High-level Action Research Articles

Hierarchical Active Tracking Control for UAVs via Deep Reinforcement Learning

Active tracking control is essential for UAVs to perform autonomous operations in GPS-denied environments. In the active tracking task, UAVs take high-dimensional raw images as input and execute motor actions to actively follow the dynamic target. Most research focuses on three-stage methods, which entail perception first, followed by high-level decision-making based on extracted spatial information of the dynamic target, and then UAV movement control, using a low-level dynamic controller. Perception methods based on deep neural networks are powerful but require considerable effort for manual ground truth labeling. Instead, we unify the perception and decision-making stages using a high-level controller and then leverage deep reinforcement learning to learn the mapping from raw images to the high-level action commands in the V-REP-based environment, where simulation data are infinite and inexpensive. This end-to-end method also has the advantages of a small parameter size and reduced effort requirements for parameter turning in the decision-making stage. The high-level controller, which has a novel architecture, explicitly encodes the spatial and temporal features of the dynamic target. Auxiliary segmentation and motion-in-depth losses are introduced to generate denser training signals for the high-level controller’s fast and stable training. The high-level controller and a conventional low-level PID controller constitute our hierarchical active tracking control framework for the UAVs’ active tracking task. Simulation experiments show that our controller trained with several augmentation techniques sufficiently generalizes dynamic targets with random appearances and velocities, and achieves significantly better performance, compared with three-stage methods.

An Action Plan to Modernize Gender, Sex, and Sexual Orientation Information Practices in Canadian Electronic Health Record Systems

This brief report contains a summary version of the proposed action plan to modernize gender, sex, and sexual orientation (GSSO) information practices in Canadian electronic health record (EHR) systems. The plan was co-created by a research team with guidance from the Infoway Sex and Gender Working Group, organizations and communities involved with sexual and gender minorities (SGM), and those with lived and living experiences as SGM. The plan describes seven high level actions needed to improve the principles, approaches, policies, and processes of defining, collecting, organizing, sharing, and using GSSO information in EHRs.

Coordinating Shared Tasks in Human-Robot Collaboration by Commands.

Human-robot collaboration is gaining more and more interest in industrial settings, as collaborative robots are considered safe and robot actions can be programmed easily by, for example, physical interaction. Despite this, robot programming mostly focuses on automated robot motions and interactive tasks or coordination between human and robot still requires additional developments. For example, the selection of which tasks or actions a robot should do next might not be known beforehand or might change at the last moment. Within a human-robot collaborative setting, the coordination of complex shared tasks, is therefore more suited to a human, where a robot would act upon requested commands.In this work we explore the utilization of commands to coordinate a shared task between a human and a robot, in a shared work space. Based on a known set of higher-level actions (e.g., pick-and-placement, hand-over, kitting) and the commands that trigger them, both a speech-based and graphical command-based interface are developed to investigate its use. While speech-based interaction might be more intuitive for coordination, in industrial settings background sounds and noise might hinder its capabilities. The graphical command-based interface circumvents this, while still demonstrating the capabilities of coordination. The developed architecture follows a knowledge-based approach, where the actions available to the robot are checked at runtime whether they suit the task and the current state of the world. Experimental results on industrially relevant assembly, kitting and hand-over tasks in a laboratory setting demonstrate that graphical command-based and speech-based coordination with high-level commands is effective for collaboration between a human and a robot.

Tam pam pam pam and mi – fa – sol: constituting musical instructions through multimodal interaction in orchestra rehearsals

Abstract Using multimodal (inter)action/conversation analysis, the present contribution inventories the repertoire of higher-level actions that constitute musical instruction in orchestra rehearsals. The study describes the modal complexity of the instructional actions as built from a varied combination of speech, gesture, gaze, vocalizing and body posture/movement. A high modal intensity of speech and vocalizing is explained with recourse to their contextually useful modal reaches. While some modes, like vocalizing and body posture appear to be action-specific, others turn out to be pervasive default modes. Besides modal intensity, the study also attends to the transitioning between higher-level actions through gaze and the role of the score as frozen action. The analyses help demystify orchestra rehearsals as a special type of professional communicative interaction, which builds on a rich multimodal texture motivated by recurring instructional functions. The methodological rationale demonstrated will be suited to exploring the social variation of instructional interaction in orchestra rehearsals.

Identifying Priorities, Targets, and Actions for the Long-term Social and Ecological Management of Invasive Non-Native Species

Formulating effective management plans for addressing the impacts of invasive non-native species (INNS) requires the definition of clear priorities and tangible targets, and the recognition of the plurality of societal values assigned to these species. These tasks require a multi-disciplinary approach and the involvement of stakeholders. Here, we describe procedures to integrate multiple sources of information to formulate management priorities, targets, and high-level actions for the management of INNS. We follow five good-practice criteria: justified, evidence-informed, actionable, quantifiable, and flexible. We used expert knowledge methods to compile 17 lists of ecological, social, and economic impacts of lodgepole pines (Pinus contorta) and American mink (Neovison vison) in Chile and Argentina, the privet (Ligustrum lucidum) in Argentina, the yellow-jacket wasp (Vespula germanica) in Chile, and grasses (Urochloa brizantha and Urochloa decumbens) in Brazil. INNS plants caused a greater number of impacts than INNS animals, although more socio-economic impacts were listed for INNS animals than for plants. These impacts were ranked according to their magnitude and level of confidence on the information used for the ranking to prioritise impacts and assign them one of four high-level actions—do nothing, monitor, research, and immediate active management. We showed that it is possible to formulate management priorities, targets, and high-level actions for a variety of INNS and with variable levels of available information. This is vital in a world where the problems caused by INNS continue to increase, and there is a parallel growth in the implementation of management plans to deal with them.

Cerberus: Applying Supervised and Reinforcement Learning Techniques to Capture the Flag Games

Applying machine learning techniques to real-time computer games is an active research field. In this paper we present Cerberus, a machine learning framework for team-based Capture The Flag (CTF) games. This framework utilizes reinforcement learning to select high-level actions that achieve best team behaviour and utilizes neural networks to control fighting behaviour of team individuals. Our proposed framework also combines waypoints and influence maps for effective path planning

Learning and planning with logical automata

We introduce a method to learn policies from expert demonstrations that are interpretable and manipulable. We achieve interpretability by modeling the interactions between high-level actions as an automaton with connections to formal logic. We achieve manipulability by integrating this automaton into planning via Logical Value Iteration, so that changes to the automaton have predictable effects on the learned behavior. These qualities allow a human user to first understand what the model has learned, and then either correct the learned behavior or zero-shot generalize to new, similar tasks. Our inference method requires only low-level trajectories and a description of the environment in order to learn high-level rules. We achieve this by using a deep Bayesian nonparametric hierarchical model. We test our model on several domains of interest and also show results for a real-world implementation on a mobile robotic arm platform for lunchbox-packing and cabinet-opening tasks.

SportsCap: Monocular 3D Human Motion Capture and Fine-Grained Understanding in Challenging Sports Videos

Markerless motion capture and understanding of professional non-daily human movements is an important yet unsolved task, which suffers from complex motion patterns and severe self-occlusion, especially for the monocular setting. In this paper, we propose SportsCap—the first approach for simultaneously capturing 3D human motions and understanding fine-grained actions from monocular challenging sports video input. Our approach utilizes the semantic and temporally structured sub-motion prior in the embedding space for motion capture and understanding in a data-driven multi-task manner. To enable robust capture under complex motion patterns, we propose an effective motion embedding module to recover both the implicit motion embedding and explicit 3D motion details via a corresponding mapping function as well as a sub-motion classifier. Based on such hybrid motion information, we introduce a multi-stream spatial-temporal graph convolutional network to predict the fine-grained semantic action attributes, and adopt a semantic attribute mapping block to assemble various correlated action attributes into a high-level action label for the overall detailed understanding of the whole sequence, so as to enable various applications like action assessment or motion scoring. Comprehensive experiments on both public and our proposed datasets show that with a challenging monocular sports video input, our novel approach not only significantly improves the accuracy of 3D human motion capture, but also recovers accurate fine-grained semantic action attribute.

What Happened to Mirror Neurons?

Ten years ago, Perspectives in Psychological Science published the Mirror Neuron Forum, in which authors debated the role of mirror neurons in action understanding, speech, imitation, and autism and asked whether mirror neurons are acquired through visual-motor learning. Subsequent research on these themes has made significant advances, which should encourage further, more systematic research. For action understanding, multivoxel pattern analysis, patient studies, and brain stimulation suggest that mirror-neuron brain areas contribute to low-level processing of observed actions (e.g., distinguishing types of grip) but not to high-level action interpretation (e.g., inferring actors’ intentions). In the area of speech perception, although it remains unclear whether mirror neurons play a specific, causal role in speech perception, there is compelling evidence for the involvement of the motor system in the discrimination of speech in perceptually noisy conditions. For imitation, there is strong evidence from patient, brain-stimulation, and brain-imaging studies that mirror-neuron brain areas play a causal role in copying of body movement topography. In the area of autism, studies using behavioral and neurological measures have tried and failed to find evidence supporting the “broken-mirror theory” of autism. Furthermore, research on the origin of mirror neurons has confirmed the importance of domain-general visual-motor associative learning rather than canalized visual-motor learning, or motor learning alone.

Ontological Knowledge for Goal-Driven Autonomy Agents in Starcraft

Starcraft, a commercial Real-Time Strategy (RTS) game that has enjoyed world-wide popularity (including televised professional matches), is a challenging domain for automated computer agents. Evidence of this difficulty comes not only from characteristics of the game (massive state space, stochastic actions, partial visibility, etc.) but also from three years of competitive entries in tournaments (i.e. AIIDE Annual Starcraft Competition) in which the best automated entry performs poorly against a human expert. We are interested in taking a new research direction: using semantic knowledge, such as description logic, to represent the game state with abstract concepts in order to perform high level actions.

Replanning for Situated Robots

Planning enables intelligent agents, such as robots, to act so as to achieve their long term goals. To make the planning process tractable, a relatively low fidelity model of the world is often used, which sometimes leads to the need to replan. The typical view of replanning is that the robot is given the current state, the goal, and possibly some data from the previous planning process. However, for robots (or teams of robots) that exist in continuous physical space, act concurrently, have deadlines, or must otherwise consider durative actions, things are not so simple. In this paper, we address the problem of replanning for situated robots. Relying on previous work on situated temporal planning, we frame the replanning problem as a situated temporal planning problem, where currently executing actions are handled via Timed Initial Literals (TILs), under the assumption that actions cannot be interrupted. We then relax this assumption, and address situated replanning with interruptible actions. We bridge the gap between the low-level model of the robot and the high-level model used for planning by the novel notion of a bail out action generator, which relies on the low-level model to generate highlevel actions that describe possible ways to interrupt currently executing actions. Because actions can be interrupted at different times during their execution, we also propose a novel algorithm to handle temporal planning with time-dependent durations.

Multimodal (inter)action analysis of task instructions in language teaching via videoconferencing: A case study

AbstractOnline language teaching is gaining momentum worldwide and an expanding body of research analyses online pedagogical interactions. However, few studies have explored experienced online teachers’ practices in videoconferencing particularly while giving instructions, which are key to success in task-based language teaching (Markee, 2015). Adopting multimodal (inter)action analysis (Norris, 2004, 2019) to investigate the multimodal construction of instructions in a single case study, we examine instruction-giving as a social practice demonstrated in a specific site of engagement (a synchronous online lesson recorded for research purposes). Drawing on the higher-level actions (instruction-giving fragments) we have identified elsewhere (Satar & Wigham, 2020), in this paper we analyse the lower-level actions (modes) that comprise these higher-level actions, specifically focusing on the print mode (task resource sheets, URLs, text chat, and online collaborative writing spaces) wherein certain higher-level actions become frozen. Our findings are unique in depicting the modal complexity of sharing task resources in synchronous online teaching due to semiotic misalignment and semiotic lag that precludes the establishment of a completely shared interactional space. We observe gaze shifts as the sole indicator for learners that the teacher is multitasking between different higher-level actions. Further research is needed to fully understand the interactional features of online language teaching via videoconferencing to inform teacher training policy and practice.

IRoPro: An interactive Robot Programming Framework

The great diversity of end-user tasks ranging from manufacturing environments to personal homes makes pre-programming robots for general purpose applications extremely challenging. In fact, teaching robots new actions from scratch that can be reused for previously unseen tasks remains a difficult challenge and is generally left up to robotics experts. In this work, we present iRoPro, an interactive Robot Programming framework that allows end-users with little to no technical background to teach a robot new reusable actions. We combine Programming by Demonstration and Automated Planning techniques to allow the user to construct the robot's knowledge base by teaching new actions by kinesthetic demonstration. The actions are generalised and reused with a task planner to solve previously unseen problems defined by the user. We implement iRoPro as an end-to-end system on a Baxter Research Robot to simultaneously teach low- and high-level actions by demonstration that the user can customise via a Graphical User Interface to adapt to their specific use case. To evaluate the feasibility of our approach, we first conducted pre-design experiments to better understand the user's adoption of involved concepts and the proposed robot programming process. We compare results with post-design experiments, where we conducted a user study to validate the usability of our approach with real end-users. Overall, we showed that users with different programming levels and educational backgrounds can easily learn and use iRoPro and its robot programming process.

IPads and interaction: a materials perspective on collaborative discourse in secondary Spanish immersion

ABSTRACT As digital technologies proliferate throughout classrooms in the United States, iPads and other mobile tablets have been heralded as tools for enhanced peer collaboration. However, little research has examined exactly how students interact with and around iPads during collaborative learning tasks. This study employs multimodal interaction analysis to investigate how secondary Spanish immersion students leveraged the physicality of the iPads as multimodal affordances when completing a world history literacy-based task in a target language. In particular, I consider how the focal students employed touch-based communicative modes to complete higher-level actions, and the potential relationship between these modes and target language development. Findings have implications for materials use research in language teaching/learning classrooms as well as progressing our understanding of the impact of iPads on classroom discourse.

Action Sequencing Using Visual Permutations

Humans can easily reason about the sequence of high level actions needed to complete tasks, but it is particularly difficult to instill this ability in robots trained from relatively few examples. This work considers the task of neural action sequencing conditioned on a single reference visual state. This task is extremely challenging as it is not only subject to the significant combinatorial complexity that arises from large action sets, but also requires a model that can perform some form of symbol grounding, mapping high dimensional input data to actions, while reasoning about action relationships. This letter takes a permutation perspective and argues that action sequencing benefits from the ability to reason about both permutations and ordering concepts. Empirical analysis shows that neural models trained with latent permutations outperform standard neural architectures in constrained action sequencing tasks. Results also show that action sequencing using visual permutations is an effective mechanism to initialise and speed up traditional planning techniques and successfully scales to far greater action set sizes than models considered previously.

Disruptive Resilience: Harnessing Leadership to Build a More Equitable Health Care System After COVID-19.

Disruptive Resilience: Harnessing Leadership to Build a More Equitable Health Care System After COVID-19.

Analisis Postur Kerja Pada Proses Pemindahan Pasir Untuk Mengidentifikasi Resiko Kecelakaan Kerja

This study aims to know about the assessment of REBA posture method on sand removal work in the prevention of Musculoskletal Disorder. Workers who perform sand removal on UD XYZ are influenced by work posture, work position, and body performance. From the measurement results obtained the value of Rapid Entire Body Assessment REBA is 10, it can be seen the value of 9 is on the score of 8-10 which means that the work of the operator when moving the sand with a shovel has a high risk level and need action as soon as possible. the results obtained in carrying out the sand shoveling activity by repeatedly scoring the REBA score is 9 and it is stated that the activity has the potential to pose a risk of injury to the worker

MULTIMODAL COMMUNICATION ANALYSIS: THE HIGHER-LEVEL ACTIONS OF WAITERS AND WAITRESSES AT THREE RESTAURANTS IN TABANAN

This study aimed to analyze the multimodal communication of waiters and waitresses at three restaurants in Tabanan, Bali Province, in applying higher-level actions in serving foreign guests and investigating the functions of the higher-level actions of waiters and waitresses in serving the foreign guests. This study used theories proposed by Fishman (1972) cited in Suandi & Indriani (2016) on functions of using verbal and nonverbal communication, Krees and Van Leewuen 2006) on multimodal communication, and Norris (2004) on higher-level actions. The subjects of this study were four waiters and two waitresses who were working at three different restaurants in Tabanan regency. The results of this study showed that in each conversation, there were many sequences of higher-level actions used by waiters or waitresses to serve the foreign guests. The waiters and waitresses used a sequence of modes alternately or simultaneously in their higher-level actions, depending on the situation. The functions of waiters or waitresses’ higher-level actions in serving foreign guests made their messages clearer to foreign guests, and this made the waiters or waitresses’ messages more believable to the foreign guests.

Decomposition Analysis of the Evolution of the Local Energy System as a Tool to Assess the Effect of Local Actions: Methodology and Example of Malmö, Sweden

This paper proposes the use of decomposition analysis to assess the effect of local energy-related actions towards climate change mitigation, and thus improve policy evaluation and planning at the local level. The assessment of the impact of local actions has been a challenge, even from a strictly technical perspective. This happens because the total change observed is the result of multiple factors influencing local energy-related greenhouse gas (GHG) emissions, many of them not even influenced by local authorities. A methodology was developed, based on a recently developed decomposition model, that disaggregates the total observed changes in the local energy system into multiple causes/effects (including local socio-economic evolution, technology evolution, higher-level governance frame and local actions). The proposed methodology, including the quantification of the specific effect associated with local actions, is demonstrated with the case study of the municipality of Malmö (Sweden) in the timeframe between 1990 and 2015.

Community Conversations on Relationship Violence: Town Variations in Prevention Perceptions through Concept Mapping.

To further develop domestic and sexual violence (DSV) prevention strategies at the community level requires an understanding of how community members in towns think about prevention. Using concept mapping, this study sought to better understand community members' perspectives about what DSV prevention strategies would be most feasible, most effective, and generate the greatest community support within their town. Data were collected across four rural Northern England towns. Participants (>90% White) in each town (total sample size=119) brainstormed, sorted, and rated between 67 and 90 statements per town (x̅=75). Based on the results, a 5-cluster solution of school settings, conversations, individual direct action, community building, and community awareness was identified across all four towns as DSV prevention strategies with one town identifying an additional cluster solution of governance. Despite identifying similar prevention strategies, participants from each town rated these clusters of strategies differently on how feasible, effective, and supported they would be in their community. Overall, our results suggest that there were interesting consistencies across four towns in a similar geographic region in terms of how DSV prevention was described. However, individual communities differed in their views of the feasibility and acceptability of the different strategies. These results suggest that different strategies and higher-level actions may be required to address and prevent DSV within different towns and communities and that community narratives can clarify which specific strategies may encounter fewer barriers to implementation.