Internal Feedback Research Articles

Physical reservoir computing with FORCE learning in a living neuronal culture

Rich dynamics in a living neuronal system can be considered as a computational resource for physical reservoir computing (PRC). However, PRC that generates a coherent signal output from a spontaneously active neuronal system is still challenging. To overcome this difficulty, we here constructed a closed-loop experimental setup for PRC of a living neuronal culture, where neural activities were recorded with a microelectrode array and stimulated optically using caged compounds. The system was equipped with first-order reduced and controlled error learning to generate a coherent signal output from a living neuronal culture. Our embodiment experiments with a vehicle robot demonstrated that the coherent output served as a homeostasis-like property of the embodied system from which a maze-solving ability could be generated. Such a homeostatic property generated from the internal feedback loop in a system can play an important role in task solving in biological systems and enable the use of computational resources without any additional learning.

Moisture recycling and the potential role of forests as moisture source during European heatwaves

Heatwaves are extreme weather events that have become more frequent and intense in Europe over the past decades. Heatwaves are often coupled to droughts. The combination of them lead to severe ecological and socio-economic impacts. Heatwaves can self-amplify through internal climatic feedback that reduces local precipitation. Understanding the terrestrial sources of local precipitation during heatwaves might help identify mitigation strategies on land management and change that alleviate impacts. Moisture recycling of local water sources through evaporation allows a region to maintain precipitation in the same region or, by being transported by winds, in adjacent regions. To understand the role of terrestrial moisture sources for sustaining precipitation during heatwaves, we backtrack and analyse the precipitation sources of Northern, Western, and Southern sub-regions across Europe during 20 heatwave periods between 1979 and 2018 using the moisture tracking model Water Accounting Model-2layers (WAM-2layers). In Northern and Western Europe, we find that stabilizing anticyclonic patterns reduce the climatological westerly supply of moisture, mainly from the North Atlantic Ocean, and enhances the moisture flow from the eastern Euro-Asian continent and from within their own regions—suggesting over 10% shift of moisture supply from oceanic to terrestrial sources. In Southern Europe, limited local moisture sources result in a dramatic decrease in the local moisture recycling rate. Forests uniformly supply additional moisture to all regions during heatwaves and thus contribute to buffer local impacts. This study suggests that terrestrial moisture sources, especially forests, may potentially be important to mitigate moisture scarcity during heatwaves in Europe.

Reconfigurable artificial microswimmers with internal feedback

Self-propelling microparticles are often proposed as synthetic models for biological microswimmers, yet they lack the internally regulated adaptation of their biological counterparts. Conversely, adaptation can be encoded in larger-scale soft-robotic devices but remains elusive to transfer to the colloidal scale. Here, we create responsive microswimmers, powered by electro-hydrodynamic flows, which can adapt their motility via internal reconfiguration. Using sequential capillary assembly, we fabricate deterministic colloidal clusters comprising soft thermo-responsive microgels and light-absorbing particles. Light absorption induces preferential local heating and triggers the volume phase transition of the microgels, leading to an adaptation of the clusters’ motility, which is orthogonal to their propulsion scheme. We rationalize this response via the coupling between self-propulsion and variations of particle shape and dielectric properties upon heating. Harnessing such coupling allows for strategies to achieve local dynamical control with simple illumination patterns, revealing exciting opportunities for developing tactic active materials.

Investigating the relationships among error culture, error reporting, and performance: A panel study

Errors can have far-reaching consequences, especially in high-reliability contexts like healthcare. While it has been highlighted long ago, that organizations are able to learn from negative performance feedback that might have its root cause in human errors, the process how management becomes aware of errors has not yet been fully understood. Whereas research mostly assumes that employees share negative feedback with their superiors in a timely and accurate manner, upward flows of negative information are often delayed, sugarcoated, or even suppressed. Guided by the behavioral theory of the firm, we stress the importance of internal performance feedback and introduce organization-level error reporting as a key mechanism that triggers problemistic search activities of management. We incorporate temporal dynamics in current theorizing and provide new insights into the interplay of organizational error culture, error reporting, and organizational performance. Based on a large, longitudinal data set of public hospitals in England, our study successfully demonstrates that, within organizations, an increase in positive error culture increases organizations’ error reporting rate which in turn reduces the number of fatal patient safety incidents over time. Moreover, the investigation of lagged relationships offers a more nuanced understanding of how this within-organizational mediation process unfolds over time.

Combination of gated recurrent unit and Network in Network for underwater acoustic target recognition

Underwater acoustic target recognition is an important part of underwater acoustic signal processing and an important technical support for underwater acoustic information acquisition and underwater acoustic information confrontation. Taking into account that the gated recurrent unit (GRU) has an internal feedback mechanism that can reflect the temporal correlation of underwater acoustic target features, a model with gated recurrent unit and Network in Network (NIN) is proposed to recognize underwater acoustic targets in this paper. The proposed model introduces NIN to compress the hidden states of GRU while retaining the original timing characteristics of underwater acoustic target features. The higher recognition rate and faster calculation speed of the proposed model are demonstrated with experiments for raw underwater acoustic signals comparing with the multi-layer stacked GRU model.

Incentives for physician teams: Effectiveness of performance feedback and payment distribution methods

Best practices in team-based incentive design remain underexplored. This study examines under group-based pay-for-performance, how managers incentivize physicians for teamwork through internal feedback and payment distribution methods. In collaboration with Taiwan Association of Family Medicine, authors conducted a national survey of physician groups, with a response rate of 48.3%. Multilevel linear regression was applied to 134 groups, collectively consisting of 1,245 physicians in Taiwan. The outcome variables were two manager-rated scores for group performance on achieving (a) comprehensive, coordinated, continuous care, and (b) patient health improvement. The results indicate that providing each physician feedback on peer performance is superior to not providing it; when providing peer information within a group, concealing identities is superior to revealing them. These findings imply that application of the principle of social comparison can be effective; however, caution should be taken when disclosure of identifiable peer performance may intensify peer competition and undermine care coordination in team-based models. Further, groups that distribute payments equally among physicians perform better than groups that distributed payment proportionally to physicians’ patient shares. The findings are germane to small teams, where physicians do not have full control over care processes and outcomes, and need to work cooperatively to maximize group-based payment.

A System Theory Approach in Management of Community Health Centers (Mid-Level Health Organizations) in India

Health-care delivery in India has been developed at three levels specifically, primary, secondary, and tertiary. The secondary level of health-care basically includes Community Health Centers (CHCs), constituting the First Referral Units. The health-care system is that the organization (e.g. hospital, clinic, and nursing home) that has infrastructure and completely different complementary resources to support the work and development of care teams and microsystems. In open system theory, each system and its subsystems adapt to internal and external demands and feedback. Demands external to the health-care organization cover environmental factors (e.g., labor market, legislation, and population characteristics). Health-care organizations try to use a system that is functioning in response to feedback and informational signals to counteract lacuna. The application of system theory at mid-level will enhance the further managerial promptness at this level. In this paper, we described system theory as the process of identifying the systems of health-care delivery at the CHC level and subsystems with in CHCs and its interactions with environment and other systems.

Using electroencephalogram for emotion recognition based on filter-bank long short-term memory networks

Emotion plays an important role in people's cognition and communication. By analyzing electroencephalogram (EEG) signals to identify internal emotions and feedback emotional information in an active or passive way, affective brain-computer interactions can effectively promote human-computer interaction. This paper focuses on emotion recognition using EEG. We systematically evaluate the performance of state-of-the-art feature extraction and classification methods with a public-available dataset for emotion analysis using physiological signals (DEAP). The common random split method will lead to high correlation between training and testing samples. Thus, we use block-wise K fold cross validation. Moreover, we compare the accuracy of emotion recognition with different time window length. The experimental results indicate that 4 s time window is appropriate for sampling. Filter-bank long short-term memory networks (FBLSTM) using differential entropy features as input was proposed. The average accuracy of low and high in valance dimension, arousal dimension and combination of the four in valance-arousal plane is 78.8%, 78.4% and 70.3%, respectively. These results demonstrate the advantage of our emotion recognition model over the current studies in terms of classification accuracy. Our model might provide a novel method for emotion recognition in affective brain-computer interactions.

Skill acquisition practices of coaches on the British Para swimming World Class Programme

This study set out to gain insight into the practices adopted by elite level Paralympic swimming coaches and to shed light on the knowledge and rationale underpinning their approaches. Coaching sessions were delivered by nine senior coaches from the British Para swimming (BPS) World Class Programme. A coaching session was observed from each coach and a semi-structured interview was recorded and transcribed to explore their intentions and rationale for the structure and content of the session delivered. Results indicated that coaches: (i) predominantly emphasised internal focus instruction and feedback cues, (ii) incorporated relatively low levels of between-skill variability and higher levels of within-skill variability, and (iii) applied mostly explicit learning techniques such as part-task training and verbal feedback, but also incorporated some implicit learning techniques such as analogies and constraints-based learning. Interview data indicated coaches had limited knowledge of key skill acquisition principles. The study serves to highlight potential gaps in understanding on the side of both research and applied practice in the hope of facilitating future collaborations between coaches and skill acquisition practitioners.

Adaptive Control Structure with Neural Data Processing Applied for Electrical Drive with Elastic Shaft

This paper presents issues related to the adaptive control of the drive system with an elastic clutch connecting the main motor and the load machine. Firstly, the problems and the main algorithms often implemented for the mentioned object are analyzed. Then, the control concept based on the RNN (recurrent neural network) for the drive system with the flexible coupling is thoroughly described. For this purpose, an adaptive model inspired by the Elman model is selected, which is related to internal feedback in the neural network. The indicated feature improves the processing of dynamic signals. During the design process, for the selection of constant coefficients of the controller, the PSO (particle swarm optimizer) is applied. Moreover, in order to obtain better dynamic properties and improve work in real conditions, one model based on the ADALINE (adaptive linear neuron) is introduced into the structure. Details of the algorithm used for the weights’ adaptation are presented (including stability analysis) to perform the shaft torque signal filtering. The effectiveness of the proposed approach is examined through simulation and experimental studies.

A Recently Quenched Isolated Dwarf Galaxy Outside of the Local Group Environment

Abstract We report the serendipitous identification of a low-mass (M * ∼ 2 × 106 M ⊙), isolated, likely quenched dwarf galaxy in the “foreground” of the COSMOS-CANDELS field. From deep Hubble Space Telescope (HST) imaging we infer a surface brightness fluctuation distance for COSMOS-dw1 of D SBF = 22 ± 3 Mpc, which is consistent with its radial velocity of cz = 1222 ± 64 km s−1 via Keck/LRIS. At this distance, the galaxy is 1.4 Mpc in projection from its nearest massive neighbor. We do not detect significant H α emission (EW(H α)= − 0.4 ± 0.5 Å), suggesting that COSMOS-dw1 is likely quenched. Very little is currently known about isolated quenched galaxies in this mass regime. Such galaxies are thought to be rare, as there is no obvious mechanism to permanently stop star formation in them; to date there are only four examples of well-studied quenched field dwarfs, only two of which appear to have quenched in isolation. COSMOS-dw1 is the first example outside of the immediate vicinity of the Local Group. COSMOS-dw1 has a relatively weak Dn4000 break and the HST data show a clump of blue stars indicating that star formation ceased only recently. We speculate that COSMOS-dw1 was quenched due to internal feedback, which was able to temporarily suspend star formation. In this scenario the expectation is that quenched isolated galaxies with masses M * = 106–107 M ⊙ generally have luminosity-weighted ages ≲1 Gyr.

A Stochastic Optimal Control Model with Internal Feedback and Velocity Tracking for Saccadic Eye Movements

A new stochastic optimal control model with internal feedback and velocity tracking is presented in this paper to study saccadic eye movement control. Recent evidence from neurophysiological studies of superior colliculus suggests the presence of a dynamic input to the saccade generation system that encodes saccade velocity, rather than just the static information of the desired saccade amplitude and direction. The new evidence makes it imperative to test if the saccadic control system can use a desired velocity input to achieve accurate behavioral outcomes. To test such a velocity-based architecture of saccade control, a new optimal control tracking model that incorporated two unique characteristics of internal feedback and stochasticity has been developed. The proposed model was validated using behavioral data of saccades generated by healthy human subjects in an experimental setup. The model was able to generate displacement and velocity trajectories of horizontal saccades made to different amplitudes and predict saccades made to vertical and oblique directions. It captured the main sequence relationship between saccade amplitude and peak velocities that were observed in the behavioral data. This paper presents a novel optimal control framework with tracking and internal feedback and proposes the first-ever model of the saccadic system that uses an alternate interpretation of velocity-based control, contrary to the dominant end-point based models available in the literature.

Human Somatosensory Cortex Is Modulated during Motor Planning.

Recent data and motor control theory argues that movement planning involves preparing the neural state of primary motor cortex (M1) for forthcoming action execution. Theories related to internal models, feedback control, and predictive coding also emphasize the importance of sensory prediction (and processing) before (and during) the movement itself, explaining why motor-related deficits can arise from damage to primary somatosensory cortex (S1). Motivated by this work, here we examined whether motor planning, in addition to changing the neural state of M1, changes the neural state of S1, preparing it for the sensory feedback that arises during action. We tested this idea in two human functional MRI studies (N = 31, 16 females) involving delayed object manipulation tasks, focusing our analysis on premovement activity patterns in M1 and S1. We found that the motor effector to be used in the upcoming action could be decoded, well before movement, from neural activity in M1 in both studies. Critically, we found that this effector information was also present, well before movement, in S1. In particular, we found that the encoding of effector information in area 3b (S1 proper) was linked to the contralateral hand, similarly to that found in M1, whereas in areas 1 and 2 this encoding was present in both the contralateral and ipsilateral hemispheres. Together, these findings suggest that motor planning not only prepares the motor system for movement but also changes the neural state of the somatosensory system, presumably allowing it to anticipate the sensory information received during movement.SIGNIFICANCE STATEMENT Whereas recent work on motor cortex has emphasized the critical role of movement planning in preparing neural activity for movement generation, it has not investigated the extent to which planning also modulates the activity in the adjacent primary somatosensory cortex. This reflects a key gap in knowledge, given that recent motor control theories emphasize the importance of sensory feedback processing in effective movement generation. Here, we find through a convergence of experiments and analyses, that the planning of object manipulation tasks, in addition to modulating the activity in the motor cortex, changes the state of neural activity in different subfields of the human S1. We suggest that this modulation prepares the S1 for the sensory information it will receive during action execution.

Making internal feedback explicit: harnessing the comparisons students make during two-stage exams

In two-stage exams, students complete some academic work individually then immediately complete the same work again in small groups. Prior research shows that students perceive the dialogue they engage in at the group stage aids their individual learning. Outcome studies also suggest that engaging in dialogue benefits individual student learning. This investigation takes a different perspective. Learning during two-stage exams is conceptualised as deriving from internal feedback processes, with comparison seen as the mechanism that fuels these processes. To investigate this interpretation, students were asked to write a feedback commentary on their own learning during a two-stage exam. The intention was to make explicit the internal feedback that at would naturally occur during the group stage, as they students compare their earlier individual performance against the group dialogue and the unfolding group output. Analysis showed that students not only generated content and process feedback but also self-regulatory feedback, and that this feedback was of a higher quality than, and went beyond, the comments the teacher provided. This investigation yields new insights into the nature, scope and power of internal feedback processes, and into how to improve students’ learning during such two-stage sequences.

Making internal feedback explicit: exploiting the multiple comparisons that occur during peer review

This article explores peer review through the lens of internal feedback. It investigates the internal feedback that students generate when they compare their work with the work of peers and with comments received from peers. Inner feedback was made explicit by having students write an account of what they were learning from making these different comparisons. This allowed evaluation of the extent to which students’ self-generated feedback comments would match the feedback comments a teacher might provide, and exploration of other variables hypothesized to influence inner feedback generation. Analysis revealed that students’ self-generated feedback became more elaborate from one comparison to the next and that this, and multiple simultaneous comparisons, resulted in students’ generating feedback that not only matched the teacher’s feedback but surpassed it in powerful and productive ways. Comparisons against received peer comments added little to the feedback students had already generated from comparisons against peer works. The implications are that having students make explicit the internal feedback they generate not only helps them build their metacognitive knowledge and self-regulatory abilities but can also decrease teacher workload in providing comments.

Mapping farmer perceptions, Conservation Agriculture practices and on-farm measurements: The role of systems thinking in the process of adoption

CONTEXTConservation Agriculture (CA) usage, particularly in Southern Africa, has remained low with lower yield, higher weed pressure and lower soil quality cited as reasons for ‘disadoption’. OBJECTIVEUsing a detailed case study of 50 farmers in two villages in Cabo Delgado (Northern Mozambique), this study seeks to test the hypothesis that farmers' perceptions of CA are associated with distinctly different ‘mental models’ and if these "ways of thinking" overlap with farmers' identified/self-identified groupings (e.g. CA users, ‘disadopters’ and conventional tillage users). Secondly, we examine whether these different mental models (perceptions) are associated with actual differences in on-farm measurements. Finally, we explore the hypothesis that ‘systems thinking’ (i.e., understanding nonlinear causal relationships and internal feedback loops that drive a complex system) and CA usage are positively associated. METHODSFuzzy Cognitive Mapping (FCM) was used to elicit representations of farmers' mental models. To explore the association between farmers' mental models of CA/conventional practices and on-farm measurements we evaluated cowpea aboveground biomass, yield, weed cover, and soil quality parameters from the farmer's main plot. We drew on network analysis to measure structural metrics of cognitive maps that provide important information about a person's mental model (perceptions) of causal interdependencies of farming dynamics. RESULTS AND CONCLUSIONSWe find evidence of two data-driven distinct clusters of farmers' mental models that are in relative alignment with farmers' identified/self-identified groupings. Cluster 1 mainly consists of conventional users and cluster 2 mainly consists of CA users/disadopters. While no significant differences in socio-demographic variables were observed, clusters of mental models were associated with key differences in on-farm measurements. Importantly, cluster 1, who tended to be conventional users, had lower yields, lower soil cover, significantly lower carbon stock and higher weed coverage than cluster 2. Soil quality indicators were higher in cluster 2 as were farmers' overall revenue per hectare. Moreover, cluster 2 had significantly higher degrees of ‘systems thinking’ (measured through complex network analysis of graphical mental models) than cluster 1 which had higher forms of linear thinking. We argue that higher forms of experiential learning and practice of CA relate to higher degrees of systems thinking and stronger positive perceptions of CA, even among the CA ‘disadopters’. SIGNIFICANCEOur findings highlight the importance of systems thinking abilities and inclusion of biophysical, socio-economic and mental modelling variables rather than simple binary measurements which may have led to erroneous conclusions on CA and thus has implications for how CA is understood and promoted in future.

Application of a novel PIPDF controller in an improved plug-in electric vehicle integrated power system for AGC operation

Previously linearised plug-in electric vehicle (PEV) models are adopted in automatic generation control functioning. However, lesser attention is given towards the exact designing of a PEV unit contemplating the suitable model parameters. Therefore, this paper presents a novel approach for determining the PEV model parameters employing a Bode plot analysis in the primary frequency control operation. Furthermore, a novel proportional–integral internal proportional-derivative feedback controller is proposed for the secondary frequency control application asserting zero frequency and tie-line power deviations under steady-state condition. The proposed controller provides better dynamic stability compared to the existing controllers by adjusting open loop poles of the system. The moth flame optimisation technique is used to tune the gain parameters of the proposed controller. Simulation results, obtained from the case studies under different power scenarios, depict a significant improvement in the dynamic responses of frequency and tie-line power deviations in the presence of the proposed PEV model and optimised controller.

DFIG-Based Wind Energy System Robust Optimal Control by Using of Novel LMI-Based Adaptive MPC

The wind turbine system is one of the most challenging control problems in power systems due to its fast and slow dynamics and the optimal and robust control of the doubly fed induction generator is particularly important in this system. To handle the desired performance under different conditions, this paper presents a new control method by combining the model predictive control method with the adaptive method. In the aforementioned method by decoupling the active and reactive power control, an adaptive method is used to estimate the uncertainty, disturbance, and nonlinear terms as well as the internal feedback design. Then, linear matrix inequality-based robust model predictive control is designed as external feedback. The Lyapunov method is used to demonstrate system stability under the proposed controller. The simulation results show the efficiency of the method under different scenarios.

Entrepreneurial Dynamics and Clusters Formation. The Combinatory Systems View

I propose an interpretation of business dynamics in terms of the spatial co-localization of firms in a circumscribed area in order to form clusters of various types. I interpret clustering by adopting the methodology of combinatory systems: that is, systems formed by collectivities of agents that, by combining their micro behaviour, give rise to a macro behaviour and a macro effect that refers to the collectivity considered as a whole. Due to the presence of an internal feedback the macro behaviours direct or condition the subsequent micro behaviours, even though they derive from these. There are two business dynamics that lead to the formation of clusters: the exogenous dynamics, where the outside entrepreneurs locate their firms in a given area, and the endogenous dynamics, where there are new entrepreneurs generated from within a pre-existing cluster. The firm is considered as an intelligent cognitive system that evaluates its own fitness on the basis of a system of performance indicators; it estimates the effective or potential fitness for various possible areas of intervention and on the basis of the fitness levels assigns an index of attractiveness to the area. I argue that if an area has advantages in terms of fitness, then clusters will form there as a result of the co-localizations of firms that assign these areas a high attractiveness index. If the attractiveness landscapes appear flat, because no element stands out from the other areas to favour fitness, then if by chance an initial core of firms co-localize in the area, and their presence produces economic advantages in terms of economic efficiency or profitability with respect to other areas, these acquire the force of attraction and a combinatory system forms that by necessity increases the cluster. This process lasts as long as recombining factors maintain or increase the perceived advantages.

Optimal anticipatory control as a theory of motor preparation: A thalamo-cortical circuit model

SummaryAcross a range of motor and cognitive tasks, cortical activity can be accurately described by low-dimensional dynamics unfolding from specific initial conditions on every trial. These “preparatory states” largely determine the subsequent evolution of both neural activity and behavior, and their importance raises questions regarding how they are, or ought to be, set. Here, we formulate motor preparation as optimal anticipatory control of future movements and show that the solution requires a form of internal feedback control of cortical circuit dynamics. In contrast to a simple feedforward strategy, feedback control enables fast movement preparation by selectively controlling the cortical state in the small subspace that matters for the upcoming movement. Feedback but not feedforward control explains the orthogonality between preparatory and movement activity observed in reaching monkeys. We propose a circuit model in which optimal preparatory control is implemented as a thalamo-cortical loop gated by the basal ganglia.