Task Representation Research Articles

Moving forward: Exploring the role of interference on prospective memory deactivation

Recent prospective memory (PM) studies have shown that an intention may be erroneously executed despite no-longer-needed (i.e., commission errors), especially under demanding ongoing activities. In the current study, we examined whether PM deactivation benefits from a retroactive interference mechanism. For this, we set up a procedure in which participants are first asked to perform a PM task which is critically declared finished afterwards. Next, they encoded a new and dissimilar PM intention to accomplish later (Experiment 1) or performed filler tasks with increased working memory difficulty levels (Experiment 2). Lastly, all participants encountered several (but irrelevant) PM cues. Together, our findings provide evidence that the efficiency of the deactivation process can be modulated by encoding novel and dissimilar PM tasks and by the type of processing after intention completion. These findings are discussed in terms of strategic or spontaneous retrieval processes and linked to a retroactive interference mechanism which helps to overwrite or deteriorate the old-PM task representation.

Context Matters in Gifted Education

Bruce M. Shore’s research contributions in gifted education have focused on three contexts that impact how giftedness is understood and the instructional environments that serve gifted learners’ educational needs. This article describes these contributions and provides selected examples plus a more complete Supplemental Online bibliography. First, giftedness benefits from being conceptualized in terms of theories that address the development of expertise. Featured expert–gifted parallels include interconnectedness of knowledge, metacognitive processes, perspective taking, active learner roles, affinity for novelty and complexity, and task representation and planning. Illustrative research is described from preschool age through higher education, including connections to creativity research. Second, gifted education benefits when guided by social-constructivist theory of education and its expression in inquiry-based instruction. Examples include building upon learner interests, question asking, collaborative inquiry, and active learner roles. Desirable specific instructional practices are framed by the above theories and by being considered in the contexts of widely recommended and best practices with their research support. Third, gifted education, at all levels including higher education and teacher education, needs to be an integral part of the context of general education. Most specific gifted education practices also work in general education, including learning high-level skills within subject matter. Nineteen examples are cited about how gifted education contributes to the quality of general education.

Review of unsupervised pretraining strategies for molecules representation.

In recent years, the computer-assisted techniques make a great progress in the field of drug discovery. And, yet, the problem of limited labeled data problem is still challenging and also restricts the performance of these techniques in specific tasks, such as molecular property prediction, compound-protein interaction and de novo molecular generation. One effective solution is to utilize the experience and knowledge gained from other tasks to cope with related pursuits. Unsupervised pretraining is promising, due to its capability of leveraging a vast number of unlabeled molecules and acquiring a more informative molecular representation for the downstream tasks. In particular, models trained on large-scale unlabeled molecules can capture generalizable features, and this ability can be employed to improve the performance of specific downstream tasks. Many relevant pretraining works have been recently proposed. Here, we provide an overview of molecular unsupervised pretraining and related applications in drug discovery. Challenges and possible solutions are also summarized.

Rule Perseveration during Task-Switching in Brain Tumor: A Severe Form of Task-Setting Impairment.

It has been proposed that at least two distinct processes are engaged during task-switching: reconfiguration of the currently relevant task-set and interference resolution arising from the competing task-set. Whereas in healthy individuals the two are difficult to disentangle, their disruption is thought to cause different impairments in brain-damaged patients. Yet, the observed deficits are inconsistent across studies and do not allow drawing conclusions regarding their independence. Forty-one brain tumor patients were tested on a task-switching paradigm. We compared their performance between switch and repeat trials (switch cost) to assess rule reconfiguration, and between trials requiring the same response (congruent) and a different response for the two tasks (incongruent) to assess interference control. In line with previous studies, we found the greatest proportion of errors on incongruent trials, suggesting an interference control impairment. However, a closer look at the distribution of errors between two task rules revealed a rule perseveration impairment: Patients with high error rate on incongruent trials often applied only one task rule throughout the task and less frequently switched to the alternative one. Multivariate lesion-symptom mapping analysis unveiled the relationship between lesions localized in left orbitofrontal and posterior subcortical regions and perseveration scores, measured as absolute difference in accuracy between two task rules. This finding points to a more severe task-setting impairment, not reflected as a mere switching deficit, but instead as a difficulty in creating multiple stable task representations, in line with recent accounts of OFC functions suggesting its critical role in representing task states.

Multi-task fully convolutional network for tree species mapping in dense forests using small training hyperspectral data

This work proposes a multi-task fully convolutional architecture for tree species mapping in dense forests from sparse and scarce polygon-level annotations using hyperspectral UAV-borne data. Our model implements a partial loss function that enables dense tree semantic labeling outcomes from non-dense training samples, and a distance regression complementary task that enforces tree crown boundary constraints and substantially improves the model performance. Our multi-task architecture uses a shared backbone network that learns common representations for both tasks and two task-specific decoders, one for the semantic segmentation output and one for the distance map regression. We report that introducing the complementary task boosts the semantic segmentation performance compared to the single-task counterpart in up to 11% reaching an average user’s accuracy of 88.63% and an average producer’s accuracy of 88.59%, achieving state-of-art performance for tree species classification in tropical forests.

The Neural Bases of Egocentric Spatial Representation for Extracorporeal and Corporeal Tasks: An fMRI Study.

(1) Background: Humans use reference frames to elaborate the spatial representations needed for all space-oriented behaviors such as postural control, walking, or grasping. We investigated the neural bases of two egocentric tasks: the extracorporeal subjective straight-ahead task (SSA) and the corporeal subjective longitudinal body plane task (SLB) in healthy participants using functional magnetic resonance imaging (fMRI). This work was an ancillary part of a study involving stroke patients. (2) Methods: Seventeen healthy participants underwent a 3T fMRI examination. During the SSA, participants had to divide the extracorporeal space into two equal parts. During the SLB, they had to divide their body along the midsagittal plane. (3) Results: Both tasks elicited a parieto-occipital network encompassing the superior and inferior parietal lobules and lateral occipital cortex, with a right hemispheric dominance. Additionally, the SLB > SSA contrast revealed activations of the left angular and premotor cortices. These areas, involved in attention and motor imagery suggest a greater complexity of corporeal processes engaging body representation. (4) Conclusions: This was the first fMRI study to explore the SLB-related activity and its complementarity with the SSA. Our results pave the way for the exploration of spatial cognitive impairment in patients.

Cognitive Capacity, Representation, and Instruction.

The central argument of the present article is that Cognitive Psychology’s problems in dealing with the concept of “cognitive capacity” is intimately linked with Cognitive Psychology’s long-lasting failure of coming to terms with the concept of “representation” in general, and “task representation” in particular. From this perspective, the role of instructions in psychological experiments is emphasised. It is argued that both a careful conceptual analysis of instruction-induced task representations as well as an experimental variation of instructions promises to broaden our understanding of the role of task representations as a determinant of limited cognitive capacity.

Walking a Thin Line of Representation: Analyzing the Behavior of Egyptian MPs

Abstract This article is part of the Special Issue “Parliaments in the Middle East and North Africa: A Struggle for Relevance.” While legislators in democratic settings have the electorate as their main principal, mps in semi- and nondemocratic settings need to serve two principals to remain in office: the regime and the active segment of the electorate. This dichotymy sometimes requires particular skills in parliamentary behavior. For the case of Egypt, we investigate how mps strike a balance between regime support and representing their constituents up to an extent that does not endanger their chances for re-election. A content analysis of session scripts of the Egyptian parliament in 2016 was conducted to examine how mps walk this – traditionally understudied – thin line. Our findings indicate that representation gets reduced to “descriptive representation,” i.e. a representation that puts more emphasis on representing local constituents and demographic segments, like Copts and women, that mps are presumably elected to represent. We therefore show that mps fulfill the important tasks of citizens representation even in semi- and nondemocratic settings.

One-Layer vs. Two-Layer SOM in the Context of Outlier Identification: A Simulation Study

In this work, we applied a stochastic simulation methodology to quantify the power of the detection of outlying mixture components of a stochastic model, when applying a reduced-dimension clustering technique such as Self-Organizing Maps (SOMs). The essential feature of SOMs, besides dimensional reduction into a discrete map, is the conservation of topology. In SOMs, two forms of learning are applied: competitive, by sequential allocation of sample observations to a winning node in the map, and cooperative, by the update of the weights of the winning node and its neighbors. By means of cooperative learning, the conservation of topology from the original data space to the reduced (typically 2D) map is achieved. Here, we compared the performance of one- and two-layer SOMs in the outlier representation task. The same stratified sampling was applied for both the one-layer and two-layer SOMs; although, stratification would only be relevant for the two-layer setting—to estimate the outlying mixture component detection power. Two distance measures between points in the map were defined to quantify the conservation of topology. The results of the experiment showed that the two-layer setting was more efficient in outlier detection while maintaining the basic properties of the SOM, which included adequately representing distances from the outlier component to the remaining ones.

Abstraction in data-sparse task transfer

When a robot adapts a learned task for a novel environment, any changes to objects in the novel environment have an unknown effect on its task execution. For example, replacing an object in a pick-and-place task affects where the robot should target its actions, but does not necessarily affect the underlying action model. In contrast, replacing a tool that the robot will use to complete a task will effectively alter its end-effector pose with respect to the robot's base coordinate system, and thus the robot's motion must be replanned accordingly.These examples highlight the relationship among (i) differences between the source and target environments, (ii) the level of abstraction at which a robot's task model should be represented to enable transfer to the target environment, and (iii) the information needed to ground the abstracted task representation in the target environment. In this article, we present a taxonomy of transfer problems based on this relationship. We also describe a knowledge representation called the Tiered Task Abstraction (TTA) and demonstrate its applicability to a variety of transfer problems in the taxonomy. Our experimental results indicate a trade-off between the generality and data requirements of a task representation, and reinforce the need for multiple transfer methods that operate at different levels of abstraction.

Majority Decision-Making Works Best Under Conditions of Leadership Ambiguity and Shared Task Representations.

The effectiveness of decision-making teams depends largely on their ability to integrate and make sense of information. Consequently, teams which more often use majority decision-making may make better quality decisions, but particularly so when they also have task representations which emphasize the elaboration of information relevant to the decision, in the absence of clear leadership. In the present study we propose that (a) majority decision-making will be more effective when task representations are shared, and that (b) this positive effect will be more pronounced when leadership ambiguity (i.e., team members’ perceptions of the absence of a clear leader) is high. These hypotheses were put to the test using a sample comprising 81 teams competing in a complex business simulation for seven weeks. As predicted, majority decision-making was more effective when task representations were shared, and this positive effect was more pronounced when there was leadership ambiguity. The findings extend and nuance earlier research on decision rules, the role of shared task representations, and leadership clarity.

Модель межотраслевого баланса Леонтьева как задача управления динамической системой

The purpose of the study was to determine the problem of control of a dynamic system of higher dimension. Relying on Leontev input-output balance, we formalized the dynamic system and synthesized its control. Within the research, we developed a mathematical model that combines different working objects that consume and release various resources. The value of the penalty for all nodes and objects is introduced into the matrix representation of the problem, taking into account various options for their interaction, i.e., the observation problem. A matrix representation of the planning task at each working object is formed. For the formed system, a control loop is created; the influencing parameters of the external environment are indicated. We calculated the system operational mode, taking into account the interaction of the nodes of objects with each other when the parameters of the external environment influence them. Findings of research show that in achieving a complex result, the system is inefficient without optimal planning and accounting for the matrix of penalties for the interaction of nodes and objects of the dynamic system with each other. In a specific example, for a dynamic system with a dimension of 4.8 million parameters, we estimated the control taking into account the penalty matrix, which made it possible to increase the inflow of additional resources from the outside by 2.4 times from 130 billion conv. units up to 310 conv. units in 5 years. Taking into account the maximum optimization of control in the nodes, an increase of 3.66 times in the inflow of additional resources was ensured --- from 200.46 to 726.62 billion rubles

Exploring BCI Control in Smart Environments

The brain–computer interface (BCI) control technology that utilizes motor imagery to perform the desired action instead of manual operation will be widely used in smart environments. However, most of the research lacks robust feature representation of multi-channel EEG series, resulting in low intention recognition accuracy. This article proposes an EEG2Image based Denoised-ConvNets (called EID) to enhance feature representation of the intention recognition task. Specifically, we perform signal decomposition, slicing, and image mapping to decrease the noise from the irrelevant frequency bands. After that, we construct the Denoised-ConvNets structure to learn the colorspace and spatial variations of image objects without cropping new training images precisely. Toward further utilizing the color and spatial transformation layers, the colorspace and colored area of image objects have been enhanced and enlarged, respectively. In the multi-classification scenario, extensive experiments on publicly available EEG datasets confirm that the proposed method has better performance than state-of-the-art methods.

Does the presence of more features in a bound representation in working memory require extra object-based attention?

Recent studies have examined the role of attention in retaining bound representations in working memory (WM) and found that object-based attention plays a pivotal role. However, no study has investigated whether maintaining bound representations with more features in WM requires extra object-based attention. We investigated this by examining whether a secondary task consuming object-based attention was more disruptive to the maintenance of bindings in WM when more features were stored per object. We instructed participants to memorize three bound representations in a WM task while manipulating the number of features (two vs. three features) contained in each representation. Moreover, we manipulated whether a secondary task consuming object-based attention was interpolated into the maintenance phase of WM. If extra object-based attention was required after the addition of an extra feature in the bound representation, the secondary task would result in a greater disruption of the three- rather than two-featured binding. In two experiments, we found that the added secondary task significantly impaired the binding performance, but the performance of the two- and three-featured bindings was disrupted to the same extent. These results suggest that the presence of more features in a bound representation in WM does not require extra object-based attention.

Theoretical Foundations for Structural Symmetries of Lifted PDDL Tasks

We transfer the notion of structural symmetries to lifted planning task representations, based on abstract structures which we define to model planning tasks. We show that symmetries are preserved by common grounding methods and we shed some light on the relation to previous symmetry concepts used in planning. Using a suitable graph representation of lifted tasks, our experimental analysis of common planning benchmarks reveals that symmetries occur in the lifted representation of many domains. Our work establishes the theoretical ground for exploiting symmetries beyond their previous scope, such as for faster grounding and mutex generation, as well as for state space transformations and reductions.

Task-order representations in dual tasks: Separate or integrated with component task sets?

In situations requiring the execution of two tasks at around the same time, we need to decide which of the tasks should be executed first. Previous research has revealed several factors that affect the outcome of such response order control processes, including bottom-up factors (e.g., the temporal order of the stimuli associated with the two tasks) and top-down factors (e.g., instructions). In addition, it has been shown that tasks associated with certain response modalities are preferably executed first (e.g., temporal prioritisation of tasks involving oculomotor responses). In this study, we focused on a situation in which task order has to be unpredictably switched from trial to trial and asked whether task-order representations are coded separately or integrated with the component task sets (i.e., in a task-specific manner). Across three experiments, we combined two tasks known to differ in prioritisation, namely an oculomotor and a manual (or pedal) task. The results indicated robust task-order switch costs (i.e., longer RTs when task order was switched vs. repeated). Importantly, the data demonstrate that it is possible to show an asymmetry of task-order switch costs: While these costs were of similar size for both task orders in one particular experimental setting with specific spatial task characteristics, two experiments consistently indicated that it was easier for participants to switch to their prioritised task order (i.e., to execute the dominant oculomotor task first). This suggests that in a situation requiring frequent task-order switches (indicated by unpredictable changes in stimulus order), task order is represented in an integrated, task-specific manner, bound to characteristics (here, associated effector systems) of the component tasks.

Learning to Pre-train Graph Neural Networks

Graph neural networks (GNNs) have become the defacto standard for representation learning on graphs, which derive effective node representations by recursively aggregating information from graph neighborhoods. While GNNs can be trained from scratch, pre-training GNNs to learn transferable knowledge for downstream tasks has recently been demonstrated to improve the state of the art. However, conventional GNN pre-training methods follow a two-step paradigm: 1) pre-training on abundant unlabeled data and 2) fine-tuning on downstream labeled data, between which there exists a significant gap due to the divergence of optimization objectives in the two steps. In this paper, we conduct an analysis to show the divergence between pre-training and fine-tuning, and to alleviate such divergence, we propose L2P-GNN, a self-supervised pre-training strategy for GNNs. The key insight is that L2P-GNN attempts to learn how to fine-tune during the pre-training process in the form of transferable prior knowledge. To encode both local and global information into the prior, L2P-GNN is further designed with a dual adaptation mechanism at both node and graph levels. Finally, we conduct a systematic empirical study on the pre-training of various GNN models, using both a public collection of protein graphs and a new compilation of bibliographic graphs for pre-training. Experimental results show that L2P-GNN is capable of learning effective and transferable prior knowledge that yields powerful representations for downstream tasks. (Code and datasets are available at https://github.com/rootlu/L2P-GNN.)

Learning Contextual Representations for Semantic Parsing with Generation-Augmented Pre-Training

Most recently, there has been significant interest in learning contextual representations for various NLP tasks, by leveraging large scale text corpora to train powerful language models with self-supervised learning objectives, such as Masked Language Model (MLM). Based on a pilot study, we observe three issues of existing general-purpose language models when they are applied in the text-to-SQL semantic parsers: fail to detect the column mentions in the utterances, to infer the column mentions from the cell values, and to compose target SQL queries when they are complex. To mitigate these issues, we present a model pretraining framework, Generation-Augmented Pre-training (GAP), that jointly learns representations of natural language utterance and table schemas, by leveraging generation models to generate high-quality pre-train data. GAP Model is trained on 2 million utterance-schema pairs and 30K utterance-schema-SQL triples, whose utterances are generated by generation models. Based on experimental results, neural semantic parsers that leverage GAP Model as a representation encoder obtain new state-of-the-art results on both Spider and Criteria-to-SQL benchmarks.

GRASP: Generic Framework for Health Status Representation Learning Based on Incorporating Knowledge from Similar Patients

Deep learning models have been applied to many healthcare tasks based on electronic medical records (EMR) data and shown substantial performance. Existing methods commonly embed the records of a single patient into a representation for medical tasks. Such methods learn inadequate representations and lead to inferior performance, especially when the patient’s data is sparse or low-quality. Aiming at the above problem, we propose GRASP, a generic framework for healthcare models. For a given patient, GRASP first finds patients in the dataset who have similar conditions and similar results (i.e., the similar patients), and then enhances the representation learning and prognosis of the given patient by leveraging knowledge extracted from these similar patients. GRASP defines similarities with different meanings between patients for different clinical tasks, and finds similar patients with useful information accordingly, and then learns cohort representation to extract valuable knowledge contained in the similar patients. The cohort information is fused with the current patient’s representation to conduct final clinical tasks. Experimental evaluations on two real-world datasets show that GRASP can be seamlessly integrated into state-of-the-art models with consistent performance improvements. Besides, under the guidance of medical experts, we verified the findings extracted by GRASP, and the findings are consistent with the existing medical knowledge, indicating that GRASP can generate useful insights for relevant predictions.

Conceptualized and Contextualized Gaussian Embedding

Word embedding can represent a word as a point vector or a Gaussian distribution in high-dimensional spaces. Gaussian distribution is innately more expressive than point vector owing to the ability to additionally capture semantic uncertainties of words, and thus can express asymmetric relations among words more naturally (e.g., animal entails cat but not the reverse. However, previous Gaussian embedders neglect inner-word conceptual knowledge and lack tailored Gaussian contextualizer, leading to inferior performance on both intrinsic (context-agnostic) and extrinsic (context-sensitive) tasks. In this paper, we first propose a novel Gaussian embedder which explicitly accounts for inner-word conceptual units (sememes) to represent word semantics more precisely; during learning, we propose Gaussian Distribution Attention over Gaussian representations to adaptively aggregate multiple sememe distributions into a word distribution, which guarantees the Gaussian linear combination property. Additionally, we propose a Gaussian contextualizer to utilize outer-word contexts in a sentence, producing contextualized Gaussian representations for context-sensitive tasks. Extensive experiments on intrinsic and extrinsic tasks demonstrate the effectiveness of the proposed approach, achieving state-of-the-art performance with near 5.00% relative improvement.