations while benefiting from the collective experiences of other agents. Experimental evaluations demonstrate that FLDQN achieves a significant average reduction of over 34.6% in travel time compared to non-cooperative methods while simultaneously lowering the computational overhead through distributed learning. FLDQN underscores the vital impact of agent cooperation and provides an innovative solution for enabling agent cooperation in a multi-agent environment.

Recent models of spiking neuronal networks have been trained to perform behaviors in static environments using a variety of learning rules, with varying degrees of biological realism. Most of these models have not been tested in dynamic visual environments where models must make predictions on future states and adjust their behavior accordingly. The models using these learning rules are often treated as black boxes, with little analysis on circuit architectures and learning mechanisms supporting optimal performance. Here we developed visual/motor spiking neuronal network models and trained them to play a virtual racket-ball game using several reinforcement learning algorithms inspired by the dopaminergic reward system. We systematically investigated how different architectures and circuit-motifs (feed-forward, recurrent, feedback) contributed to learning and performance. We also developed a new biologically-inspired learning rule that significantly enhanced performance, while reducing training time. Our models included visual areas encoding game inputs and relaying the information to motor areas, which used this information to learn to move the racket to hit the ball. Neurons in the early visual area relayed information encoding object location and motion direction across the network. Neuronal association areas encoded spatial relationships between objects in the visual scene. Motor populations received inputs from visual and association areas representing the dorsal pathway. Two populations of motor neurons generated commands to move the racket up or down. Model-generated actions updated the environment and triggered reward or punishment signals that adjusted synaptic weights so that the models could learn which actions led to reward. Here we demonstrate that our biologically-plausible learning rules were effective in training spiking neuronal network models to solve problems in dynamic environments. We used our models to dissect the circuit architectures and learning rules most effective for learning. Our model shows that learning mechanisms involving different neural circuits produce similar performance in sensory-motor tasks. In biological networks, all learning mechanisms may complement one another, accelerating the learning capabilities of animals. Furthermore, this also highlights the resilience and redundancy in biological systems.

The road sign in dynamic traffic system is an important element which affects on human cognitive performance on driving. Web-based vision system simulator was developed to examine the cognition time of the road sign in dynamic environment. This experiment was designed in within-subject design with two factors: vehicle speed and the amount of information of the traffic sign. It measured the cognition time of the road sign through two evaluation methods: the subjective test with vision system simulator and computational cognitive model. In these two evaluations of human cognitive performance under the dynamic traffic environment. It demonstrated that subject`s cognition time was affected by both the amount of information of traffic sign and driving speed.

Botball is a national program in which teams of middle and high school students design, build, and program small autonomous mobile robots to compete in a highly charged interactive (but non-destructive) tournament. Botball students learn to program in c, construct feedback and control loops, create electromechanical systems, and integrate it all together while they work on a team. Botball takes place in regional tournaments across the country and culminates in a National Botball Tournament traditionally hosted by the American Association for Artificial Intelligence at its annual conference. This program puts reusable equipment into schools and, at the Botball Teacher Workshops, trains teachers in robotics and the integration of robotics into their curriculum. Botball appeals to a wide variety of students and brings out the best in each, challenging them to solve real-world problems in a dynamic environment at their own level.

Many applications can be formalized as dynamic problems and their main goals are to track the moving optimum. Many algorithms that address efficiently static problems, fail when applied to dynamic problems due to their inability to adapt and respond to changes in the environment. A different topology multi-swarm PSO is proposed to deal with problems in dynamic environment. The proposed algorithm integrates two different topology subswarms and the two subswarms exchange their best particle at checkpoints. One rapidly convergence to the optima and the other is responsible for local search and diversity reservation. The experiments demonstrate the proposed algorithm is effective and stable in dynamic environment.

PurposeThe purpose of this paper is to improve the traditional bug algorithms for the navigation of mobile robots in unknown environments by considering the limitations in previous works such as generating long path, limited to static environments as well as ignoring implementation issues. With this purpose, a new bug‐type algorithm termed Distance Histogram Bug (DH‐Bug) is proposed for overcoming these limitations.Design/methodology/approachDH‐Bug redefines the traditional motion modes and switching criteria to shorten the path length. In order to extend the framework of the traditional bug algorithms to tackle the navigation problem in dynamic environments, a new mode and the related switching conditions are designed for dealing with moving obstacles. Moreover, a realization method termed the Distance Histogram (DH) method which takes many implementation issues into full account is proposed for implementing DH‐Bug on real robots.FindingsDH‐Bug is convergent in static environments and it also guarantees “approximate” convergence based on several reasonable assumptions when there are moving obstacles. Simulations results show that DH‐Bug generates shorter average path length than some classical Bug algorithms in static environments and it also performs well in most simulations that contain moving obstacles except for some extremely adverse scenarios which have been discussed in the paper. Experiments on real robots further verify its applicability in both static environments and dynamic environments containing moving obstacles.Originality/valueCompared with previous works, DH‐Bug has three main contributions. First, in static environments, it can shorten the average path length than many classical bug algorithms in the premise of guaranteeing convergence. Second, it can be applied in dynamic environments containing moving obstacles. Third, unlike the previous bug algorithms that always ignore the practical implementation issues, DH‐Bug presents not only an abstract concept, but also a realization approach for realizing this concept on real robots.

Visual localization has been well studied in recent decades and applied in many fields as a fundamental capability in robotics. However, the success of the state of the arts usually builds on the assumption that the environment is static. In dynamic scenarios where moving objects are present, the performance of the existing visual localization systems degrades a lot due to the disturbance of the dynamic factors. To address this problem, we propose a novel sparse motion removal (SMR) model that detects the dynamic and static regions for an input frame based on a Bayesian framework. The similarity between the consecutive frames and the difference between the current frame and the reference frame are both considered to reduce the detection uncertainty. After the detection process is finished, the dynamic regions are eliminated while the static ones are fed into a feature-based visual simultaneous localization and mapping (SLAM) system for further visual localization. To verify the proposed method, both qualitative and quantitative experiments are performed and the experimental results have demonstrated that the proposed model can significantly improve the accuracy and robustness for visual localization in dynamic environments.<;/p><;p><;i>Note to Practitioners<;/i>-This article was motivated by the visual localization problem in dynamic environments. Visual localization is well applied in many robotic fields such as path planning and exploration as the basic capability for a mobile robot. In the GPS-denied environments, one robot needs to localize itself through perceiving the unknown environment based on a visual sensor. In real-world scenes, the existence of the moving objects will significantly degrade the localization accuracy, which makes the robot implementation unreliable. In this article, an SMR model is designed to handle this problem. Once receiving a frame, the proposed model divides it into dynamic and static regions through a Bayesian framework. The dynamic regions are eliminated, while the static ones are maintained and fed into a feature-based visual SLAM system for further visual localization. The proposed method greatly improves the localization accuracy in dynamic environments and guarantees the robustness for robotic implementation.

In this study, we propose a solution to the simultaneous localization and mapping (SLAM) problem in low dynamic environments by using a pose graph and an RGB-D (red-green-blue depth) sensor. The low dynamic environments refer to situations in which the positions of objects change over long intervals. Therefore, in the low dynamic environments, robots have difficulty recognizing the repositioning of objects unlike in highly dynamic environments in which relatively fast-moving objects can be detected using a variety of moving object detection algorithms. The changes in the environments then cause groups of false loop closing when the same moved objects are observed for a while, which means that conventional SLAM algorithms produce incorrect results. To address this problem, we propose a novel SLAM method that handles low dynamic environments. The proposed method uses a pose graph structure and an RGB-D sensor. First, to prune the falsely grouped constraints efficiently, nodes of the graph, that represent robot poses, are grouped according to the grouping rules with noise covariances. Next, false constraints of the pose graph are pruned according to an error metric based on the grouped nodes. The pose graph structure is reoptimized after eliminating the false information, and the corrected localization and mapping results are obtained. The performance of the method was validated in real experiments using a mobile robot system.

Many artificial intelligence (AI) theoretical models are inspired by human/natural/social intelligence mechanisms. Three main schools of artificial intelligence have been formed, that is, symbolism, connectionism and behaviorism. Cognitive computing is one of the key fields of AI. It is a critical task for AI researchers to develop advanced cognitive computing models. Cognitive computing is the third and the most transformational phase in computing’s evolution, after the Tabulating Era and Programming Era. Inspired by human’s granularity thinking based problem solving mechanism and the cognition law of “global precedence”, a data-driven granular cognitive computing model (DGCC) is proposed in this paper. It integrates two contradictory mechanisms, namely, human’s cognition mechanism of “global precedence” which is a cognition process of “from coarser to finer” and the information processing mechanism of machine learning systems which is “from finer to coarser”. According to DGCC, deep learning is taken as a combination of symbolism and connectionism, and named hierarchical structuralism in this paper.

To address the resource allocation problem in dynamic environments where multiple unmanned aerial vehicle base stations (UAV-BSs) provide efficient downlink services to ground users, this paper proposes a novel hierarchical decision-making mechanism based on the Proximal Policy Optimization (PPO) algorithm. The proposed method optimizes time-frequency resource allocation in the downlink, aiming to maximize the total user throughput over multiple time slots. By constructing channel and interference models, the complex multi-channel resource allocation problem is decomposed into a series of single-channel decision subproblems, significantly reducing the action space complexity. Specifically, the original exponential complexity O(NM) (where N is the number of users and M is the number of channels) is reduced to a linear complexity O(N), effectively alleviating the curse of dimensionality. Simulation results demonstrate that the proposed hierarchical architecture, integrated with the PPO algorithm, achieves superior performance in terms of total throughput, convergence speed, and stability compared to existing methods. This study provides new insights and technical support for efficient resource management in UAV-BS systems operating in complex and dynamic environments.

This article proposes two modifications of Gaussian processes, which aim to deal with dynamic environments. One is a weight decay method that gradually forgets the old data, and the other is a time stamp method that regards the time course of data as a Gaussian process. We show experimental results when these modifications are applied to regression problems in dynamic environments.