Predicting the long-term collective behaviour of fish pairs with deep learning.

Modern computing has enhanced our understanding of how social interactions shape collective behaviour in animal societies. Although analytical models dominate in studying collective behaviour, this study introduces a deep learning model to assess social interactions in the fish species Hemigrammus rhodostomus. We compare the results of our deep learning approach with experiments and with the results of a state-of-the-art analytical model. To that end, we propose a systematic methodology to assess the faithfulness of a collective motion model, exploiting a set of stringent individual and collective spatio-temporal observables. We demonstrate that machine learning (ML) models of social interactions can directly compete with their analytical counterparts in reproducing subtle experimental observables. Moreover, this work emphasizes the need for consistent validation across different timescales, and identifies key design aspects that enable our deep learning approach to capture both short- and long-term dynamics. We also show that our approach can be extended to larger groups without any retraining, and to other fish species, while retaining the same architecture of the deep learning network. Finally, we discuss the added value of ML in the context of the study of collective motion in animal groups and its potential as a complementary approach to analytical models.

Modelling the sustainability of a primary school digital education curricular reform and professional development program.

Sustaining changes in teachers' practices is a challenge that determines the success of curricular reforms, from which Digital Education (DE) is not exempt. As the literature on sustainability is considered "scarce" and "scattered", long-term studies modelling the factors impacting teachers' sustained uptake of DE pedagogical content are lacking. Thus, we investigate whether and how 287 in-service teachers sustained a primary school DE curricular reform over a year after they completed their two-year DE professional development program. We model the sustainability of the reform through Structural Equation Modelling, and identify critical sustainability-factors. The validated Sustainable Adoption of Digital Education (SADE) model confirms that sustainability in the fourth year of the reform depends on perceived usefulness of teaching the new content, ease of implementation, and access to sufficient support in schools. Such factors should thus be evaluated, accounted for in the implementation phase of the reform, and sustained over time. The findings confirm that the DE curricular reform model contributes to positive self-efficacy to teach DE, provides sufficient in-school support, and promotes increasing adoption over time. However, as teachers' practices have not yet stabilised, and teachers may still adopt more to cover the breadth of DE-concepts, it is important to remain attentive to remaining sustainability barriers: lack of time, effort required to teach DE with teachers preferring to delegate, and lack of student-learning evidence, the latter being a significant challenge to address in the literature. These barriers must therefore be jointly addressed by researchers and practitioners in the field in order to promote the sustainability of the reform.

A robotic honeycomb for interaction with a honeybee colony.

Robotic technologies have shown the capability to interact with living organisms and even to form integrated mixed societies composed of living and artificial agents. Biocompatible robots, incorporating sensing and actuation capable of generating and responding to relevant stimuli, can be a tool to study collective behaviors previously unattainable with traditional techniques. To investigate collective behaviors of the western honeybee (Apis mellifera), we designed a robotic system capable of observing and modulating the bee cluster using an array of thermal sensors and actuators. We initially integrated the system into a beehive populated with about 4000 bees for several months. The robotic system was able to observe the colony by continuously collecting spatiotemporal thermal profiles of the winter cluster. Furthermore, we found that our robotic device reliably modulated the superorganism's response to dynamic thermal stimulation, influencing its spatiotemporal reorganization. In addition, after identifying the thermal collapse of a colony, we used the robotic system in a "life-support" mode via its thermal actuators. Ultimately, we demonstrated a robotic device capable of autonomous closed-loop interaction with a cluster comprising thousands of individual bees. Such biohybrid societies open the door to investigation of collective behaviors that necessitate observing and interacting with the animals within a complete social context, as well as for potential applications in augmenting the survivability of these pollinators crucial to our ecosystems and our food supply.

Social Integrating Robots Suggest Mitigation Strategies for Ecosystem Decay.

We develop here a novel hypothesis that may generate a general research framework of how autonomous robots may act as a future contingency to counteract the ongoing ecological mass extinction process. We showcase several research projects that have undertaken first steps to generate the required prerequisites for such a technology-based conservation biology approach. Our main idea is to stabilise and support broken ecosystems by introducing artificial members, robots, that are able to blend into the ecosystem's regulatory feedback loops and can modulate natural organisms' local densities through participation in those feedback loops. These robots are able to inject information that can be gathered using technology and to help the system in processing available information with technology. In order to understand the key principles of how these robots are capable of modulating the behaviour of large populations of living organisms based on interacting with just a few individuals, we develop novel mathematical models that focus on important behavioural feedback loops. These loops produce relevant group-level effects, allowing for robotic modulation of collective decision making in social organisms. A general understanding of such systems through mathematical models is necessary for designing future organism-interacting robots in an informed and structured way, which maximises the desired output from a minimum of intervention. Such models also help to unveil the commonalities and specificities of the individual implementations and allow predicting the outcomes of microscopic behavioural mechanisms on the ultimate macroscopic-level effects. We found that very similar models of interaction can be successfully used in multiple very different organism groups and behaviour types (honeybee aggregation, fish shoaling, and plant growth). Here we also report experimental data from biohybrid systems of robots and living organisms. Our mathematical models serve as building blocks for a deep understanding of these biohybrid systems. Only if the effects of autonomous robots onto the environment can be sufficiently well predicted can such robotic systems leave the safe space of the lab and can be applied in the wild to be able to unfold their ecosystem-stabilising potential.

The symbiotic relationship between educational robotics and computer science in formal education.

Educational Robotics (ER) has the potential to provide significant benefits to education, provided an increase in outreach by transitioning from the extra-curricular initiatives in which ER has thrived to formal education. As Computer Science (CS) Education is undergoing curricular reforms worldwide, the present study addresses the case of a Digital Education reform that included ER as a means to teach core CS concepts. Approximately 350 teachers from the first four grades of primary school participated in a mandatory two-year continuing professional development (CPD) program. The first year of the program was dedicated to CS and introduced teachers to CS Unplugged (CSU) and Robotics Unplugged (RU) activities. As such, we analyse the interplay between these activities and focus on teachers' voluntary adoption of the proposed content in classrooms. This is complemented by an analysis of their perception and recommendation of ER. The findings highlight three main points. Firstly, ER benefits from the integration in the CS CPD, as this provides the necessary traction to introduce ER into teacher practices (the teachers freely devoted 2275 h to ER activities in their classrooms, over two years). Secondly, the presence of ER activities in the CS-CPD allows a higher proportion of teachers to adopt the CS content, as there are teachers that favour one type of activity over the other. Finally, the globally positive perception of ER registered in this study is relevant for two reasons: teachers were not voluntarily participating in the CPD, and results did not differ between pioneers and novices.

A computer science and robotics integration model for primary school: evaluation of a large-scale in-service K-4 teacher-training program.

Integrating computer science (CS) into school curricula has become a worldwide preoccupation. Therefore, we present a CS and Robotics integration model and its validation through a large-scale pilot study in the administrative region of the Canton Vaud in Switzerland. Approximately 350 primary school teachers followed a mandatory CS continuing professional development program (CPD) of adapted format with a curriculum scaffolded by instruction modality. This included CS Unplugged activities that aim to teach CS concepts without the use of screens, and Robotics Unplugged activities that employed physical robots, without screens, to learn about robotics and CS concepts. Teachers evaluated positively the CPD and their representation of CS improved. Voluntary adoption rates reached 97% during the CPD and 80% the following year. These results combined with the underpinning literature support the generalisability of the model to other contexts.

Strategies to modulate zebrafish collective dynamics with a closed-loop biomimetic robotic system.

The objective of this study is to integrate biomimetic robots into small groups of zebrafish and to modulate their collective behaviours. A possible approach is to have the robots behave like sheepdogs. In this case, the robots would behave like a different species than the fish and would present different relevant behaviours. In this study, we explore different strategies that use biomimetic zebrafish behaviours. In past work, we have shown that robots biomimicking zebrafish can be socially integrated into zebrafish groups. We have also shown that a fish-like robot can modulate the rotation choice of zebrafish groups in a circular set-up. Here, we further study the modulation capabilities of such robots in a more complex set-up. To do this, we exploit zebrafish social behaviours we identified in previous studies. We first modulate collective departure by replicating the leadership mechanisms with the robot in a set-up composed of two rooms connected by a corridor. Then, we test different behavioural strategies to drive the fish groups towards a predefined target room. To drive the biohybrid groups towards a predefined choice, they have to adopt some specific fish-like behaviours. The first strategy is based on a single robot using the initiation behaviour. In this case, the robot keeps trying to initiate a group transition towards the target room. The second strategy is based on two robots, one initiating and one staying in the target room as a social attractant. The third strategy is based on a single robot behaving like a zebrafish but staying in the target room as a social attractant. The fourth strategy uses two robots behaving like zebrafish but staying in the target room. We conclude that robots can modulate zebrafish group behaviour by adopting strategies based on existing fish behaviours. Under these conditions, robots enable the testing of hypotheses about the behaviours of fish.

Bidirectional interactions facilitate the integration of a robot into a shoal of zebrafish Danio rerio.

Many studies on collective animal behavior seek to identify the individual rules that underlie collective patterns. However, it was not until the recent advancements of micro-electronic and embedded systems that scientists were able to create mixed groups of sensor-rich robots and animals and study collective interactions from the within a bio-hybrid group. In recent work, scientists showed that a robot-controlled lure is capable of influencing the collective decisions of zebrafish Danio rerio shoals moving in a ring and a two-room setup. Here, we study a closely related topic, that is, the collective behavior patterns that emerge when different behavioral models are reproduced through the use of a robotic lure. We design a behavioral model that alternates between obeying and disobeying the collective motion decisions in order to become socially accepted by the shoal members. Subsequently, we compare it against two extreme cases: a reactive and an imposing decision model. For this, we use spatial, directional and information theoretic metrics to measure the degree of integration of the robotic agent. We show that our model leads to similar information flow as in freely roaming shoals of zebrafish and exhibits leadership skills more often than the open-loop models. Thus, in order for the robot to achieve higher degrees of integration in the zebrafish shoal, it must, like any other shoal member, be bidirectionally involved in the decision making process. These findings provide insight on the ability to form mixed societies of animals and robots and yield promising results on the degree to which a robot can influence the collective decision making.

Autonomous Exploration for Radioactive Hotspots Localization Taking Account of Sensor Limitations.

Effective radioactive hotspot localization and detection is limited by sensor characteristics (i.e., the long acquisition time and poor angular resolution AR of a gamma camera) that significantly degrade the performance of autonomous exploration in terms of the completion time and accuracy. The goal of this research is to study effective exploration algorithms that take into account these specific sensor limitations. These exploration algorithms are adapted and implemented based on behaviour-based and multi-criteria decision making MCDM approaches on an autonomous robot. The algorithms were also tested in simulation and validated by experiments performed on a real robot. According to the results, the algorithms demonstrate the ability to mitigate the unfavourable effects of the limitations.

Robots mediating interactions between animals for interspecies collective behaviors.

Self-organized collective behavior has been analyzed in diverse types of gregarious animals. Such collective intelligence emerges from the synergy between individuals, which behave at their own time and spatial scales and without global rules. Recently, robots have been developed to collaborate with animal groups in the pursuit of better understanding their decision-making processes. These biohybrid systems make cooperative relationships between artificial systems and animals possible, which can yield new capabilities in the resulting mixed group. However, robots are currently tailor-made to successfully engage with one animal species at a time. This limits the possibilities of introducing distinct species-dependent perceptual capabilities and types of behaviors in the same system. Here, we show that robots socially integrated into animal groups of honeybees and zebrafish, each one located in a different city, allowing these two species to interact. This interspecific information transfer is demonstrated by collective decisions that emerge between the two autonomous robotic systems and the two animal groups. The robots enable this biohybrid system to function at any distance and operates in water and air with multiple sensorimotor properties across species barriers and ecosystems. These results demonstrate the feasibility of generating and controlling behavioral patterns in biohybrid groups of multiple species. Such interspecies connections between diverse robotic systems and animal species may open the door for new forms of artificial collective intelligence, where the unrivaled perceptual capabilities of the animals and their brains can be used to enhance autonomous decision-making, which could find applications in selective "rewiring" of ecosystems.

How mimetic should a robotic fish be to socially integrate into zebrafish groups?

Biomimetic robots are promising tools in animal behavioural studies. If they are socially integrated in a group of animals, they can produce calibrated social stimuli to test the animal responses. However, the design of such social robots is challenging as it involves both a luring capability including appropriate robot behaviours, and the acceptation of the robots by the animals as social companions. Here, we investigate the integration of a biomimetic robot driven by biomimetic behavioural models into a group of zebrafish (Danio rerio). The robot behaviours are based on a stochastic model linking zebrafish visual perception to individual behaviour and calibrated experimentally to correspond to the behaviour of zebrafish. We show that our robot can be integrated into a group of zebrafish, mimic their behaviour and exhibit similar collective dynamics compared to fish-only groups. This study shows that an autonomous biomimetic robot was enhanced by a biomimetic behavioural model so that it can socially integrate into groups of fish.

Publisher Correction: Mergeable nervous systems for robots.

The original version of this Article contained an error in the author contributions section, whereby credit for design of the experiments was not attributed to N.M. This error has now been corrected in both the PDF and HTML versions of the Article.

Mergeable nervous systems for robots.

Robots have the potential to display a higher degree of lifetime morphological adaptation than natural organisms. By adopting a modular approach, robots with different capabilities, shapes, and sizes could, in theory, construct and reconfigure themselves as required. However, current modular robots have only been able to display a limited range of hardwired behaviors because they rely solely on distributed control. Here, we present robots whose bodies and control systems can merge to form entirely new robots that retain full sensorimotor control. Our control paradigm enables robots to exhibit properties that go beyond those of any existing machine or of any biological organism: the robots we present can merge to form larger bodies with a single centralized controller, split into separate bodies with independent controllers, and self-heal by removing or replacing malfunctioning body parts. This work takes us closer to robots that can autonomously change their size, form and function.Robots that can self-assemble into different morphologies are desired to perform tasks that require different physical capabilities. Mathews et al. design robots whose bodies and control systems can merge and split to form new robots that retain full sensorimotor control and act as a single entity.

Self-organized coordinated motion in groups of physically connected robots.

An important goal of collective robotics is the design of control systems that allow groups of robots to accomplish common tasks by coordinating without a centralized control. In this paper, we study how a group of physically assembled robots can display coherent behavior on the basis of a simple neural controller that has access only to local sensory information. This controller is synthesized through artificial evolution in a simulated environment in order to let the robots display coordinated-motion behaviors. The evolved controller proves to be robust enough to allow a smooth transfer from simulated to real robots. Additionally, it generalizes to new experimental conditions, such as different sizes/shapes of the group and/or different connection mechanisms. In all these conditions the performance of the neural controller in real robots is comparable to the one obtained in simulation.