AI Research Research Articles

Categorical and Continuous Features in Counterfactual Explanations of AI Systems

Recently, eXplainable AI (XAI) research has focused on the use of counterfactual explanations to address interpretability, algorithmic recourse, and bias in AI system decision-making. The developers of these algorithms claim they meet user requirements in generating counterfactual explanations with “plausible”, “actionable” or “causally important” features. However, few of these claims have been tested in controlled psychological studies. Hence, we know very little about which aspects of counterfactual explanations really help users understand the decisions of AI systems. Nor do we know whether counterfactual explanations are an advance on more traditional causal explanations that have a longer history in AI (e.g., in expert systems). Accordingly, we carried out three user studies to (i) test a fundamental distinction in feature-types, between categorical and continuous features, and (ii) compare the relative effectiveness of counterfactual and causal explanations. The studies used a simulated, automated decision-making app that determined safe driving limits after drinking alcohol, based on predicted blood alcohol content, where users’ responses were measured objectively (using predictive accuracy) and subjectively (using satisfaction and trust judgments). Study 1 (N = 127) showed that users understand explanations referring to categorical features more readily than those referring to continuous features. It also discovered a dissociation between objective and subjective measures: counterfactual explanations elicited higher accuracy than no-explanation controls but elicited no more accuracy than causal explanations, yet counterfactual explanations elicited greater satisfaction and trust than causal explanations. In Study 2 (N = 136) we transformed the continuous features of presented items to be categorical (i.e., binary) and found that these converted features led to highly accurate responding. Study 3 (N = 211) explicitly compared matched items involving either mixed features (i.e., a mix of categorical and continuous features) or categorical features (i.e., categorical and categorically-transformed continuous features), and found that users were more accurate when categorically-transformed features were used instead of continuous ones. It also replicated the dissociation between objective and subjective effects of explanations. The findings delineate important boundary conditions for current and future counterfactual explanation methods in XAI.

Advancing LGBTQ+ inclusion in STEM education and AI research

The authors emphasize diversity, equity, and inclusion in STEM education and artificial intelligence (AI) research, focusing on LGBTQ+ representation. They discuss the challenges faced by queer scientists, educational resources, the implementation of National AI Campus, and the notion of intersectionality. The authors hope to ensure supportive and respectful engagement across all communities.

AI rising in higher education: opportunities, risks and limitations

PurposeThe aim of this paper is twofold: to explore the significance and implications of the rise of AI technology for the field of tertiary education in general and, in particular, to answer the question of whether teachers can be replaced by intelligent AI systems such as androids, what that requires in terms of human capabilities and what that might mean for teaching and learning in higher education.Design/methodology/approachGiven the interdisciplinary nature of this conceptual paper, a literature review serves as a methodological tool to access data pertaining to the research question posed in the paper.FindingsThis exploratory paper gathers a range of evidence from the philosophy of mind (the mind-body problem), Kahneman’s (2011) System 1 and System 2 models of the mind, Gödel’s (1951) Two Incompleteness Theorems, Polanyi’s (1958, 1966) theory of tacit knowing and Searle’s (1980) Chinese Room thought experiment to the effect that no AI system can ever fully replace a human being because no machine can replicate the human mind and its capacity for intelligence, consciousness and highly developed social skills such as empathy and cooperation.Practical implicationsAI is rising, but there are inherent limits to what machines can achieve when compared to human capabilities. An android can at most attain “weak AI”, that is, it can be smart but lack awareness or empathy. Therefore, an analysis of good teaching at the tertiary level shows that learning, knowledge and understanding go far beyond any quantitative processing that an AI machine does so well, helping us to appreciate the qualitative dimension of education and knowledge acquisition. ChatGPT is robotic, being AI-generated, but human beings thrive on the human-to-human interface – that is, human relationships and meaningful connections – and that is where the true qualitative value of educational attainment will be gauged.Social implicationsThis paper has provided evidence that human beings are irreplaceable due to our unique strengths as meaning-makers and relationship-builders, our capacity for morality and empathy, our creativity, our expertise and adaptability and our capacity to build unity and cooperate in building social structures and civilization for the benefit of all. Furthermore, as society is radically automated, the purpose of human life and its reevaluation will also come into question. For instance, as more and more occupations are replaced by ChatGPT services, more and more people will be freed up to do other things with their time, such as caring for relatives, undertaking creative projects, studying further and having children.Originality/valueThe investigation of the scope and limitations of AI is significant for two reasons. First, the question of the nature and functions of a mind becomes critical to the possibility of replication because if the human mind is like a super-sophisticated computer, then the relationship between a brain and mind is similar (if not identical) to the relationship between a computer as machine hardware and its programme or software (Dreyfus, 1979). [ ] If so, it should be theoretically possible to understand its mechanism and reproduce it, and then it is just a matter of time before AI research and development can replicate the human mind and eventually replace a human teacher, especially if an AI machine can teach just as intelligently yet more efficiently and economically. But if AI has inherent limitations that preclude the possibility of ever having a human-like mind and thought processes, then our investigation can at least clarify in what ways AI/AGI – such as ChatGPT – could support teaching and learning at universities.

How Does the Use of Artificial Intelligence Reflect on Business Administration and Management? A Perspective on Knowledge Production at the Postgraduate Level

Objectives: The extensive adoption of AI applications in many industries raises concerns over their potential to significantly transform employment in various aspects, including job creation, automation, and decision-making procedures. The impact that AI has on employment dynamics presents an array of obstacles and opportunities for the fields of business administration and management. Given all these advancements, the present study conducts a thorough examination of postgraduate theses, recognizing them as a systematic representation of the expanding influence of AI in the fields of business administration and management. Postgraduate research, which entails a thorough academic investigation, provides a crucial perspective for understanding the present condition and future course of AI use in business environments. Through an in-depth examination of postgraduate theses, this research attempts to reveal trends in the application of AI technologies in various organizational contexts. Design/methodology/approach: The 73 master's and doctoral theses used in this study were obtained from the repository of the Council of Higher Education National Thesis Center. The sample spans a range of institutions, demonstrating the diversity and depth of AI research across the academic landscape of management studies. To comprehensively examine and understand the information categories encompassed in these theses, the research utilizes two techniques: document analysis and descriptive content analysis. Document analysis systematically examines the theses as data sources, enabling a thorough investigation of the material that focuses on the identification, assessment, and synthesis of information pertaining to AI in business administration and management. Descriptive content analysis classifies and facilitates a methodical examination of the data. The distribution of theses subjected to document analysis was based on thesis type, universities and sub-disciplines, publication year and language, sample characteristics, methodology, theories, AI application areas, and keywords. Results: The findings show that studies on AI have gained momentum in the last four years, with a high percentage of theses focusing on management, organization, and marketing. The quantitative research method has been the most preferred for postgraduate theses. Additionally, human resource management, machine learning, and artificial neural networks constitute most of the research focus. The data indicates that most master's theses concentrate on human resource management and marketing. The finance and information technology sectors are predominant in terms of industry focus. Practical implications: The findings have the potential to significantly assist practitioners in understanding the ongoing research on AI, enabling them to align their strategic planning with the latest discoveries and approaches in higher education. The prevalence of machine learning and artificial neural networks signifies an inclination towards increasingly complex AI implementations within organizations. Organizations may leverage this understanding to increase innovation, enhance decision-making procedures, and sustain a competitive advantage through the implementation of cutting-edge AI. Furthermore, this study can assist researchers, such as master's and doctoral students, with the topic, research question, data source, data collection tool, and analysis type selection in future studies. Originality/value: The study is expected to provide insights into research focus, scope, and methodology for future research by revealing the current state of AI research in business administration and management domains in Turkey. Through the examination of postgraduate theses, this study not only highlights the increasing academic attention towards AI but also establishes a foundation upon which subsequent research can be built. This study represents an initial effort to gain deeper insights into AI research in Turkey, specifically focusing on knowledge production within universities.

Artificial intelligence and management education: A conceptualization of human-machine interaction

The increasing use of Advanced Natural Language Processing (ANLP) models, particularly ChatGPT-4, presents opportunities and challenges to management education and research. These models can enhance the style, creativity, and analytical power of research papers, potentially shifting human scholars' roles from creators to ‘prompters’. If machines can perform educational and research tasks more effectively the role of human educators becomes a salient question in a world in which ANLP models offer clear, coherent, and polished insights, the use of which has potentially paradoxical possibilities. From one perspective, a new type of high-quality scholarship and education characterized by strong human involvement that synergistically leverages ANLP models' analytical capabilities, enabling human scholars to probe complex phenomena and make management research truly meaningful and impactful for broader audiences, is possible. We explore these questions through an ‘ideal type’ conceptualization of the possible relations between AI and management education and research.

Exploring Trade-offs in Explainable AI

Machine Learning (ML) models are increasingly used in systems that involve physical human interaction or decision-making systems that impact human health and safety. Ensuring that these systems are safe and reliable is an important topic of current AI research. For many ML models it is unclear how a prediction (output) is arrived at from the provided features (input). Critical systems cannot blindly trust the predictions of such "black box" models, but instead need additional reassurance via insight into the model's reasoning. A range of methods exist within the field of Explainable AI (XAI) to make the reasoning of black box ML models more understandable and transparent. The explanations provided by XAI methods may be evaluated in a number of (competing) ways. In this paper, we investigate the trade-off between selected metrics for an XAI method called UnRAvEL, which is similar to the popular LIME approach. Our results show that by weighting the terms within the acquisition function used in UnRAvEL, different trade-offs can be achieved.

A Taxonomy of Explanation Types and Need Indicators in Human–Agent Collaborations

In recent years, explanations have become a pressing matter in AI research. This development was caused by the increased use of black-box models and a realization of the importance of trustworthy AI. In particular, explanations are necessary for human–agent interactions to ensure that the user can trust the agent and that collaborations are effective. Human–agent interactions are complex social scenarios involving a user, an autonomous agent, and an environment or task with its own distinct properties. Thus, such interactions require a wide variety of explanations, which are not covered by the methods of a single AI discipline, such as computer vision or natural language processing. In this paper, we map out what types of explanations are important for human–agent interactions, surveying the field via a scoping review. In addition to the typical introspective explanation tackled by explainability researchers, we look at assistive explanations, aiming to support the user with their task. Secondly, we survey what causes the need for an explanation in the first place. We identify a variety of human–agent interaction-specific causes and categorize them by whether they are centered on the agent’s behavior, the user’s mental state, or an external entity. Our overview aims to guide robotics practitioners in designing agents with more comprehensive explanation-related capacities, considering different explanation types and the concrete times when explanations should be given.

Responsible Artificial Intelligence for Climate Action: A Theoretical Framework for Sustainable Development

Climate change poses an urgent and significant challenge, with far-reaching impacts already affecting our planet, and projections indicating worsening conditions in the future. The concept of sustainable development aims to meet present needs while safeguarding the ability of future generations to meet their own requirements. However, climate change's effects on sustainable development are of paramount concern, as they amplify issues like poverty, food insecurity, and environmental degradation, affecting economic growth, social progress, and environmental protection. Taking immediate action to mitigate climate change and implement sustainable practices is crucial to ensuring a habitable planet for future generations. In this context, Responsible Artificial Intelligence (RAI) emerges as a promising direction, striving for ethical and responsible technology use in diverse sustainable development tasks. RAI proves to be a robust candidate for empowering climate change mitigation and adaptation efforts. This study introduces a theoretical RAI framework designed to support climate action by responsibly enabling more accurate predictions and analysis of climate data, enhancing energy efficiency, and reducing greenhouse gas emissions. The framework emphasizes the need for interdisciplinary collaboration among policymakers, scientists, and technicians to develop RAI solutions that advance sustainable development and alleviate the adverse impacts of climate change. Unlike previous works, this research presents a novel perspective on the principles of RAI that explicitly consider climate-related aspects. By laying the foundations of AI research to bolster our fight against climate change, this article establishes essential pillars that encourage further advancements in this critical endeavor.

Artificial Intelligence in MOBA Games: A Multivocal Literature Mapping

Esports - games played competitively - comprise a major sector of the global games industry. Esports has been used as a testbed for game AI and game analytics for two decades. This article presents a multivocal literature mapping of available research that focuses strictly on the use of artificial intelligence approaches in Multiplayer Online Battle Arena (MOBA) games, one of the most popular esports genres and the one most widely used for game AI and game analytics research. A mapping is performed on relevant publications published between 2011 and 2022 and systematically examines them to extract similarities, gaps, and main findings. We analyzed 124 publications to identify the most studied topics, the most commonly used techniques, and the most commonly applied evaluation methods. The results show that League of Legends and DOTA are the most studied games, with outcome prediction being the most popular research topic. Finally, we provide an analysis of the potential future flagship areas for research in the domain, considering the gaps found in the white and grey literature.

A Collaborative Code Platform with Advanced AI Features and Real-Time Collaboration Tools

Abstract: Good collaboration and code management are crucial in today's software development environment. Our research has launched a new collaborative process that combines AI resources, including AI and global research, to improve products and enhance and facilitate collaboration. The platform features real-time collaboration tools such as dynamic text with multilanguage support, dialog functionality, and notepad functionality. With OAuth authentication, users can seamlessly perform CRUD operations on files and directories, share code repositories, and participate in live coding sessions. The platform supports compilation and execution of a code file and follows good design principles such as caching, microservices architecture, and containerization to ensure scalability and performance. The platform focuses on user experience by providing simple user interface design and easy interface navigation. Users can create workspaces, share them collaboratively, and seamlessly import repositories from Git. The platform's intelligent data manager helps manage files and directories, while the collaborative platform allows developers to collaborate on code libraries efficiently. The platform is modern web-based, using React on the frontend and Node.js on the backend; Supported with syntax highlighting and smart initialization along with Socket.io for runtime with SQL and NoSQL databases. Provides comprehensive language support. All progress is saved, providing users with peace of mind. While the basic design is for desktops, future iterations are expected to support mobile devices.

A practical guide to the implementation of AIin orthopaedic research, Part 6: How to evaluate the performance of AI research?

Level V.

Neural Circuit Policies for Virtual Character Control

The development of high-stakes decision-making neural agents that interact with complex environments, such as video games, is an important aspect of AI research with numerous potential applications. Reinforcement learning combined with deep learning architectures (DRL) has shown remarkable success in various genres of games. The performance of DRL is heavily dependent upon the neural networks resides within them. Although these algorithms perform well in offline testing but the performance deteriorates in noisy and sub-optimal conditions, creating safety and security issues. To address these, we propose a hybrid deep learning architecture that combines a traditional convolutional neural network with worm brain-inspired neural circuit policies. This allows the agent to learn key coherent features from the environment and interpret its dynamics. The obtained DRL agent was not only able to achieve an optimal policy quickly, but it was also the most noise-resilient with the highest success rate. Our research indicates that only 20 control neurons (12 inter-neurons and 8 command neurons) are sufficient to achieve competitive results. We implemented and analyzed the agent in the popular video game Doom, demonstrating its effectiveness in practical applications.

Enhancing AI Research for Breast Cancer: A Comprehensive Review of Tumor-Infiltrating Lymphocyte Datasets.

The field of immunology is fundamental to our understanding of the intricate dynamics of the tumor microenvironment. In particular, tumor-infiltrating lymphocyte (TIL) assessment emerges as essential aspect in breast cancer cases. To gain comprehensive insights, the quantification of TILs through computer-assisted pathology (CAP) tools has become a prominent approach, employing advanced artificial intelligence models based on deep learning techniques. The successful recognition of TILs requires the models to be trained, a process that demands access to annotated datasets. Unfortunately, this task is hampered not only by the scarcity of such datasets, but also by the time-consuming nature of the annotation phase required to create them. Our review endeavors to examine publicly accessible datasets pertaining to the TIL domain and thereby become a valuable resource for the TIL community. The overall aim of the present review is thus to make it easier to train and validate current and upcoming CAP tools for TIL assessment by inspecting and evaluating existing publicly available online datasets.

REAL-Colon: A dataset for developing real-world AI applications in colonoscopy

Detection and diagnosis of colon polyps are key to preventing colorectal cancer. Recent evidence suggests that AI-based computer-aided detection (CADe) and computer-aided diagnosis (CADx) systems can enhance endoscopists' performance and boost colonoscopy effectiveness. However, most available public datasets primarily consist of still images or video clips, often at a down-sampled resolution, and do not accurately represent real-world colonoscopy procedures. We introduce the REAL-Colon (Real-world multi-center Endoscopy Annotated video Library) dataset: a compilation of 2.7 M native video frames from sixty full-resolution, real-world colonoscopy recordings across multiple centers. The dataset contains 350k bounding-box annotations, each created under the supervision of expert gastroenterologists. Comprehensive patient clinical data, colonoscopy acquisition information, and polyp histopathological information are also included in each video. With its unprecedented size, quality, and heterogeneity, the REAL-Colon dataset is a unique resource for researchers and developers aiming to advance AI research in colonoscopy. Its openness and transparency facilitate rigorous and reproducible research, fostering the development and benchmarking of more accurate and reliable colonoscopy-related algorithms and models.

Cell Maps for Artificial Intelligence: AI-Ready Maps of Human Cell Architecture from Disease-Relevant Cell Lines.

This article describes the Cell Maps for Artificial Intelligence (CM4AI) project and its goals, methods, standards, current datasets, software tools , status, and future directions. CM4AI is the Functional Genomics Data Generation Project in the U.S. National Institute of Health's (NIH) Bridge2AI program. Its overarching mission is to produce ethical, AI-ready datasets of cell architecture, inferred from multimodal data collected for human cell lines, to enable transformative biomedical AI research.

Now, Later, and Lasting: 10 Priorities for AI Research, Policy, and Practice

Shaping the future of artificial intelligence.

The Impacts of Text-to-Image Generative AI on Creative Professionals According to Prospective Generative AI Researchers: Insights from Japan

The growing interest in Japan to implement text-to-image (T2I) generative artificial intelligence (GenAI) technologies in creative workflows has raised concern over what ethical and social implications these technologies will have on creative professionals. Our pilot study is the first to discuss what social and ethical oversights may emerge regarding such issues from prospective Japanese researchers – computer science (CS) graduate students studying in Japan. Given that these students are the primary demographic hired to work at research and development (R&D) labs at the forefront of such innovations in Japan, any social and ethical oversight on such issues may unequip them as future knowledge experts who will play a pivotal role in helping shape Japan’s policies regarding image generating AI technologies.

The extimate core of understanding: absolute metaphors, psychosis and large language models

AbstractThis paper delves into the striking parallels between the linguistic patterns of Large Language Models (LLMs) and the concepts of psychosis in Lacanian psychoanalysis. Lacanian theory, with its focus on the formal and logical underpinnings of psychosis, provides a compelling lens to juxtapose human cognition and AI mechanisms. LLMs, such as GPT-4, appear to replicate the intricate metaphorical and metonymical frameworks inherent in human language. Although grounded in mathematical logic and probabilistic analysis, the outputs of LLMs echo the nuanced linguistic associations found in metaphor and metonymy, suggesting a mirroring of human linguistic structures. A pivotal point in this discourse is the exploration of “absolute metaphors”—core gaps in reasoning discernible in both AI models and human thought processes and central to the Lacanian conceptualization of psychosis. Despite the traditional divide between AI research and continental philosophy, this analysis embarks on an innovative journey, utilizing Lacanian philosophy to unravel the logic of AI, using concepts established in the continental discourse on logic, rather than the analytical tradition.

Revisiting Aristotle’s Master-Slave Relationship: A Casual Evaluation in the Context of Human-AI Dynamics

This paper sought to explore Aristotle's handling of the notion of natural slavery and its relevance in ongoing discussions concerning the ethics of artificial intelligence. Using an extensive literature review covering his discussions on slavery in the Politics and the Nicomachean Ethics, this study delved into the philosophical underpinnings of slavery within Aristotle's political theory and examines its ramifications within the domain of ethical technology. Consequently, while Aristotle's theory of natural slavery suggests that certain individuals are predisposed to service due to deficiencies in rational capacity, this concept was analysed in the context of modern AI technology, raising ethical concerns about autonomy, control and social hierarchy in human AI interactions. By relating Aristotle's philosophical framework to contemporary AI ethics, this study offered an understanding of the ethical implications of advancing technology and encouraged interdisciplinary dialogue between philosophy, ethics and AI research.

2760: Data-centric AI and cancer research: constructing a research data access pipeline using XNAT

2760: Data-centric AI and cancer research: constructing a research data access pipeline using XNAT