Distinguishing between early-stage brain tumors (EBTs) and progressive-stage brain tumors (PBTs) from magnetic resonance imaging (MRI) scans is pivotal, as interpretation complexity challenges neuro-oncologists and impacts timely treatment decisions. Existing deep learning (DL)-based approaches typically handle brain tumor classification and segmentation separately, emphasizing multi-level feature extraction, but neglecting the benefits of integrated feature fusion. This study introduces a novel DL-based clinical pipeline, enhanced with generative artificial intelligence (GenAI), that explicitly fuses features from classification and segmentation models for stage-specific tumor analysis. Our fusion-based framework first classifies brain tumors into EBTs and PBTs using a dedicated classification network (C-Net), incorporating an enhanced-pooled attention block and a dilated fusion block to selectively extract and fuse multi-level features, balancing computational efficiency with feature relevance. Subsequently, stage-specific segmentation network 1 (S1-Net) and segmentation network 2 (S2-Net) leverage hierarchical feature fusion through progressive upsampling, capturing distinct tumor characteristics at multiple abstraction levels. Finally, a clinician-validated GenAI module provides post-segmentation semantic interpretation by analyzing segmentation masks and patient metadata to describe tumor morphology and explicitly report limitations. Aligned with information fusion principles, this integration combines the precision of task-specific models (S1-Net, S2-Net) with expert-guided GenAI reasoning, enhancing interpretability while preserving clinical safety. The effectiveness of the proposed pipeline is validated with open dataset of brain tumor progression: 1) C-Net achieves an accuracy of 85.06%, precision of 86.73%, recall of 85%, and harmonic mean of recall and precision (F1-score) of 85.84%; 2) S1-Net attains a Dice score (DS) of 79.96% and intersection over union (IoU) of 71.22%; and 3) S2-Net achieves 81.14% DS and 72.44% IoU, significantly outperforming state-of-the-art methods.

With the availability of egocentric 3D hand-object interaction datasets, there is increasing interest in developing unified models for hand-object pose estimation and action recognition. However, existing methods still struggle to recognise seen actions on unseen objects due to the limitations in representing object shape and movement using 3D bounding boxes. Additionally, the reliance on object templates at test time limits their generalisability to unseen objects. To address these challenges, we propose to leverage superquadrics as an alternative 3D object representation to bounding boxes and demonstrate their effectiveness on both template-free object reconstruction and action recognition tasks. Moreover, as we find that pure appearance-based methods can outperform the unified methods, the potential benefits from 3D geometric information remain unclear. Therefore, we study the compositionality of actions by considering a more challenging task where the training combinations of verbs and nouns do not overlap with the testing split. We extend H2O and FPHA datasets with compositional splits and design a novel collaborative learning framework that can explicitly reason about the geometric relations between hands and the manipulated object. Through extensive quantitative and qualitative evaluations, we demonstrate significant improvements over the state-of-the-arts in (compositional) action recognition.

Social norms govern culturally appropriate behavior in communication, enabling dialogue systems to produce responses that are not only coherent but also socially acceptable. We present NormGenesis, a multicultural framework for generating and annotating socially grounded dialogues across English, Chinese, and Korean. To model the dynamics of social interaction beyond static norm classification, we propose a novel dialogue type, Violation-to-Resolution (V2R), which models the progression of conversations following norm violations through recognition and socially appropriate repair. To improve pragmatic consistency in underrepresented languages, we implement an exemplar-based iterative refinement early in the dialogue synthesis process. This design introduces alignment with linguistic, emotional, and sociocultural expectations before full dialogue generation begins. Using this framework, we construct a dataset of 10,800 multi-turn dialogues annotated at the turn level for norm adherence, speaker intent, and emotional response. Human and LLM-based evaluations demonstrate that NormGenesis significantly outperforms existing datasets in refinement quality, dialogue naturalness, and generalization performance. We show that models trained on our V2R-augmented data exhibit improved pragmatic competence in ethically sensitive contexts. Our work establishes a new benchmark for culturally adaptive dialogue modeling and provides a scalable methodology for norm-aware generation across linguistically and culturally diverse languages.

Suchae Jeong, Inseong Choi, Youngsik Yun, Jihie Kim. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Zero-shot learning (ZSL) is an important and rapidly growing area of machine learning that aims to recognize new classes without prior training data. Despite its significance, ZSL has faced challenges with overfitting in embedding-based methods and limitations in traditional one-directional attention (ODA) based approaches. To bridge these gaps, this paper proposes the use of bi-directional attention (BDA) to integrate insights from both embedding and attention-based approaches. The proposed BDA system consists of a bi-directional attention network (BDAN) and a synthesized visual embedding network (SVEN) that facilitates visual-semantic interaction for ZSL classification. More specifically, the BDAN employs region self-attention (RSA), semantic synthesis attention (SSA), and visual synthesis attention (VSA) to overcome the overfitting issue in embedding methods and enhance transferability, to associate visual features with semantic property information, and to learn locally improved visual features. Extensive testing on CUB, SUN, and AWA2 datasets confirm the superiority of our proposed method over traditional approaches. Code is available at https://github.com/JunseokLee3/BDA.

Abstract When stress causes negative behaviours to emerge in our daily lives, it is important to intervene quickly and appropriately to control the negative problem behaviours. Questionnaires, a common method of information gathering, have the disadvantage that it is difficult to get the exact information needed due to defensive or insincere responses from subjects. As an alternative to these drawbacks, projective testing using pictures can provide the necessary information more accurately than questionnaires because the subject responds subconsciously and the direct experience expressed through pictures can be more accurate than questionnaires. Analysing hand‐drawn image data with the Wartegg Zeichen Test (WZT) is not easy. In this study, we used deep learning to analyse image data represented as pictures through WZT to predict early depression. We analyse the data of 54 people who were judged as early depression and 54 people without depression, and increase the number of people without depression to 100 and 500, and aim to study in unbalanced data. We use CNN and CNN‐SVM to analyse the drawing images of WZT's initial depression with deep learning and predict the outcome of depression. The results show that the initial depression is predicted with 92%–98% accuracy on the image data directly drawn by WZT. This is the first study to automatically analyse and predict early depression in WZT based on hand‐drawn image data using deep learning models. The extraction of features from WZT images by deep learning analysis is expected to create more research opportunities through the convergence of psychotherapy and Information and Communication Technology (ICT) technology, and is expected to have high growth potential.

Accurate representation in media is known to improve the well-being of the people who consume it. Generative image models trained on large web-crawled datasets such as LAION are known to produce images with harmful stereotypes and misrepresentations of cultures. We improve inclusive representation in generated images by (1) engaging with communities to collect a culturally representative dataset that we call the Cross-Cultural Understanding Benchmark (CCUB) and (2) proposing a novel Self-Contrastive Fine-Tuning (SCoFT, pronounced /sô ft/) method that leverages the model's known biases to self-improve. SCoFT is designed to prevent overfitting on small datasets, encode only high-level information from the data, and shift the generated distribution away from misrepresentations encoded in a pretrained model. Our user study conducted on 51 participants from 5 different countries based on their self-selected national cultural affiliation shows that fine-tuning on CCUB consistently generates images with higher cultural relevance and fewer stereotypes when compared to the Stable Diffusion baseline, which is further improved with our SCoFT technique. Resources and code are at https://ariannaliu.github.io/SCoFT.

This thesis deals with the problem of negative behaviour due to changes in mental and physical stress in adolescence.In particular, it is a study to solve the health care problem of students exposed to violence.Among the problematic behaviours, students exposed to violence, especially, have many problems with healthcare.A projective test using pictures can elicit information from adolescents through direct experiences represented by pictures to which the subject unconsciously reacts.Few methods analyse images drawn by adolescents as image data.This study analyses data from 134 violent students who received fifth-degree punishment for violent behaviour and 134 nonviolent students.We use the convolutional neural network (CNN)(softmax), CNN (support vector machine (SVM)), with the style transfer generative adversarial network, and ensemble techniques to analyse drawn images using WZT and predict violence through deep learning.We predict violence from pictures with an accuracy of 93%-98%.This study is the first to automatically analyze and predict violence with a deep learning model in images drawn by adolescents on WZT.It also features WZT to proactively conduct violence investigations to improve health care for students.Advances in deep learning for image feature extraction are expected to provide more research opportunities.

Knowledge-based visual question answering (KVQA) task aims to answer questions that require additional external knowledge as well as an understanding of images and questions. Recent studies on KVQA inject an external knowledge in a multi-modal form, and as more knowledge is used, irrelevant information may be added and can confuse the question answering. In order to properly use the knowledge, this study proposes the following: 1) we introduce a novel semantic inconsistency measure computed from caption uncertainty and semantic similarity; 2) we suggest a new external knowledge assimilation method based on the semantic inconsistency measure and apply it to integrate explicit knowledge and implicit knowledge for KVQA; 3) the proposed method is evaluated with the OK-VQA dataset and achieves the state-of-the-art performance.

Multimodal sentiment analysis has become an increasingly popular research area as the demand for multimodal online content is growing. For multimodal sentiment analysis, words can have different meanings depending on the linguistic context and non-verbal information, so it is crucial to understand the meaning of the words accordingly. In addition, the word meanings should be interpreted within the whole utterance context that includes nonverbal information. In this paper, we present a Context-driven Modality Shifting BERT with Contrastive Learning for linguistic, visual, acoustic Representations (CMSBERT-CLR), which incorporates the whole context's non-verbal and verbal information and aligns modalities more effectively through contrastive learning. First, we introduce a Context-driven Modality Shifting (CMS) to incorporate the non-verbal and verbal information within the whole context of the sentence utterance. Then, for improving the alignment of different modalities within a common embedding space, we apply contrastive learning. Furthermore, we use an exponential moving average parameter and label smoothing as optimization strategies, which can make the convergence of the network more stable and increase the flexibility of the alignment. In our experiments, we demonstrate that our approach achieves state-of-the-art results.

Training chatbots using the reinforcement learning paradigm is challenging due to high-dimensional states, infinite action spaces and the difficulty in specifying the reward function. We address such problems using clustered actions instead of infinite actions, and a simple but promising reward function based on human-likeness scores derived from human-human dialogue data. We train Deep Reinforcement Learning (DRL) agents using chitchat data in raw text—without any manual annotations. Experimental results using different splits of training data report the following. First, that our agents learn reasonable policies in the environments they get familiarised with, but their performance drops substantially when they are exposed to a test set of unseen dialogues. Second, that the choice of sentence embedding size between 100 and 300 dimensions is not significantly different on test data. Third, that our proposed human-likeness rewards are reasonable for training chatbots as long as they use lengthy dialogue histories of ≥10 sentences.

Machine reading comprehension helps machines learn to utilize most of the human knowledge written in the form of text.Existing approaches made a significant progress comparable to human-level performance, but they are still limited in understanding, up to a few paragraphs, failing to properly comprehend lengthy document.In this paper, we propose a novel deep neural network architecture to handle a long-range dependency in RC tasks.In detail, our method has two novel aspects: (1) an advanced memory-augmented architecture and (2) an expanded gated recurrent unit with dense connections that mitigate potential information distortion occurring in the memory.Our proposed architecture is widely applicable to other models.We have performed extensive experiments with well-known benchmark datasets such as TriviaQA, QUASAR-T, and SQuAD.The experimental results demonstrate that the proposed method outperforms existing methods, especially for lengthy documents.

In this paper, we propose a self-learning architecture for generating natural language templates for conversational assistants.Generating templates to cover all the combinations of slots in an intent is time consuming and labor-intensive.We examine three different models based on our proposed architecture -Rule-based model, Sequence-to-Sequence (Seq2Seq) model and Semantically Conditioned LSTM (SC-LSTM) model for the IoT domain -to reduce the human labor required for template generation.We demonstrate the feasibility of template generation for the IoT domain using our self-learning architecture.In both automatic and human evaluation, the self-learning architecture outperforms previous works trained with a fully human-labeled dataset.This is promising for commercial conversational assistant solutions.

Adaptive learning is the core technology behind intelligent tutoring systems, which are responsible for estimating student knowledge and providing personalized instruction to students based on their skill level. In this paper, we present a new adaptive learning system architecture, which uses Artificial Neural Network to construct the Learner Model, which automatically models relationship between different concepts in the curriculum and beats Knowledge Tracing in predicting student performance. We also propose a novel method for selecting items of optimal difficulty, personalized to student's skill level and learning rate, which decreases their learning time by 26.5% as compared to standard pre-defined curriculum sequence item selection policy.

In this paper, we introduce a new problem: automatically capturing programming content in online discussions. We expect solving this problem helps enhance visual presentation of programming forum content, qualitative analysis of forum contributions, and forum text preprocessing and normalization. We map this problem to a sequence learning problem and use Conditional Random Fields to solve it. We compare the performance with a word-feature based baseline and a nonsequence classification method (Naïve Bayes). The best results are produced by CRF method with an F1-Score as of 86.9%. Moreover, we demonstrate that the CRF classifier maintains a good accuracy across different domains; a model learned from a C++ forum performs almost as well on other programming language forums for Java and Python. As a demonstration of how captured information can be used, we provide an example of user profiling with programming content. In particular, we correlate the percentage of programming content in student answers to the student's course performance.

Social networking has gained immense traction in many areas, including teaching and learning. Networking sites for teachers aim to facilitate teacher communication and information sharing, but fall short of their potential. In order to support more effective use of online resources and better communication among teachers, we develop a suite of new user modeling and recommendation capabilities within a middle school teacher networking site. We foster collaboration among novice and experienced teachers when they share similar interests, enabling new mentoring relationships, and promote the use of relevant educational resources. We illustrate our approach with an implemented system called PedConnect that analyzes user activities and presents intelligent suggestions for collaboration and resource use.

Computational workflows are a powerful paradigm to represent and manage complex applications, particularly in large-scale distributed scientific data analysis. Workflows represent application components that result in individual computations as well as their interdependences in terms of dataflow. Workflow systems use these representations to manage various aspects of workflow creation and execution for users, such as the automatic assignment of execution resources. This article describes an approach to automating a new aspect of the process: the selection of application components and data sources. We present a novel approach that enables users to specify varying degrees of detail and amount of constraints in a workflow request, including the specification of constraints on input, intermediate or output data in the workflow, abstract workflow component classes rather than specific component implementations, and generic reusable workflow templates that express a pre-defined combination of components. The algorithm elaborates the user request into a set of fully ground workflows with specific choices of data sources and codes to be used so that they can be submitted for mapping and execution. The algorithm searches through the space of possible candidate workflows by creating increasingly more specialized versions of the original template and eliminating candidates that violate constraints cumulated in the candidate workflow as components and data sources are selected. A novel feature of our approach is that it assumes a distributed architecture where data and component catalogues are separate from the workflow system. The algorithm explicitly poses queries to external catalogues, and therefore any reasoning regarding data or component properties is not assumed to occur within the workflow system. We describe our implementation of this approach in the Wings workflow system. This implementation uses the W3C Web Ontology Language and associated reasoners to implement the workflow system as well as the data and component catalogues. This research demonstrates the use of artificial intelligence techniques to support the kinds of automation envisioned by the scientific community for large-scale distributed scientific data analysis.

Workflows, also known as process models, are essential in many science and engineering fields. Workflows express compositions of individual steps or tasks that assembled together account for various aspects of an overall process. When workflows include dozens of components and many links among them, the creation of valid workflows becomes challenging since users have to track many interdependencies and constraints. This article describes principles for assisting users to create valid workflows that are based on two knowledge acquisition systems that we have developed. A shared goal in these projects was to enable end users who do not have computer science backgrounds, such as biologists, military officers or engineers, to create valid end-to-end process models or workflows. Our approach exploits knowledge-rich descriptions of the individual components and their constraints in order to validate the composition, and uses artificial intelligence planning techniques in order to systematically verify formal properties of valid workflows. Both systems analyse partial workflows created by the user, determine whether they are consistent with the background knowledge that the system has, notifies the user of issues to be resolved in the current workflow, and suggests to the user what actions could be taken to correct those issues.

Automatic tools for analyzing student online discussions are highly desirable for providing better assistance and encouraging participation. This paper presents an approach for automatically identifying student discussions with unresolved issues or unanswered questions. We apply a two-phase classification algorithm. First, we classify “speech acts” of individual messages to identify the roles that the messages play, such as question, answer, issue raising, or acknowledgement. We then use the resulting speech acts as features for identifying discussion threads with unresolved issues or questions. We performed a preliminary analysis of the classifiers and achieved an average accuracy of 78%.

In this paper, we present a novel approach to scaffolding student participation and learning within discussion forums using student mentors, i.e., course peers with relatively good understanding of a particular domain topic. First, we identify mentors using domain topic models, student discussion profiles, and a similarity comparison of the topics being discussed. Second, we provide an interface that encourages classmates to invite mentors to participate. The feature, named MentorMatch, was integrated into an undergraduate course discussion board. Some results are reported.

Data analysis processes in scientific applications can be expressed as coarse-grain workflows of complex data processing operations with data flow dependencies between them. Performance optimization of these workflows can be viewed as a search for a set of optimal values in a multi-dimensional parameter space. While some performance parameters such as grouping of workflow components and their mapping to machines do not a ect the accuracy of the output, others may dictate trading the output quality of individual components (and of the whole workflow) for performance. This paper describes an integrated framework which is capable of supporting performance optimizations along multiple dimensions of the parameter space. Using two real-world applications in the spatial data analysis domain, we present an experimental evaluation of the proposed framework.

Workflows are becoming an increasingly more common paradigm to manage scientific analyses. As workflow repositories start to emerge, workflow retrieval and discovery becomes a challenge. Studies have shown that scientists wish to discover workflows given properties of workflow data inputs, intermediate data products, and data results. However, workflows typically lack this information when contributed to a repository. Our work addresses this issue by augmenting workflow descriptions with constraints derived from properties about the workflow components used to process data as well as the data itself. An important feature of our approach is that it assumes that component and data properties are obtained from catalogs that are external to the workflow system, consistent with current architectures for computational science.

This paper presents an approach for identifying student discussions with unresolved issues or unanswered questions. In order to handle highly incoherent data, we perform several data processing steps. We then apply a two-phase classification algorithm. First, we classify "speech acts" of individual messages to identify the roles that the messages play, such as question, issue raising, and answers. We then use the resulting speech acts as features for classifying discussion threads with unanswered questions or unresolved issues. We performed a preliminary analysis of the classifiers and the system shows an average F score of 0.76 in discussion thread classification.

This paper describes our experience to date employing the systematic mapping and optimization of large- scale scientific application workflows to current and future parallel platforms. The overall goal of the project is to integrate a set of system layers - application program, compiler, run-time environment, knowledge representation, optimization framework, and workflow manager - and through a systematic strategy for workflow mapping, our approach will exploit the vast machine resources available in such parallel platforms to dramatically increase the productivity of application programmers. In this paper, we describe the representation of a biomedical imaging application as a workflow, our early experiences in integrating the set of tools brought together for this project, and implications for future applications.

The International Workshop on Recommendation and Collaboration (ReColl 2008) aims to identify emerging trends in recommendation technology and collaborative environments in the context of intelligent user interfaces. We explore these two topics separately and the synergies between them.

Abstract Our research focuses on creating and executing large‐scale scientific workflows that often involve thousands of computations over distributed, shared resources. We describe an approach to workflow creation and refinement that uses semantic representations to (1) describe complex scientific applications in a data‐independent manner, (2) automatically generate workflows of computations for given data sets, and (3) map the workflows to available computing resources for efficient execution. Our approach is implemented in the Wings/Pegasus workflow system and has been demonstrated in a variety of scientific application domains. This paper illustrates the application‐level provenance information generated Wings during workflow creation and the refinement provenance by the Pegasus mapping system for execution over grid computing environments. We show how this information is used in answering the queries of the First Provenance Challenge. Copyright © 2007 John Wiley & Sons, Ltd.

Abstract The first Provenance Challenge was set up in order to provide a forum for the community to understand the capabilities of different provenance systems and the expressiveness of their provenance representations. To this end, a functional magnetic resonance imaging workflow was defined, which participants had to either simulate or run in order to produce some provenance representation, from which a set of identified queries had to be implemented and executed. Sixteen teams responded to the challenge, and submitted their inputs. In this paper, we present the challenge workflow and queries, and summarize the participants' contributions. Copyright © 2007 John Wiley & Sons, Ltd.

This paper describes a discussion-bot that provides answers to students' discussion board questions in an unobtrusive and human-like way. Using information retrieval and natural language processing techniques, the discussion-bot identifies the questioner's interest, mines suitable answers from an annotated corpus of 1236 archived threaded discussions and 279 course documents and chooses an appropriate response. A novel modeling approach was designed for the analysis of archived threaded discussions to facilitate answer extraction. We compare a self-out and an all-in evaluation of the mined answers. The results show that the discussion-bot can begin to meet students' learning requests. We discuss directions that might be taken to increase the effectiveness of the question matching and answer extraction algorithms. The research takes place in the context of an undergraduate computer science course.

Current knowledge acquisition tools have limited understanding of how users enter knowledge and how acquired knowledge is used, and provide limited assistance in organizing various knowledge authoring tasks. Users have to make up for these shortcomings by keeping track of past mistakes, current status, potential new problems, and possible courses of actions by themselves. In this paper, we present a novel extension to existing knowledge acquisition tools where the system organizes the episodes of past interactions through a set of declarative meta-level patterns and improves its suggestions based on relevant episodes. In particular, we focus on 1) assessing the level of confidence in suggesting an action, 2) suggesting how a knowledge authoring action can be done based on successful past actions, and 3) monitoring dynamic changes in the environment to suggest relevant modifications in the knowledge base. A preliminary study with varying synthetic user interactions shows that this meta-level assessment may reduce the number of incorrect suggestions, prevent some of the user mistakes and improve the overall problem solving results.

Complex applications in many areas, including scientific computations and business-related web services, are created from collections of components to form computational workflows. In many cases end users have requirements and preferences that depend on how the workflow unfolds, and that cannot be specified beforehand. Workflow editors enable users to formulate workflows, but the editors need to be augmented with intelligent assistance in order to help users in several key aspects of the task, namely: 1) keeping track of detailed constraints across selected components and their connections; 2) specifying the workflow flexibly, e.g., top-down, bottom-up, from requirements, or from available data; and 3) taking partial or incomplete descriptions of workflows and understanding the steps needed for their completion. We present an approach that combines knowledge bases (that have rich representations of components) together with planning techniques (that can track the relations and constraints among individual steps). We illustrate the approach with an implemented system called CAT (Composition Analysis Tool) that analyzes workflows and generates error messages and suggestions in order to help users compose complete and consistent workflows.

Interactive tools to help users author plans or processes are essential in a variety of domains. KANAL helps users author sound plans by simulating them, checking for a variety of errors and presenting the results in an accessible format that allows the user to see an overview of the plan steps or timelines of objects in the plan. From our experience in two domains, users tend to interleave plan authoring and plan checking while extending background knowledge of actions. This has led us to refine KANAL to provide a high-level overview of plans and integrate a tool for refining the background knowledge about actions used to check plans. We report on these lessons learned and new directions in KANAL.

Eliciting complex logical rules directly from logic-naive subject matter experts (SMEs) is a challenging knowledge capture task. We describe a large-scale experiment to evaluate tools designed to produce SME-authored rule bases. We assess the quality of the rule bases with respect to the: 1) performance on the addressed functional task (military course of action (COA) critiquing); and 2) intrinsic knowledge representation quality. In the course of this assessment, we note both strengths and weaknesses in the state of the art, and accordingly suggest some foci for future development in this important technology area.

This paper describes an integrated acquisition interface that includes several techniques previously developed to support users in various ways as they add new knowledge to an intelligent system. As a result of this integration, the individual techniques can take better advantage of the context in which they are invoked and provide stronger guidance to users. We describe the current implementation using examples from a travel planning domain, and demonstrate how users can add complex knowledge to the system.

Knowledge acquisition research concerned with the development of knowledge acquisition tools is in need of a methodological approach to evaluation. This paper describes experimental methodology to conduct studies and experiments of users modifying knowledge bases with knowledge acquisition tools. The paper also reports on the lessons learned from several experiments that have been performed using this methodology. The hope is that it will help others design user evaluations of knowledge acquisition tools. Ideas are discussed for improving the current methodology and some open issues that remain.

This paper describes a series of experiments with a range of users to evaluate an intelligent interface for acquiring problem-solving knowledge to describe how to accomplish a task. The tool derives the interdependencies between different pieces of knowledge in the system and uses them to guide the user in completing the acquisition task. The paper describes results obtained when the tool was tested with a wide range of users, including end users. The studies show that our acquisition interface saves users an average of 32% of the time it takes to add new knowledge, and highlight some interesting differences across user groups. The paper also describes what are the areas that need to be addressed in future research in order to make these tools usable by end users.