Search

Predicting the consequences of the agent's actions on its environment is a pivotal challenge in robotic learning, which plays a key role in developing higher cognitive skills for intelligent robots. While current methods have predominantly relied on vision and motion data to generate the predicted videos, more comprehensive sensory perception is required for complex physical interactions such as contact-rich manipulation or highly dynamic tasks. In this work, we investigate the interdependence between vision and tactile sensation in the scenario of dynamic robotic interaction. A multi-modal fusion mechanism is introduced to the action-conditioned video prediction model to forecast future scenes, which enriches the single-modality prototype with a compressed latent representation of multiple sensory inputs. Additionally, to accomplish the interactive setting, we built a robotic interaction system that is equipped with both web cameras and vision-based tactile sensors to collect the dataset of vision-tactile sequences and the corresponding robot action data. Finally, through a series of qualitative and quantitative comparative study of different prediction architecture and tasks, we present insightful analysis of the cross-modality influence between vision, tactile and action, revealing the asymmetrical impact that exists between the sensations when contributing to interpreting the environment information. This opens possibilities for more adaptive and efficient robotic control in complex environments, with implications for dexterous manipulation and human-robot interaction. [ABSTRACT FROM AUTHOR]

Generalization enables individuals to respond to novel stimuli based on previous experiences. The degree to which organisms respond is determined by their physical resemblance to the original conditioned stimulus (CS+), with a stronger response elicited by more similar stimuli, resulting in similarity-based generalization gradients. Recent research showed that cognitive or conceptual dimensions also result in gradients similar to those observed with manipulations of physical dimensions. Such findings suggest that attributes beyond physical similarity play a role in shaping generalization gradients. However, despite its adaptive relevance for survival, there is no study exploring the effectiveness of affective dimensions in shaping generalization gradients. In two experiments (135 Spanish and 150 English participants, respectively), we used an online predictive learning task, in which different stimuli (words and Gabor patches) were paired with the presence – or absence – of a fictitious shock. After training, we assessed whether valence (i.e., hedonic experience) conveyed by words shape generalization gradients. In Experiment 1, the outcome expectancy decreased monotonically with variations in valence of Spanish words, mirroring the gradient obtained with the physical dimension (line orientation). In Experiment 2, conducted with English words, a similar gradient was observed when non-trained (i.e., generalization) words varied along the valence dimension, but not when words were of neutral valence. The consistency of these findings across two different languages strengthens the reliability and validity of the affective dimension as a determinant of generalization gradients. Furthermore, our data highlight the importance of considering the role of affective features in generalization responses, advancing the interplay between emotion, language, and learning. [ABSTRACT FROM AUTHOR]

Introduction: It is known that cognitive deficits are a core feature of schizophrenia and that in the general population, prior beliefs significantly influence learning and reasoning processes. However, the interaction of prior beliefs with cognitive deficits and their impact on performance in schizophrenia patients is still poorly understood. This study investigates the role of beliefs and cognitive variables (CVs) like working memory, associative learning, and processing speed on learning processes in individuals with schizophrenia. We hypothesize that beliefs will influence the ability to learn correct predictions and that first-episode schizophrenia patients (FEP) will show impaired learning due to cognitive deficits. Methods: We used a predictive-learning task to examine how FEP (n = 23) and matched controls (n = 23) adjusted their decisional criteria concerning physical properties during the learning process when predicting the sinking behavior of two transparent containers filled with aluminum discs when placed in water. Results: On accuracy, initial differences by group, trial type, and interaction effects of these variables disappeared when CVs were controlled. The differences by conditions, associated with differential beliefs about why the objects sink slower or faster, were seen in patients and controls, despite controlling the CVs' effect. Conclusions: Differences between groups were mainly explained by CVs, proving that they play an important role than what is assumed in this type of task. However, beliefs about physical events were not affected by CVs, and beliefs affect in the same way the decisional criteria of the control or FEP patients' groups.

Predicting the consequences of the agent’s actions on its environment is a pivotal challenge in robotic learning, which plays a key role in developing higher cognitive skills for intelligent robots. While current methods have predominantly relied on vision and motion data to generate the predicted videos, more comprehensive sensory perception is required for complex physical interactions such as contact-rich manipulation or highly dynamic tasks. In this work, we investigate the interdependence between vision and tactile sensation in the scenario of dynamic robotic interaction. A multi-modal fusion mechanism is introduced to the action-conditioned video prediction model to forecast future scenes, which enriches the single-modality prototype with a compressed latent representation of multiple sensory inputs. Additionally, to accomplish the interactive setting, we built a robotic interaction system that is equipped with both web cameras and vision-based tactile sensors to collect the dataset of vision-tactile sequences and the corresponding robot action data. Finally, through a series of qualitative and quantitative comparative study of different prediction architecture and tasks, we present insightful analysis of the cross-modality influence between vision, tactile and action, revealing the asymmetrical impact that exists between the sensations when contributing to interpreting the environment information. This opens possibilities for more adaptive and efficient robotic control in complex environments, with implications for dexterous manipulation and human-robot interaction.

Today, there is a visible reduction in productivity of common crops like rice and wheat in agricultural sector. At the same time, millet crops have emerged as an alternative to counter global food security issues because of their adaptability and nutritional value that can combat malnutrition. Also, because of its low moisture requirement and tolerance to extreme climatic conditions, it is perceived as a stable cereal. But its cultivation is affected by diseases like rust and blast, thereby causing harm to the farming economy. Conventional operational methods of disease detection require regular manual intervention which are also costly. The correctness of expert suggestions is also under scrutiny. Thus, a sustainable, robust and low cost modern approach for millet crop monitoring and disease detection is required. This research aims to develop a smart and sustainable framework by integrating Internet of things (IoT) and deep learning (DL). In the presented framework, a sensory module based automated crop health data gathering system with an improved deep learning-assisted intelligent disease recognition model is developed for the millet crop. Sensors collect data from millet farmland transfer the crop data readings to the cloud server for storage and Raspberry Pi for further detection. Any abnormality in reading leads to an alert notification communicated to the farmer. A hybrid predictive model, Customized Convolutional neural network (Customized-CNN) model works with the Raspberry Pi to predict the presence of blast and rust disease symptoms in millet. The proposed sustainable model is implemented, and it generates productive outcomes. The recorded accuracy, precision, recall, and f-score values with the customized CNN model were 98.8%, 98.2%, 97.4%, and 97.7%, respectively. The training and testing delays were only 67 s and 88 s, respectively. A sample of yearly sensor readings was generated to show the accuracy and reliability of the model's data collection. Also, the model proved to be scalable, as it gave noteworthy performance with diverse crops. The evaluation shows the reliability of the model, and it can be used by farming societies to enhance millet cereal yield in a cost-effective way. [ABSTRACT FROM AUTHOR]

Predictive learning is receiving growing interests with wide applications. Combining RNN and CNN, recent works attempt to capture temporal dependencies and spatial correlations simultaneously. However, these methods tend to be deeper networks without feature compression in scale and easily suffer from high computational costs, especially for high-resolution frames. To reduce resource burden, we introduce a novel and efficient network named CASTnet for spatiotemporal predictive learning. In this paper, we present a concise architecture cascading multiple stages, which can directly accept the raw frames without information loss in the first stage and compress features in the following stages leading to low computation relatively. Then, we adopt a special prediction head to aggregate multi-level features into predictions, which helps our model capture multi-scale targets. As for the spatiotemporal block, we adopt bidirectional convolutional attention (BCA) operations to capture the local and long-range information simultaneously without quadratic calculation complexity. In order to further improve the model performance while not causing much additional computation burden, we propose frame-wise knowledge distillation, which enables each frame at low-level to learn from each one at high-level. To evaluate our model, we conduct quantitative qualitative and ablation experiments on the MovingMNIST and Radar Echo datasets. The results show that our CASTnet achieves competitive results with lower computational costs and fewer network parameters than the state-of-the-art models. [ABSTRACT FROM AUTHOR]

Knowledge of predictive relations is a core aspect of learning. Beyond individual relations, we also represent intuitive theories of the world, which include interrelated sets of relations. We asked whether individual predictive relations learned incidentally in the same context become associatively bound and whether they spontaneously influence later learning. Participants performed a cover task while watching three sequences of events. Each sequence contained the same set of events, but differed in how the events related to each other. The first two sequences each had two strong predictive relations (R1 & R2, and R3 & R4). The third contained either a consistent pairing of relations (R1 & R2) or an inconsistent pairing (R1 & R3). We found that participants learning of the individual relations in the third sequence was affected by pairing consistency, suggesting the mind associates relations to each other as part of the intrinsic way it learns about the world. This was despite participants minimal ability to verbally describe most of the relations they had learned. Thus, participants spontaneously developed the expectation that pairs of relations should cohere, and this affected their ability to learn new evidence. Such associative binding of relational information may help us build intuitive theories.

When learning about events through visual experience, one must not only identify which events are visually similar but also retrieve those events associates-which may be visually dissimilar-and recognize when different events have similar predictive relations. How are these demands balanced? To address this question, we taught participants the predictive structures among four events, which appeared in four different sequences, each cued by a distinct object. In each, one event (cause) was predictably followed by another (effect). Sequences in the same relational category had similar predictive structure, while across categories, the effect and cause events were reversed. Using functional magnetic resonance imaging data, we measured associative coding, indicated by correlated responses between effect and cause events; perceptual coding, indicated by correlated responses to visually similar events; and relational category coding, indicated by correlated responses to sequences in the same relational category. All three models characterized responses within the right middle temporal gyrus (MTG), but in different ways: Perceptual and associative coding diverged along the posterior to anterior axis, while relational categories emerged anteriorly in tandem with associative coding. Thus, along the posterior-anterior axis of MTG, the representation of the visual attributes of events is transformed to a representation of both specific and generalizable relational attributes.

This paper proposes a new planning model to coordinate the expansion of electric vehicle charging infrastructure (EVCI) and renewables in power grids. Firstly, individual electric vehicle (EV) charging behaviours are modeled considering EV customers adopting smart charging services as the main charging method and those using fast charging, super fast charging and battery swapping services as a complementary charging approach. Next, EV aggregation and the associated system economic dispatch model are built. A novel model predictive control (MPC) learning approach is then proposed to iteratively learn the correlation between different types of EV charging loads and the EV interactions with renewables and other generating units in modern power grids of the future. The simulation results demonstrate that the proposed approach can be used to quantify the ratio of different types of charging loads in a region and strategically guide on the integration of EVs and renewables to achieve the clean energy transition goals. The proposed framework can also be used to decide charging capacity needs in a charging demand zone.

Human analogical ability involves the re-use of abstract, struc-tured representations within and across domains. Here, wepresent a generative neural network that completes analogiesin a 1D metric space, without explicit training on analogy.Our model integrates two key ideas. First, it operates overrepresentations inspired by properties of the mammalian En-torhinal Cortex (EC), believed to extract low-dimensional rep-resentations of the environment from the transition probabil-ities between states. Second, we show that a neural networkequipped with a simple predictive objective and highly generalinductive bias can learn to utilize these EC-like codes to com-pute explicit, abstract relations between pairs of objects. Theproposed inductive bias favors a latent code that consists ofanti-correlated representations. The relational representationslearned by the model can then be used to complete analogiesinvolving the signed distance between novel input pairs (1:3:: 5:? (7)), and extrapolate outside of the network’s trainingdomain. As a proof of principle, we extend the same architec-ture to more richly structured tree representations. We suggestthat this combination of predictive, error-driven learning andsimple inductive biases offers promise for deriving and utiliz-ing the representations necessary for high-level cognitive func-tions, such as analogy.

Predictive graph learning approaches have been bringing significant advantages in many real-life applications, such as social networks, recommender systems, and other social-related downstream tasks. For those applications, learning models should be able to produce a great prediction result to maximize the usability of their application. However, the paradigm of current graph learning methods generally neglects the differences in link strength, leading to discriminative predictive results, resulting in different performance between tasks. Based on that problem, a fairness-aware predictive learning model is needed to balance the link strength differences and not only consider how to formulate it. To address this problem, we first formally define two biases (i.e., Preference and Favoritism) that widely exist in previous representation learning models. Then, we employ modularity maximization to distinguish strong and weak links from the quantitative perspective. Eventually, we propose a novel predictive learning framework entitled ACE that first implements the link strength differentiated learning process and then integrates it with a dual propagation process. The effectiveness and fairness of our proposed ACE have been verified on four real-world social networks. Compared to nine different state-of-the-art methods, ACE and its variants show better performance. The ACE framework can better reconstruct networks, thus also providing a high possibility of resolving misinformation in graph-structured data. [ABSTRACT FROM AUTHOR]

General value functions (GVFs) in the reinforcement learning (RL) literature are long-term predictive summaries of the outcomes of agents following specific policies in the environment. Affordances as perceived action possibilities with specific valence may be cast into predicted policy-relative goodness and modeled as GVFs. A systematic explication of this connection shows that GVFs and especially their deep-learning embodiments (1) realize affordance prediction as a form of direct perception, (2) illuminate the fundamental connection between action and perception in affordance, and (3) offer a scalable way to learn affordances using RL methods. Through an extensive review of existing literature on GVF applications and representative affordance research in robotics, we demonstrate that GVFs provide the right framework for learning affordances in real-world applications. In addition, we highlight a few new avenues of research opened up by the perspective of "affordance as GVF," including using GVFs for orchestrating complex behaviors. [ABSTRACT FROM AUTHOR]

An experiment evaluated whether a stimulus associated with extinction can attenuate the reinstatement of a previously extinguished predictive learning relationship in humans. Participants learned a specific relationship between two cues (X and Y) and two outcomes (O1 and O2) during the first phase. Throughout extinction, both cues were presented without outcomes. Then, testing was conducted after exposure to the original outcomes. We found a reduction of the reinstatement effect when participants received a cue associated with extinction, but not when testing involved a novel cue. This result indicates that the reductive effect depends on the cue's specific association with extinction. The findings are consistent with the theoretical view that explains reinstatement as a failure to retrieve the extinction learning. [ABSTRACT FROM AUTHOR]

Identifying responses/sensitivity of drugs to explore personalized therapy related to drugs has been a significant issue in pharmacogenomic-related studies. Cancer is an acute disease involving heterogeneous behavior of similar tumor types of human's data related to drug therapy. Some large-scale pharmacogenomic studies related to cancer cell lines have been emerging pharma clinical systems to predicting advanced genetic sub relations like sensitivity for influence therapeutic approach. Advanced machine learning approaches related to pharmacogenomic data studies have evaluated the ability to identify drug/compound sensitivity. Recently, drug sensitivity prediction has been critical in drug identification and design because some of the pharmacogenomic online data contribute as open access. These databases are used to identify drug sensitivity by developing computational methodologies. However, all these methodologies are not sufficient to predict the sensitivity of drugs because of multi-features. So that in this paper, we propose a Novel Regression-based Ensemble Convolution Neural Network Model (NRECNNM) to identify the sensitivity of drugs based on multiple pharma omics data & addresses heterogeneity in the selection of features to sub pharmacogenomic parameters. This approach also evaluates the number of components with the regression calculation method. Ensemble convolution network describes the significance of distance metrics relates to neighborhood dependencies related to drug therapy relations. The proposed approach uses Cancer cell line Encyclopedia (CCLE) data sets, and experimental results examine the performance in predicting sensitivity can enhance with conventional approaches. [ABSTRACT FROM AUTHOR]

Group work can inspire students, encourage constructive learning, and improve essential critical thinking, communication, and decision-making in the present competitive world. The risk factors in group learning include students who prefer working alone and strongly despise dealing with things created by gathering in teams. Online learners often have problems locating lasting peace times for group therapy sessions are considered an essential factor. A predictive Group Learning Behavior Approach (PGLBA) has been proposed. Students who commute to college agree to group conferences and workgroup learning and Inquiry learning for higher education. The grounded Inquiry Learning Approach is invented to strengthen students' enjoyment of active group learning, and the students find times for group meetings that are often mutually advantageous. The simulation analysis is performed based on performance, accuracy, and efficiency proves the proposed framework's reliability. The experimental results show that the proposed PGLBA-IL model enhances the accuracy ratio of 81.2%, an efficiency ratio of the number of students 86.4%, and the overall performance analysis ratio of 85.1% compared to others existing approaches. [ABSTRACT FROM AUTHOR]

Predicting spatial behaviors of an individual (e.g., frequent visits to specific locations) is important to improve our understanding of the complexity of human mobility patterns, and to capture anomalous behaviors in an individual’s spatial movements, which can be particularly useful in situations such as those induced by the COVID-19 pandemic. We propose a system called Deep Spatio-Temporal Predictor (DST-Predict), that can predict the future visit frequency of an individual based on one’s past mobility behaviour patterns using GPS trace data collected from mobile phones. Predicting such spatial behavior is challenging, primarily because individuals’ patterns of location visits for each individual consists of both systematic and random components, which vary across the spatial and temporal scales of analysis. To address these issues, we propose a novel multi-view sequence-to-sequence model that uses Convolutional Long-short term memory (ConvLSTM) where the past history of frequent visit patterns features is used to predict individuals’ future visit patterns in a multi-step manner. Using the GPS survey data obtained from 1,464 participants in western New York, US, we demonstrated that the proposed system is capable of predicting individuals’ frequency of visits to common places in an urban setting, with high accuracy.

Humans create internal models of an environment (i.e. cognitive maps) through subjective sensorimotor experiences and can also understand spatial locations by looking at an external map as a symbol of an environment. We simulate the development of the cognitive map from sensorimotor experiences and grounding of the external map in a single deep neural network model. Our proposed network has a shared module that processes the features of multiple modalities (i.e. vision, hearing, and touch) and even external maps in the same manner. The multiple modalities are encoded into feature vectors by modality-specific encoders, and the encoded features are processed by the same shared module. The proposed network was trained to predict the sensory inputs of a simulated mobile robot. After the predictive learning, the spatial representation was developed in the internal states of the shared module, and the same spatial representation was used for predicting multiple modalities, including the external map. The network can also perform spatial navigation by associating the external map with the cognitive map. This implies that the external maps are grounded in subjective sensorimotor experiences, being bridged through the developed internal spatial representation in the shared module. [ABSTRACT FROM AUTHOR]

Human emotions affect psychological health to a great level. Positive emotions relate to health improvement; whereas negative emotions may aggravate psychological disorders such as anxiety, stress, and depression. Although there exist several computational methods to predict psychological disorders, most of them provide a black-box view of uncertainty. This research involves developing a novel predictive model for multi class psychological risk recognition with an accurate explainable interface. Standard questionnaires are utilized as data set and a new approach called a Q-Prioritization is employed to drop insignificant questions from the data set. Moreover, a novel balanced decision tree method based on repetitive oversampling is applied for the training and testing of the model. Predictive nature along with its contributing factors are interpreted with three techniques such as permuted feature importance, contrastive explanation, and counterfactual method, which together form a reasoning engine. The prediction outcome generated an impressive performance with an aggregated accuracy of 98.25%. The mean precision, recall, and F-score metric recorded were 0.98, 0.977, and 0.979, respectively. Also, it was noted that without applying Q-Prioritization, the accuracy significantly drops to 90.25%. The error rate observed with our model was only 0.026. The proposed multi-level psychological disorder predictive model can successfully serve as an assistive deployment for medical experts in the effective treatment of mental health.

Outcome predictability effects in associative learning paradigms describe better learning about outcomes with a history of greater predictability in a similar but unrelated task compared with outcomes with a history of unpredictability. Inspired by the similarities between this phenomenon and the effect of uncontrollability in learned helplessness paradigms, here, we investigate whether learning about unpredictability decreases outcome-specific motivation to learn. We used a modified version of the allergy task, in which participants first observe the foods eaten by a fictitious patient, followed by allergic reactions that he subsequently suffers, some of which are perfectly predictable and others unpredictable. We then implemented an active learning method in a second task in which participants could only learn about either the previously predictable or unpredictable outcomes on each trial. At the beginning of each trial, participants had to decide whether they wanted to learn about one outcome category or the other. Participants at the beginning of the second task chose to learn about the previously predictable outcomes first and to learn about the previously unpredictable outcomes in later trials. This showed that unpredictability affects future motivation to learn in other circumstances. Interestingly, we did not find any sign of outcome predictability effect at the end of the second phase, suggesting that participants compensate for biased outcome sampling when making overt choices in ways that they may not when learning about both outcome categories simultaneously. [ABSTRACT FROM AUTHOR]

Deep robotic learning by learning from demonstration allows robots to mimic a given demonstration and generalize their performance to unknown task setups. However, this generalization ability is heavily affected by the number of demonstrations, which can be costly to manually generate. Without sufficient demonstrations, robots tend to overfit to the available demonstrations and lose the robustness offered by deep learning. Applying the concept of motor babbling – a process similar to that by which human infants move their bodies randomly to obtain proprioception – is also effective for allowing robots to enhance their generalization ability. Furthermore, the generation of babbling data is simpler than task-oriented demonstrations. Previous researches use motor babbling in the concept of pre-training and fine-tuning but have the problem of the babbling data being overwritten by the task data. In this work, we propose an RNN-based robot-control framework capable of leveraging targetless babbling data to aid the robot in acquiring proprioception and increasing the generalization ability of the learned task data by learning both babbling and task data simultaneously. Through simultaneous learning, our framework can use the dynamics obtained from babbling data to learn the target task efficiently. In the experiment, we prepare demonstrations of a block-picking task and aimless-babbling data. With our framework, the robot can learn tasks faster and show greater generalization ability when blocks are at unknown positions or move during execution. [ABSTRACT FROM AUTHOR]

In this paper, traffic and vehicle density optimization issues are discussed, and a suitable solution for autonomous driving environments is proposed. The problem of scalable traffic flow management is hindered by the unpredictable vehicle density and its navigation. This problem increases the flow congestion at the intersection during the change of lane in particular. This paper introduces density-scaling traffic management for autonomous vehicle driving scenarios. The proposed technique relies on navigation output and the autonomous vehicles' trajectory decisions for predicting their availability and time duration. The traffic flow is streamlines based on the predictions; the prediction error is mitigated by rescheduling the successive traffic flow. In this process, predictive iterated learning is employed for improving the scalable traffic flows in an intersection. The predictive learning handles navigation assistance output change of trajectory independently and correlates the decisions' match. The process of prediction matching is performed for improving the flow management with the time factor. The traffic management technique is useful in providing latency-less travel and errorless decision and flow complexity and collision probability. [ABSTRACT FROM AUTHOR]

Recent research has identified key differences in the wayindividuals make decisions in predictive learning tasks,including the use of feature- and rule-based strategies incausal learning and model-based versus model-free choices inreinforcement learning. These results suggest that people relyto varying degrees on separable psychological processes.However, the relationship between these types of learningstrategies has not been explored in any depth. This studyinvestigated the relationship between feature- vs rule-basedstrategies in a causal learning task and indices of model-freeand model-based choice in a two-step reinforcement learningprocedure. We found that rule-based transfer was associatedwith the use of model-based, but not model-free responding ina two-step task.

• Experiments demonstrate correlations between long-term stress and renewal. • Long term stress had no relation to conditioning or extinction. • Perceived control of stressful events was related to greater renewal. • Effects of long-term stress and acute stress prior to test are similar. • The PSS-10 as a measure of stress may predict susceptibility to relapse. Two online experiments evaluated the relationship between long-term stress, as measured with the Perceived Stress Scale-10, and the Renewal Effect. In the first experiment renewal was assessed with a behavioral suppression task in a science-fiction based video game. Participants learned to suppress mouse clicking during a signal for an upcoming attack to avoid losing points. The signal was first paired with an attack in Context A and extinguished in Context B and tested back in Context A. The contexts were different space galaxies where the gameplay took place. Experiment 2 used a food/illness predictive-learning paradigm. Two food items were paired with stomachache in one restaurant (A) and extinguished in Context B prior to testing in both contexts without feedback. Positive correlations were obtained between renewal and stress in each experiment. Unlike acute stress (Drexler et al., 2017), long term stress was associated with greater renewal. The effects of stress, both chronic and punctual, on renewal are discussed. [ABSTRACT FROM AUTHOR]

In contrast to the large body of work demonstrating second-order conditioning (SOC) in non-human animals, the evidence for SOC in humans is scant. In this review, I examine the existing literature and suggest theoretical and procedural explanations for why SOC has been so elusive in humans. In particular, I discuss potential interactions with conditioned inhibition, whether SOC is rational, and propose critical parameters needed to obtain the effect. I conclude that SOC is a real but difficult phenomenon to obtain in humans, and suggest directions for future research.

In contrast to the large body of work demonstrating second-order conditioning (SOC) in non-human animals, the evidence for SOC in humans is scant. In this review, I examine the existing literature and suggest theoretical and procedural explanations for why SOC has been so elusive in humans. In particular, I discuss potential interactions with conditioned inhibition, whether SOC is rational, and propose critical parameters needed to obtain the effect. I conclude that SOC is a real but difficult phenomenon to obtain in humans, and suggest directions for future research. [ABSTRACT FROM AUTHOR]

Predictive learning endows stimuli with the capacity to signal both the sensory‐specific and general motivational properties of their associated rewards or outcomes. These two signals can be distinguished behaviorally by their influence on the selection and performance of instrumental actions, respectively. This review focuses on how sensory‐specific predictive learning guides choice between actions that earn otherwise equally desirable outcomes. We describe evidence that outcome‐specific predictive learning is encoded in the basolateral amygdala and drives the accumulation of delta‐opioid receptors on the surface of cholinergic interneurons located in the nucleus accumbens shell. This accumulation constitutes a novel form of cellular memory, not for outcome‐specific predictive learning per se but for the selection of, and choice between, future instrumental actions. We describe recent evidence regarding the cascade of events necessary for the formation and expression of this cellular memory and point to open questions for future research into this process. Beyond these mechanistic considerations, the discovery of this new form of memory is consistent with recent evidence suggesting that intracellular rather than synaptic changes can mediate learning‐related plasticity to modify brain circuitry to prepare for future significant events. [ABSTRACT FROM AUTHOR]

Predictive graph learning approaches have been bringing significant advantages in many real-life applications, such as social networks, recommender systems, and other social-related downstream tasks. For those applications, learning models should be able to produce a great prediction result to maximize the usability of their application. However, the paradigm of current graph learning methods generally neglects the differences in link strength, leading to discriminative predictive results, resulting in different performance between tasks. Based on that problem, a fairness-aware predictive learning model is needed to balance the link strength differences and not only consider how to formulate it. To address this problem, we first formally define two biases (i.e., Preference and Favoritism) that widely exist in previous representation learning models. Then, we employ modularity maximization to distinguish strong and weak links from the quantitative perspective. Eventually, we propose a novel predictive learning framework entitled ACE that first implements the link strength differentiated learning process and then integrates it with a dual propagation process. The effectiveness and fairness of our proposed ACE have been verified on four real-world social networks. Compared to nine different state-of-the-art methods, ACE and its variants show better performance. The ACE framework can better reconstruct networks, thus also providing a high possibility of resolving misinformation in graph-structured data.

Aims: To examine two explanations for the observation that cue–exposure treatment has not been clearly effective in the treatment of alcohol dependence: do alcohol‐dependent individuals have either (1) slower extinction and/or (2) greater contextual specificity of extinction than non‐dependent individuals? Design In two exploratory laboratory experiments we used mixed factorial designs with two‐group between‐subjects factors and within‐subjects factors corresponding to performance in different parts of a computer‐based learning task. Setting: University of Southampton psychology research laboratories and two addiction treatment services in the city of Southampton, UK. Participants: Experiment 1: 74 (54 female) undergraduates from the University of Southampton (age mean = 20.4 years). Experiment 2: 102 (40 female) participants from the University of Southampton, the local community, and from two Southampton alcohol treatment services (age mean = 41.3 years). Measurements: The Alcohol Use Disorders Identification Test, a 1‐week time‐line follow‐back alcohol consumption questionnaire, the Barratt Impulsiveness Scale (11th edn), and a computerized learning task. Experiment 2 additionally used the 44‐item Big Five Inventory, a drug use history checklist, and the Hospital Anxiety and Depression Scale. Findings Experiment 1: light and heavy drinkers did not differ significantly in extinction [extinction block × drinking status interaction, P = 0.761, ηp2=0.005, 95% confidence interval (CI) = (0,0.028)] or on contextual control of extinction [recovery block × drinking status interaction, P = 0.514, ηp2=0.009, 95% CI =(0, 0.084)]. Experiment 2: slower extinction in abstinent alcohol‐dependent participants compared with light drinkers [extinction block × drinking status interaction, P = 0.023, ηp2=0.031, 95% CI = 0, 0.069)] but no significant difference on contextual control of extinction [recovery block × drinking status interaction, P = 0.069, ηp2=0.033, 95% CI = (0, 0.125)]. Conclusion: Abstinent alcohol‐dependent people may have slower extinction learning for alcohol‐related cues than non‐dependent light drinkers. [ABSTRACT FROM AUTHOR]

Introduction: It is known that cognitive deficits are a core feature of schizophrenia and that in the general population, prior beliefs significantly influence learning and reasoning processes. However, the interaction of prior beliefs with cognitive deficits and their impact on performance in schizophrenia patients is still poorly understood. This study investigates the role of beliefs and cognitive variables (CVs) like working memory, associative learning, and processing speed on learning processes in individuals with schizophrenia. We hypothesize that beliefs will influence the ability to learn correct predictions and that first-episode schizophrenia patients (FEP) will show impaired learning due to cognitive deficits., Methods: We used a predictive-learning task to examine how FEP ( n  = 23) and matched controls (n = 23) adjusted their decisional criteria concerning physical properties during the learning process when predicting the sinking behavior of two transparent containers filled with aluminum discs when placed in water., Results: On accuracy, initial differences by group, trial type, and interaction effects of these variables disappeared when CVs were controlled. The differences by conditions, associated with differential beliefs about why the objects sink slower or faster, were seen in patients and controls, despite controlling the CVs' effect., Conclusions: Differences between groups were mainly explained by CVs, proving that they play an important role than what is assumed in this type of task. However, beliefs about physical events were not affected by CVs, and beliefs affect in the same way the decisional criteria of the control or FEP patients' groups., Competing Interests: The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Ramon Castillo reports financial support was provided by University of Talca Faculty of Psychology. Ramon D. Castillo reports a relationship with National Agency for Research and Development that includes: funding grants. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2024 The Author(s).)

Background Learning to predict threatening events enables an organism to engage in protective behavior and prevent harm. Failure to differentiate between cues that truly predict danger and those that do not, however, may lead to indiscriminate fear and avoidance behaviors, which in turn may contribute to disability in people with persistent pain. We aimed to test whether people with persistent neck pain exhibit contingency learning deficits in predicting pain relative to pain-free, gender-and age-matched controls. Method We developed a differential predictive learning task with a neck pain-relevant scenario. During the acquisition phase, images displaying two distinct neck positions were presented and participants were asked to predict whether these neck positions would lead to pain in a fictive patient with persistent neck pain (see fictive patient scenario details in Appendix A). After participants gave their pain-expectancy judgment in the hypothetical scenario, the verbal outcome (PAIN or NO PAIN) was shown on the screen. One image (CS+) was followed by the outcome “PAIN”, while another image (CS−) was followed by the outcome “NO PAIN”. During the generalization phase, novel but related images depicting neck positions along a continuum between the CS+ and CS− images (generalization stimuli; GSs) were introduced to assess the generalization of acquired predictive learning to the novel images; the GSs were always followed by the verbal outcome “NOTES UNREADABLE” to prevent extinction learning. Finally, an extinction phase was included in which all images were followed by “NO PAIN” assessing the persistence of pain-expectancy judgments following disconfirming information. Results Differential pain-expectancy learning was reduced in people with neck pain relative to controls, resulting from patients giving significantly lower pain-expectancy judgments for the CS+, and significantly higher pain-expectancy judgments for the CS−. People with neck pain also demonstrated flatter generalization gradients relative to controls. No differences in extinction were noted. Discussion The results support the hypothesis that people with persistent neck pain exhibit reduced differential pain-expectancy learning and flatter generalization gradients, reflecting deficits in predictive learning. Contrary to our hypothesis, no differences in extinction were found. These findings may be relevant to understanding behavioral aspects of chronic pain.

We provide high-probability sample complexity guarantees for exact structure recovery and accurate predictive learning using noise-corrupted samples from an acyclic (tree-shaped) graphical model. The hidden variables follow a tree-structured Ising model distribution, whereas the observable variables are generated by a binary symmetric channel taking the hidden variables as its input (flipping each bit independently with some constant probability q 2 [0; 1=2)). In the absence of noise, predictive learning on Ising models was recently studied by Bresler and Karzand (2020); this paper quantifies how noise in the hidden model impacts the tasks of structure recovery and marginal distribution estimation by proving upper and lower bounds on the sample complexity. Our results generalize state-of-the-art bounds reported in prior work, and they exactly recover the noiseless case (q = 0). In fact, for any tree with p vertices and probability of incorrect recovery ffi > 0, the sufficient number of samples remains logarithmic as in the noiseless case, i.e., O(log(p=)), while the dependence on q is O 1=(1 2q)4, for both aforementioned tasks. We also present a new equivalent of Isserlis' Theorem for sign-valued tree-structured distributions, yielding a new low-complexity algorithm for higher-order moment estimation. [ABSTRACT FROM AUTHOR]

Events—the experiences we think we are having and recall having had—are constructed; they are not what actually occurs. What occurs is ongoing dynamic, multidimensional, sensory flow, which is somehow transformed via psychological processes into structured, describable, memorable units of experience. But what is the nature of the redescription processes that fluently render dynamic sensory streams as event representations? How do such processes cope with the ubiquitous novelty and variability that characterize sensory experience? How are event‐rendering skills acquired and how do event representations change with development? This review considers emerging answers to these questions, beginning with evidence that an implicit tendency to monitor predictability structure via statistical learning is key to event rendering. That is, one way that the experience of bounded events (e.g., actions within behavior, words within speech) arises is with the detection of "troughs" in sensory predictability. Interestingly, such troughs in predictability are often predictable; these regions of predictable‐unpredictability provide articulation points to demarcate one event from another in representations derived from the actual streaming information. In our information‐optimization account, a fluent event‐processor predicts such troughs and selectively attends to them—while suppressing attention to other regions—as sensory streams unfold. In this way, usage of attentional resources is optimized for efficient sampling of the most relevant, information‐rich portions of the unfolding flow of sensation. Such findings point to the development of event‐processing fluency—whether in action, language, or other domains—depending crucially on rapid and continual cognitive reorganization. As knowledge of predictability grows, attention is adaptively redeployed. Accordingly, event experiences undergo continuous alteration. Baldwin and Kosie study the processes that 'construct' events from raw sensory experience, focusing on statistical methods for identifying event boundaries. They extend the observation that event boundaries often coincide with troughs in predictability of sensory inputs, by noting that these troughs are often themselves predictable, and develop an account of the 'predictability structure' of events. [ABSTRACT FROM AUTHOR]

A flexible and dynamically adjustable behavior is crucial to adapt to a continuously changing environment. In order to optimally adapt, we need to learn from the consequences of our behavior. We usually learn through different kinds of prediction errors, which occur when we experience unexpected situations due to false predictions. With this literature review, we intended to contribute to current etiological models that ascribe various positive symptoms (particularly delusions and hallucinations) in patients with schizophrenia to false prediction errors and deficient predictive learning. We discuss alterations in the electrophysiological measure of the error‐related negativity/error negativity (ERN/Ne) as a global deficit and a trait in schizophrenia, as they have been observed in different samples of patients with schizophrenia, in individuals at high‐risk and individuals with subclinical schizotypal traits. As the ERN/Ne can itself be considered the result of predictive processes (evaluation of current action outcomes as worse than expected), we propose that the reported alterations indicate that patients suffering from schizophrenic illnesses fail to adequately classify the outcomes of their actions as better or worse than expected due to a deficit in self‐monitoring. Furthermore, we discuss results in further action‐monitoring components, such as the correct response negativity (CRN)—a smaller negativity elicited by correct responses; and error positivity (Pe)—a later positivity assumed to reflect conscious error processing. The reported results show normal Pe amplitudes and normal post‐error adjustments (adaptations after committed error to improve performance), indicating an intact later and conscious processing. From the results of diminished differences between ERN/Ne and CRN amplitudes, we conclude a general predictive deficit in early aspects of self‐monitoring associated with positive symptoms in patients with schizophrenia. Storchak, Ehlis, and Fallgatter provide an extensive literature review on electrophysiological measurements, which indicate that general predictive deficits in self‐monitoring are associated with various positive symptoms in patients with schizophrenia. [ABSTRACT FROM AUTHOR]

In this paper, multivalued data or multiple values variables are defined. They are typical when there is some intrinsic uncertainty in data production, as the result of imprecise measuring instruments, such as in image recognition, in human judgments and so on. So far, contributions in symbolic data analysis literature provide data preprocessing criteria allowing for the use of standard methods such as factorial analysis, clustering, discriminant analysis, tree-based methods. As an alternative, this paper introduces a methodology for supervised classification, the so-called Dynamic CLASSification TREE (D-CLASS TREE), dealing simultaneously with both standard and multivalued data as well. For that, an innovative partitioning criterion with a tree-growing algorithm will be defined. Main result is a dynamic tree structure characterized by the simultaneous presence of binary and ternary partitions. A real world case study will be considered to show the advantages of the proposed methodology and main issues of the interpretation of the final results. A comparative study with other approaches dealing with the same types of data will be also shown. The comparison highlights that, even if the results are quite similar in terms of error rates, the proposed D-CLASS tree returns a more interpretable tree-based structure. [ABSTRACT FROM AUTHOR]

Background: Learning to predict threatening events enables an organism to engage in protective behavior and prevent harm. Failure to differentiate between cues that truly predict danger and those that do not, however, may lead to indiscriminate fear and avoidance behaviors, which in turn may contribute to disability in people with persistent pain. We aimed to test whether people with persistent neck pain exhibit contingency learning deficits in predicting pain relative to pain-free, gender-and age-matched controls. Method: We developed a differential predictive learning task with a neck pain-relevant scenario. During the acquisition phase, images displaying two distinct neck positions were presented and participants were asked to predict whether these neck positions would lead to pain in a fictive patient with persistent neck pain (see fictive patient scenario details in Appendix A). After participants gave their pain-expectancy judgment in the hypothetical scenario, the verbal outcome (PAIN or NO PAIN) was shown on the screen. One image (CS+) was followed by the outcome "PAIN", while another image (CS-) was followed by the outcome "NO PAIN". During the generalization phase, novel but related images depicting neck positions along a continuum between the CS+ and CS- images (generalization stimuli; GSs) were introduced to assess the generalization of acquired predictive learning to the novel images; the GSs were always followed by the verbal outcome "NOTES UNREADABLE" to prevent extinction learning. Finally, an extinction phase was included in which all images were followed by "NO PAIN" assessing the persistence of pain-expectancy judgments following disconfirming information. Results: Differential pain-expectancy learning was reduced in people with neck pain relative to controls, resulting from patients giving significantly lower pain-expectancy judgments for the CS+, and significantly higher pain-expectancy judgments for the CS-. People with neck pain also demonstrated flatter generalization gradients relative to controls. No differences in extinction were noted. Discussion: The results support the hypothesis that people with persistent neck pain exhibit reduced differential pain-expectancy learning and flatter generalization gradients, reflecting deficits in predictive learning. Contrary to our hypothesis, no differences in extinction were found. These findings may be relevant to understanding behavioral aspects of chronic pain. [ABSTRACT FROM AUTHOR]

During the mineral flotation process, the surface froth image contains characteristic information that is closely related to the production index of the process. However, a delay often exists between the state of the flotation tank and the taste calculation results, which hinders real-time process regulation. To address this challenge, a video sequence prediction algorithm is proposed to forecast the future sequence of flotation froth, and a Spatiotemporal Motion-Attentive Long Short-Term Memory (STMA-LSTM) network model is introduced to estimate and predict the future image information of the froth by integrating spatiotemporal information. After obtaining the future froth image, an image velocity measurement algorithm is utilized to achieve the future velocity prediction of the flotation froth. To address the ambiguity of future froth images and the real-time demand of future velocity estimation, a Fast Froth Velocity Estimation (FFVE) algorithm is introduced. This algorithm is suitable for fast estimation of flotation froth velocity due to its low sensitivity to image blurring factors and fast feature matching. The proposed approach implements future motion velocity prediction of froth images, thereby addressing the delay in froth taste prediction and enhancing the overall efficiency of the process control in the flotation field. • Proposed STMA-LSTM and FFVE methods for flotation froth future velocity prediction. • STMA-LSTM efficiently captures sequence images' long-term time information. • The captured information generates flotation froth future sequence image predictions. • FFVE accurately measures the flotation froth future sequence image velocity. • Future froth velocity provides a more real-time indicator for flotation monitoring. [ABSTRACT FROM AUTHOR]

Spatiotemporal prediction is a challenging topic because of uncertainty. Existing works attempt to design complicated systems to learn short-term and long-term dynamics. However, these models suffer from heavy computational burden on spatiotemporal data, especially for high-resolution frames. To reduce resource dependence, we propose FastNet, a novel and light encoder–decoder for predictive learning. We stack four ConvLSTM (Convolutional Long Short-Term Memory) based layers to construct a hierarchical framework. Based on this architecture, the feature aggregation module first aligns the temporal context, then decouples different frequency information, next gathers multi-level features and last synthesizes new feature maps alternately. We aggregate various and hierarchical features into predictions bringing two benefits: rich multi-level features and low resource usage. As for the unit blocks, depth-wise separable convolution is used to improve model efficiency and compress model size. Besides, we adopt perceptual loss as the cost function between ground truths and predictions, which helps our model to get higher similarity with true frames. In experiments, we evaluate the performances of FastNet on the MovingMNIST (Mixed National Institute of Standards and Technology) and Radar Echo datasets to verify its effectiveness. The quantitative metrics on the Radar Echo dataset show that FastNet reaches a slight increase in accuracy but up to 84% decline in computation compared with PredRNN-V2. Therefore, our FastNet achieves competitive results with lower resource usage and fewer parameters than the state-of-the-art model. [ABSTRACT FROM AUTHOR]

In property induction tasks, encountering a diverse range of instances (e.g., hippos and hamsters) with a given property usually increases our willingness to generalise that property to a novel instance, relative to non-diverse evidence (e.g., hippos and rhinos). Although generalisation in property induction and predictive learning tasks share conceptual similarities, it is unknown whether this diversity principle applies to generalisation of a predictive association. We tested this hypothesis in two predictive learning experiments using differential training where one category of stimuli (e.g., fruits) predicted an outcome and another category (e.g., vegetables) predicted no outcome. We compared generalisation between a Non-Diverse group who were presented with non-diverse evidence in both positive (predicted the outcome) and negative (predicted no outcome) categories, and two groups who received the same training as the Non-Diverse group but with a more diverse range of exemplars in the positive (Diverse+ group) or negative (Diverse– group) category. Diversity effects were found for both positive and negative categories, in that learning about a diverse range of exemplars increased generalisation of a predictive association to novel exemplars from that same category. The results suggest that diversity, a key principle describing how we reason inductively, also applies to generalisation in associative learning tasks. [ABSTRACT FROM AUTHOR]

What is a fundamental ability for cognitive development? Although many researchers have been addressing this question, no shared understanding has been acquired yet.We propose that predictive learning of sensorimotor signals plays a key role in early cognitive development. The human brain is knownto represent sensorimotor signals in a predictive manner, i.e. it attempts to minimize prediction error between incoming sensory signals and top-down prediction. We extend this view and suggest that two mechanisms for minimizing prediction error lead to the development of cognitive abilities during early infancy. The first mechanism is to update animmature predictor. The predictor must be trained through sensorimotor experiences because it does not inherently have prediction ability. The second mechanism is to execute an action anticipated by the predictor. Interacting with other individuals often increases prediction error, which can be minimized by executing one's own action corresponding to others' action. Our experiments using robotic systems replicated developmental dynamics observed in infants. The capabilities of self-other cognition and goal-directed action were acquired based on the first mechanism, whereas imitation and prosocial behaviours emerged based on the second mechanism. Our theory further provides a potential mechanism for autism spectrum condition. Atypical tolerance for prediction error is hypothesized to be a cause of perceptual and social difficulties. [ABSTRACT FROM AUTHOR]

Predicting the futures from previous spatiotemporal data remains a challenging topic. There have been many previous works on predictive learning. However, mainstream models suffer from huge memory usage or the gradient vanishing problem. Enlightened by the idea from the resnet, we propose CostNet, a novel recursive neural network (RNN)-based network, which has a horizontal and vertical cross-connection. The core of this network is a concise unit, named Horizon LSTM with a fast gradient transmission channel, which can extract spatial and temporal representations effectively to alleviate the gradient propagation difficulty. In the vertical direction outside of the unit, we add overpass connections from unit output to the bottom layer, which can capture the short-term dynamics to generate precise predictions. Our model achieves better prediction results on moving-mnist and radar datasets than the state-of-the-art models.

The treatment of anxiety-based psychopathology often hinges upon extinction learning. Research in nutritional neuroscience has observed that the regular consumption of perilla oil (50% alpha-linolenic acid (ALA)) facilitates extinction learning in rats (Yamamoto et al., 1988). However, acute facilitation of extinction learning by oils rich in ALA has not been reported for rats or humans, though the acute consumption of rapeseed oil (10% ALA) has been observed to improve cognitive processing speed in humans (Jones, Sünram-Lea, & Wesnes, 2012). For this reason, the present laboratory work examined the effects of adding walnut oil (12% ALA) to a chocolate milkshake on the acquisition, generalization, and extinction of a fear-based prediction in young adults. It compared performance between subjects. The other participants consumed a similar milkshake with either an equicaloric amount of cream (saturated fat), or with no added fat (control). Acquisition and generalization of the fear-based prediction were similar for all groups. However, those who consumed walnut oil extinguished most rapidly and profoundly. Implications for extinction learning are discussed. [ABSTRACT FROM AUTHOR]