Cognitive Overhead Research Articles

Automatic topic labeling using graph-based pre-trained neural embedding

It is necessary to reduce the cognitive overhead of interpreting the native topic term list of the Latent Dirichlet Allocation (LDA) style topic model. In this regard, automatic topic labeling has become an effective approach to generate meaningful alternative representations of topics discovered for end-users. In this study, we introduced a novel two-phase neural embedding framework with the redundancy-aware graph-based ranking process. It demonstrated how pre-trained neural embedding could be usefully applied in topic terms, sentence presentations, and automatic topic labeling tasks. Moreover, reranking the topic terms optimized the discovered topics with fewer yet more representative terms while retaining the topic information integrality and fidelity. It further decreased the burden of computation caused by neural embedding and improved the overall effectiveness of the labeling system. Compared with the prevailing state-of-the-art and classical labeling systems, our efficient model boosted the quality of the topic labels generated and discovered more meaningful topic labels.

Automated Social Text Annotation With Joint Multilabel Attention Networks.

Automated social text annotation is the task of suggesting a set of tags for shared documents on social media platforms. The automated annotation process can reduce users' cognitive overhead in tagging and improve tag management for better search, browsing, and recommendation of documents. It can be formulated as a multilabel classification problem. We propose a novel deep learning-based method for this problem and design an attention-based neural network with semantic-based regularization, which can mimic users' reading and annotation behavior to formulate better document representation, leveraging the semantic relations among labels. The network separately models the title and the content of each document and injects an explicit, title-guided attention mechanism into each sentence. To exploit the correlation among labels, we propose two semantic-based loss regularizers, i.e., similarity and subsumption, which enforce the output of the network to conform to label semantics. The model with the semantic-based loss regularizers is referred to as the joint multilabel attention network (JMAN). We conducted a comprehensive evaluation study and compared JMAN to the state-of-the-art baseline models, using four large, real-world social media data sets. In terms of F1 , JMAN significantly outperformed bidirectional gated recurrent unit (Bi-GRU) relatively by around 12.8%-78.6% and the hierarchical attention network (HAN) by around 3.9%-23.8%. The JMAN model demonstrates advantages in convergence and training speed. Further improvement of performance was observed against latent Dirichlet allocation (LDA) and support vector machine (SVM). When applying the semantic-based loss regularizers, the performance of HAN and Bi-GRU in terms of F1 was also boosted. It is also found that dynamic update of the label semantic matrices (JMANd) has the potential to further improve the performance of JMAN but at the cost of substantial memory and warrants further study.

Heimdall, a Computer Program for Electronic Health Records Data Visualization.

Electronic health records (EHR) comprehend structured and unstructured data, that are usually time dependent, enabling the use of timelines. However, it is often difficult to display all data without inducing information overload. In both clinical usual care and medical research, users should be able to quickly find relevant information, with minimal cognitive overhead. Our goal was to devise simple visualization techniques for handling medical data in both contexts. An abstraction layer for structured EHR data was devised after an informal literature review and discussions between authors. The "Heimdall" prototype was developed. Two experts evaluated the tool by answering 5 questions on 24 clinical cases. Temporal data was abstracted in three simple types: events, states and measures, with appropriate visual representations for each type. Heimdall can load and display complex heterogeneous structured temporal data in a straightforward way. The main view can display events, states and measures along a shared timeline. Users can summarize data using temporal, hierarchical compression and filters. Default and custom views can be used to work in problem- oriented ways. The evaluation found conclusive results. The "Heimdall" prototype provides a comprehensive and efficient graphical interface for EHR data visualization. It is open source, can be used with an R package, and is available at https://koromix.dev/files/R.

A Tool for C++ Header Generation

This paper presents a novel approach in the field of C++ development for increasing performance by reducing cognitive overhead and complexity, which results in lower costs. C++ code is split into header and cpp files. This split induces code redundancy. In addition, there are (commonly used) features for classes in C++ that are not supported by recent compilers. The developer must maintain two different files for one single content and implements unsupported features by hand. This leads to the unnecessary cognitive overhead and complex sources. The result is low development performance and high development cost. Our approach utilizes an enhanced syntax inside cpp files. It allows header file generation and therefore obsoletes the need to main-tain a header file. It also enables the generation of fea-tures/methods for classes. It aims to decrease cognitive overhead and complexity, so developers can focus on more sophisticated tasks. This will lead to increased performance and lower costs.

Design and Implementation of TeleAdvisor: a Projection-Based Augmented Reality System for Remote Collaboration

TeleAdvisor is a versatile projection-based augmented reality system designed for remote collaboration. It allows a remote expert to naturally guide a local user in need of assistance in carrying out physical tasks around real-world objects. The system consists of a small projector and two cameras mounted on top of a tele-operated robotic arm at the worker's side, and an interface to view the camera stream, control the point-of-view and gesture using projected annotations at the remote expert's side. TeleAdvisor provides a hands-free, mobile, low-cost solution that supports gesturing by the remote expert while minimizing the cognitive overhead of the local worker. We describe the challenges, design considerations and implementation details of the two phases of the TeleAdvisor prototype, as well as its evaluation and deployment at an industrial manufacturing center. We summarize our understandings from our experiences during the project and discuss the general implications for design of augmented reality remote collaboration systems.

HoneyString: an improved methodology over tag digit-based honeypot to detect shoulder surfing attack

Shoulder surfing attack is often a matter of concern if one is using a public computer system to submit her login credentials. Many methodologies have been proposed by the researchers to prevent such attack. Most of the schemes require high cognitive skills from user end and due to that these schemes are less implementable in real life scenario. So instead of prevention, we work on developing detection of shoulder surfing attack as the detection scheme requires less cognitive overhead than prevention schemes. In this paper, we have proposed a detection mechanism termed as HoneyString which overcomes the limitation of previously proposed tag digit-based scheme. HoneyString provides robust security against DoS attack which was a limitation in the previously proposed scheme. A comparative analysis shows that the proposed scheme has higher detection rate and requires less login time than the existing scheme.

A Resource Constrained Project Scheduling Problem with Bounded Multitasking

Abstract The Resource Constrained Project Scheduling Problem (RCPSP) is the sub-class of the Scheduling Problem that deals with scenarios where personnel or workforce employed to perform the tasks are limited and each job has an arrival time, a due date, and a penalty associated to delays. In this article, we consider a special case of the RCPSP, which we will denote as the Resource Constrained Project Scheduling Problem with Bounded Multitasking (RCPSPBM), where human analysts are able to perform a limited number of tasks in parallel in exchange for cognitive overhead costs. We present a description of the problem and introduce an exact Mixed-Integer Linear Program formulation of the (RCPSPBM), used to evaluate the efficiency of (typically online) algorithms employed by dispatchers.

Improving habitability of natural language interfaces for querying ontologies with feedback and clarification dialogues

Natural Language Interfaces (NLIs) are a viable, human-readable alternative to complex, formal query languages like SPARQL, which are typically used for accessing semantically structured data (e.g. RDF and OWL repositories). However, in order to cope with natural language ambiguities, NLIs typically support a more restricted language. A major challenge when designing such restricted languages is habitability–how easily, naturally and effectively users can use the language to express themselves within the constraints imposed by the system. In this paper, we investigate two methods for improving the habitability of a Natural Language Interface: feedback and clarification dialogues. We model feedback by showing the user how the system interprets the query, thus suggesting repair through query reformulation. Next, we investigate how clarification dialogues can be used to control the query interpretations generated by the system. To reduce the cognitive overhead, clarification dialogues are coupled with a learning mechanism. Both methods are shown to have a positive effect on the overall performance and habitability.

Improving Habitability of Natural Language Interfaces for Querying Ontologies With Feedback and Clarifcation Dialogues

Natural Language Interfaces (NLIs) are a viable, human-readable alternative to complex, formal query languages like SPARQL, which are typically used for accessing semantically structured data (e.g.RDF and OWL repositories). However, in order to cope with natural language ambiguities, NLIs typically support a more restricted language. A major challenge when designing such restricted languages is habitability–how easily, naturally and effectively users can use the language to express themselves within the constraints imposed by the system. In this paper, we investigate two methods for improving the habitability of a Natural Language Interface: feedback and clarifcation dialogues. We model feedback by showing the user how the system interprets the query,thus suggesting repair through query reformulation. Next, we investigate how clarifcation dialogues can be used to control the query interpretations generated by the system. To reduce the cognitive overhead, clarifcation dialogues are coupled with a l earning mechanism. Both methods are shown to have a positive effect on the overall performance and habitability.

The Consumer Choice of E-Channels as a Purchasing Avenue: An Empirical Investigation of the Communicative Aspects of Information Quality

The vast majority of retail customers use electronic channels (e-channels) to search for product information, but do not complete the purchasing process online. This research investigates the role of information quality in transitioning retail consumers to complete the purchasing process online, thereby replacing physical channels with online channels. This research makes three contributions: 1) it highlights the importance of information quality and its influence on a consumer's choice to use e-channels to purchase online; 2) it supports the premise that high-quality information can convert experience attributes into search attributes; and 3) it identifies four antecedents that increase perceived information quality: higher telepresence, screening capability, channel trustworthiness, and lower cognitive overhead. Data from 309 consumers were analyzed using structural equation modeling and regression. Implications and future research avenues are discussed.

A Method for Reducing Disorientation in Hypermedia Educational Systems

Learners in hypermedia educational systems (HES) often become disorientated and experience cognitive overhead. This study presents a method for reducing disorientation in HES. The study uses two usability methods, retrospective thinking aloud (RTA) and performance measure (PM), to find usability problems in HES that impact participant disorientation. The participants had field independent (FI) and field dependent (FD) cognitive styles. The results indicate that the participants' disorientation degree using the RTA method is significantly less than those using the PM method. For the effect of cognitive styles, only the average search steps index showed that FI learners are significantly less disorientated than FD learners. The results can help designers of hypermedia systems design tools to reduce user disorientation.

Cognitive load in hypermedia reading comprehension: Influence of text type and linearity

In this paper the assumption of cognitive overhead in hypermedia learning is specified by cognitive load theory. This analysis is based on different types of cognitive load, the dimension of linearity/non-linearity as well as text characteristics. We propose a model stating that extraneous cognitive load in hypermedia learning is basically determined by the interaction of text presentation format (linear/non-linear) with text type (text with and without narrative structures). This assumption was tested by means of a 2 × 2 experimental design. Sixty participants completed a computer-based learning program that contained a narrative text or an encyclopaedia text in either linear or non-linear presentation format. Results confirm the suggested interaction hypothesis postulating that non-linear information presentation of narrative text structure increases cognitive load and decreases knowledge acquisition. However, for encyclopaedia text participants’ knowledge acquisition was not affected by linear or non-linear presentation format. Furthermore, results suggest a cross-validation of cognitive load measures and propositional analysis.

From DQ to EQ

A fix for irrelevant information, cognitive overhead, and disorientation---common gremlins endured by every e-business system user.

Using knowledge anchors to reduce cognitive overhead

Cognitive overhead, a well-known Web design problem, was defined by Conklin as the additional effort and concentration necessary to maintain several tasks or trails at one time. Cognitive overhead can be reduced using bidirectional knowledge anchors, points within a frame that have significance to the author-and later to the reader. A knowledge anchor is the point within the frame from which the user can trigger a link. An anchor designates either an area from which a reader may want to branch to investigate other related material or an area to which a reader may want to jump.

Concepts for improved visualization of Web link attributes

This paper discusses methods to generate and display automatically additional hyperlink information to the users of the World Wide Web. Current Web browsers make it hard to predict what will happen if a link is followed: users get different information than they expect, a new window may be opened, a download starts, or the destination object is just not available. Instead of giving an appropriate notification in advance, users have to follow a link, check whether the document contains the expected information, get back, try another link etc. However, usually it is possible to obtain additional hyperlink information from several sources like link anchor tags, the user’s history and Web servers. Furthermore, with little enhancements, Web servers may include even more additional information to the hyperlinks in Web documents. These can be displayed before users select a link to improve navigation and reduce the cognitive overhead. In this paper several types of Web hyperlink information are listed, potential methods to present these facts are compared, the prototype implementation of the proposed concept — called by us HyperScout — is presented, and further developments are discussed.

Fuzzy management of user actions during hypermedia navigation

The recent dramatic advances in the field of multimedia systems has made pacticable the development of an Intelligent Tutoring Multimedia (ITM). In these systems are present hypertextual structures that belongs to the class hypermedia systems. ITM development involves the definition of a suitable navigation model in addition to the other modules of an Intelligent Tutoring System (ITS), i.e. Database module, User module, Interface module, Teaching module. The navigation module receives as inputs the state of the system and the user's current assessment and tries to optimize the fruition of the knowledge base. Moreover, this module is responsible for managing the effects of disorientation and cognitive overhead. In this paper we deal essentially with four topics: 1. (i) to define a fuzzy-based user model able to manage adequately the user's cognitive state, the orientation, and the cognitive overhead; 2. (ii) to introduce fuzzy tools within the navigation module in order to carry out moves on the grounds of meaningful data; 3. (iii) to define a set of functions that can dynamically infer new states concerning user's interests; 4. (iv) to classify the hypermedia actions according to their semantics.

The spatial relationship between cues and targets mediates inhibition of return.

Abstract Although inhibition of return (IOR) has been examined in a number of experimental circumstances, it is not known if the number of potential target locations affects the magnitude of the inhibition at a cued location. To investigate this issue, participants were randomly presented with displays of two, four, or six potential target locations in a typical IOR detection task. The findings indicated that number of target locations did not affect IOR and that RTs at uncued locations varied systematically as a function of the spatial relationship between the cued and uncued location. Resume Bien que l'inhibition du retour (IOR) ait ete etudiee dans differents contextes experimentaux, on ne sait pas si le nombre d'emplacements possibles des cibles a un effet sur l'importance de l'inhibition a un emplacement marque par un indice. Pour approfondir l'examen de la question, on a presente aux participants des series de 2, 4 ou 6 emplacements possibles de cibles dans le contexte d'une tache typique de detection IOR. Les resultats indiquent que le nombre d'emplacements des cibles n'influe pas sur l'inhibition du retour et que le temps de reponse, pour les emplacements sans indice, varie systematiquement en fonction de la relation spatiale entre les emplacements accompagnes et non accompagnes d'indices. A considerable amount of research has examined inhibition of return (IOR) since the discovery of the phenomenon by Posner and Cohen in 1984. Posner and Cohen found that targets were detected more slowly at locations that had previously been cued when the delay between the presentation of the cue and the target exceeded 300 ms. They coined the term IOR for the effect, referring to their original notion that attention was inhibited to return to previously attended locations. Presumably, this inhibition occurred to make visual or attentional searches more efficient by biasing attention toward novel locations and away from previously inspected locations. It is because of this implication that IOR has been a topic for researchers in visual attention for more than a decade. Despite the wealth of research that has been conducted on IOR, some basic questions regarding the inhibitory effect have remained unasked and, consequently, unanswered. One of these questions is: Does the number of potential target locations affect IOR? This is a pertinent question because studies that have examined IOR have used anywhere from two (e.g., Maylor, 1984) to eight (e.g., Wright & Richard, 1996) potential target locations. Without knowing whether the number of potential target locations does or does not affect IOR, it is difficult to examine and interpret the different inhibitory effects produced between various experimental conditions. In addition to the pragmatic reasons for examining the effect of various possible target locations on IOR, there are also theoretical reasons to suggest that the number of locations may, or may not, differently affect IOR. On the one hand, if IOR does reflect the output of an attentional mechanism that improves searches, one might expect that the number of locations where the target might potentially occur would have a negative influence on the magnitude of IOR. Assuming that some amount of attention is allocated to each target location placeholder in a display, larger number of placeholders should require a greater proportion of the available attentional resources. This leads to a situation in which the amount of inhibition might be expected to decrease with increasing numbers of potential target locations because it would become more difficult to keep track of novel and previously attended locations. On the other hand, it may be that previously attended locations are tagged with inhibitory tags that require relatively little cognitive overhead, making the number of potential target locations largely irrelevant. For example, Klein and Taylor (1994) suggested that IOR occurs when eye movements are planned, but not executed, toward a specific location. …

HyperTree: a structural approach to web authoring

The expansion of the World Wide Web (WWW) has created an increasing need for tools capable of supporting WWW authors in composing documents using the HyperText Markup Language (HTML). Currently, most web authors use tools which are basically ordinary text editors and have additional features to facilitate the easy and correct use of HTML tags. This approach places the burden on the web author to design and then create the entire web site in a top-down fashion, without any explicit support for the structural design of the site. In this paper we discuss an alternative structural approach to Web authoring, which is based on the use of the HyperTree hypermedia system as the central authoring tool. The advantages of using HyperTree are two-dimensional. Firstly, web authors can manage a web site as a single complete hypermedia database. For example, HyperTree provides facilities like the automatic creation of indices and the discovery of link inconsistencies. Additionally, it organizes the web pages in an easy to understand hierarchy without using any HTML directly. Secondly, web end-users can benefit from the use of HyperTree, since seeking information in structured web sites is generally less disorientating and develops fewer cognitive overheads. ©1997 John Wiley & Sons, Ltd.

Disorientation and Cognitive Overhead in Hypertext Systems

Disorientation and cognitive overhead are two interrelated problems that may in fact ultimately limit the usefulness of hypertext. Cognitive overhead upon the author arises from a viewpoint that sees the author as the person solely responsible for inserting all links within her work. This includes all links between the anchors in her own work and those in external frames, as well as cross-links between the anchors within her own work. The need for an author to be aware, for each anchor within her work, of all possible other anchors makes the insertion of such links an O(N2) problem where N is the number of possible anchors in the hypertext universe. To reduce this to an O(N) problem this paper proposes the use of bi-directional intelligent anchors. These are anchors that have both a knowledge of the concepts that a potential reader would be seeking were she to arrive at the anchor, together with a knowledge of the concepts that a reader would be seeking were she to jump from that anchor to another within the hypertext. Disorientation or the lost in hyperspace effect on the reader can be addressed in many cases by a fundamental re-definition of the relationship between hypertext authors and their readers. To address this issue, this paper proposes a unit of authorship based more on the treatise than on the traditional fragmented model of hypertext that we call a scenic route and which is based loosely on Bush (1945)'s Trails and Trigg (1988 and 1991)'s Guided Tours.

The role of working memory on graphical information processing

This research extends previous graphics research by examining how individual differences in working memory (WM) capacity and changes in graphic design influence graphical information processing. An experiment compared decision accuracy of two graphic decision aids and an unaided group for a task at two levels of complexity. There were no accuracy differences for the low complexity task. At high levels of task complexity, accuracy depended upon WM capacity and how the graphic aid influenced perception. Eye movement data show information processing differences also are contingent upon graphic design features and WM capacity. We postulate that graphs reduce cognitive overhead by shifting some of the cognitive burden to our visual perception system. More efficient graphical perceptual will improve decision performance only if our cognitive resources are capacity constrained and those cognitive resources are used elsewhere in the problem solving process.