AHFE Open AccessArtificial Intelligence and Social Computing
Editors: Tareq Ahram, Jay Kalra, Waldemar Karwowski
Topics: Artificial Intelligence & Computing
Publication Date: 2023
ISBN: 978-1-958651-48-3
DOI: 10.54941/ahfe1003267
Articles
A machine learning approach for optimizing waiting times in a hand surgery operation center
For patients scheduled for surgery, long waiting times are unpleasant. However, scheduling that is too patient-oriented can lead to friction losses in the operating room and waiting times for the medical personnel. We have conducted an analysis of historical hand surgery data to improve forecasting of hand surgery durations, optimize operation room scheduling for physicians and patients and reduce overall waiting times. Several models have been evaluated to forecast surgery durations. A quantile-based approach based on the distribution of surgery durations has been tested in a scheduling simulation. This approach has indicated possibilities to gradually balance waiting times between patients and medical personnel. Within a field trial, a trained regression model has been successfully deployed in a hand surgery operation center.
Andreas Schuller, Marc Braun, Peter Hahn
Open Access
Article
Conference Proceedings
Automated Decision Support for Collaborative, Interactive Classification
Traditional classification approaches are straightforward: collect data, apply classification algorithms, then generate classification results. However, such approaches depend on data being amply available, which is not always the case. This paper describes an approach to maximize the utility of collected data through intelligent guidance of the data collection process. We present the development and evaluation of a knowledge-based decision-support system: the Logical Reasoner (LR), which guides data collection by unmanned ground and air assets to improve behavior classification. The LR is a component of a Human Directed and Controlled AI system (or “Human-AI” system) aimed at semi-autonomous classification of potential threat and non-threat individuals in a complex urban setting. The setting provides little to no pre-existing data; thus, the system collects, analyzes, and evaluates real-time human behavior data to determine whether the observed behavior is indicative of threat intent. The LR’s purpose is to produce contextual knowledge to help make productive decisions about where, when, and how to guide the vehicles in the data collection process. It builds a situational-awareness picture from the observed spatial relationships, activities, and interim classifications, then uses heuristics to generate new information-gathering goals, as well as to recommend which actions the vehicles should take to better achieve these goals. The system uses these recommendations to collaboratively help the operator direct the autonomous assets to individuals or places in the environment to maximize the effectiveness of evidence collection. LR is based on the Soar Cognitive Architecture which excels in supporting Human-AI collaboration. The described DoD-sponsored system has been developed and extensively tested for over three years, in simulation and in the field (with role-players). Results of these experiments have demonstrated that the LR decision support contributes to automated data collection and overall classification accuracy by the Human-AI team. This paper describes the development and evaluation of the LR based on multiple test events.The research reported in this document was performed under Defense Advanced Research Projects Agency (DARPA) contract #HR001120C0180, Urban Reconnaissance through Supervised Autonomy (URSA). The views, opinions, and/or findings expressed are those of the author(s) and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon. Many thanks to Robert Marinier and Kris Kearns for their assistance in the preparation of this manuscript, as well as the entire ISOLATE R&D team.Distribution Statement “A” (Approved for Public Release, Distribution Unlimited)
Randolph Jones, Robert Bixler, Robert Marinier, Lilia Moshkina
Open Access
Article
Conference Proceedings
Dynamically monitoring crowd-worker's reliability with interval-valued labels
Crowdsourcing has rapidly become a computing paradigm in machine learning and artificial intelligence. In crowdsourcing, multiple labels are collected from crowd-workers on an instance usually through the Internet. These labels are then aggregated as a single label to match the ground truth of the instance. Due to its open nature, human workers in crowdsourcing usually come with various levels of knowledge and socio-economic backgrounds. Effectively handling such human factors has been a focus in the study and applications of crowdsourcing. For example, Bi et al studied the impacts of worker's dedication, expertise, judgment, and task difficulty (Bi et al 2014). Qiu et al offered methods for selecting workers based on behavior prediction (Qiu et al 2016). Barbosa and Chen suggested rehumanizing crowdsourcing to deal with human biases (Barbosa 2019). Checco et al studied adversarial attacks on crowdsourcing for quality control (Checco et al 2020). There are many more related works available in literature. In contrast to commonly used binary-valued labels, interval-valued labels (IVLs) have been introduced very recently (Hu et al 2021). Applying statistical and probabilistic properties of interval-valued datasets, Spurling et al quantitatively defined worker's reliability in four measures: correctness, confidence, stability, and predictability (Spurling et al 2021). Calculating these measures, except correctness, does not require the ground truth of each instance but only worker’s IVLs. Applying these quantified reliability measures, people have significantly improved the overall quality of crowdsourcing (Spurling et al 2022). However, in real world applications, the reliability of a worker may vary from time to time rather than a constant. It is necessary to monitor worker’s reliability dynamically. Because a worker j labels instances sequentially, we treat j’s IVLs as an interval-valued time series in our approach. Assuming j’s reliability relies on the IVLs within a time window only, we calculate j’s reliability measures with the IVLs in the current time window. Moving the time window forward with our proposed practical strategies, we can monitor j’s reliability dynamically. Furthermore, the four reliability measures derived from IVLs are time varying too. With regression analysis, we can separate each reliability measure as an explainable trend and possible errors. To validate our approaches, we use four real world benchmark datasets in our computational experiments. Here are the main findings. The reliability weighted interval majority voting (WIMV) and weighted preferred matching probability (WPMP) schemes consistently overperform the base schemes in terms of much higher accuracy, precision, recall, and F1-score. Note: the base schemes are majority voting (MV), interval majority voting (IMV), and preferred matching probability (PMP). Through monitoring worker’s reliability, our computational experiments have successfully identified possible attackers. By removing identified attackers, we have ensured the quality. We have also examined the impact of window size selection. It is necessary to monitor worker’s reliability dynamically, and our computational results evident the potential success of our approaches.This work is partially supported by the US National Science Foundation through the grant award NSF/OIA-1946391.
Chenyi Hu, Makenzie Spurling
Open Access
Article
Conference Proceedings
Perceptions, attitudes and trust toward artificial intelligence — An assessment of the public opinion
Over the last couple of years, artificial intelligence (AI)—namely machine learning algorithms—has rapidly entered our daily lives. Applications can be found in medicine, law, finance, production, education, mobility, and entertainment. To achieve this, a large amount of research has been undertaken, to optimize algorithms that by learning from data are able to process natural language, recognize objects through computer vision, interact with their environment with the help of robotics, or take autonomous decisions without the help of human input. With that, AI is acquiring core human capabilities raising the question of the impact of AI use on our society and its individuals. To form a basis for addressing those questions, it is crucial to investigate the public perception of artificial intelligence. This area of research is however often overlooked as with the fast development of AI technologies demands and wishes of individuals are often neglected. To counteract this, our study focuses on the public's perception, attitudes, and trust towards artificial intelligence. To that end, we followed a two-step research approach. We first conducted semi-structured interviews which laid the foundation for an online questionnaire. Building upon the interviews, we designed an online questionnaire (N=124) in which in addition to user diversity factors such as belief in a dangerous world, sensitivity to threat, and technology optimism, we asked respondents to rate prejudices, myths, risks, and chances about AI. Our results show that in general respondents view AI as a tool that can act independently, adapt, and help them in their daily lives. With that being said, respondents also indicate that they are not able to understand the underlying mechanisms of AI, and with this doubt, the maturity of the technology, leading to privacy concerns, fear of misuse, and security issues. While respondents are willing to use AI nevertheless, they are less willing to place their trust in the technology. From a user diversity point of view, we found, that both trust and use intention are correlated to the belief in a dangerous world and technology optimism. In summary, our research shows that while respondents are willing to use AI in their everyday lives, still some concerns remain that can impact their trust in the technology. Further research should explore the mediation of concerns to include them in a responsible development process that ensures a positive impact of AI on individuals' lives and our society.
Gian Luca Liehner, Alexander Hick, Hannah Biermann, Philipp Brauner, Martina Ziefle
Open Access
Article
Conference Proceedings
Artificial Empathy: Exploring the Intersection of Digital Art and Emotional Responses to the COVID-19 Pandemic
The COVID-19 pandemic has caused widespread emotional and psychological impacts globally, leading to feelings of isolation, separation, and disconnection among individuals. In response to this, the present study seeks to explore and document the emotional experience of the COVID-19 pandemic through the creation of an art project titled "All those days in isolation". Using mixed media and collage techniques, the study seeks to create a visual representation of the collective experiences and emotions of a community during the pandemic. This project was inspired by the feelings of isolation and separation that many people have experienced and sought to explore and express these emotions through art. Through a comprehensive review of existing literature and qualitative research, including semi-structured interviews with a group of participants who have experienced the pandemic, this thesis will examine how digital art has been used to record and express emotions.The study found that the COVID-19 pandemic has had a significant impact on mental health and well-being, with high levels of anxiety, stress, and depression reported among individuals who have been directly or indirectly affected by the virus. Additionally, the pandemic has been associated with feelings of loneliness and social isolation, as well as with an increased risk of domestic violence and other forms of abuse.The art project was successful in exploring and expressing the complex emotions of the COVID-19 pandemic, offering a nuanced and well-rounded perspective on the emotional impact of the pandemic on individuals. The study highlights the importance of art in documenting and preserving collective experiences and emotions, as well as its potential to serve as a reflection of society and a tool for coping with stress and traumatic events. Overall, the art project demonstrates the power of art in exploring and expressing complex emotions and providing a space for people to connect with and understand the experiences of others.
Mingzhu Li, Ming Zhong
Open Access
Article
Conference Proceedings
Machine Reading Comprehension and Expert System technologies for social innovation in the drug excipient selection process
The growth of the global population together with several unpredicted crises such as political, health, and financial, create an environment of uncertainty in which social innovations can be developed to offer stability in people’s lives and create new business development opportunities for the benefit of the economy and the society. One of the undoubted rights of every human being is access to affordable medical treatment. However, the costs and time needed for research and development on new or specialized drugs are not often covered by governmental budgets and initiatives that could make such medicines accessible to all who needed them. Private companies invest tremendous amounts and expect returns on their investments. This gap, between the availability of a drug and its accessibility, created the social need for a generic drug market and the inspiration for advanced innovations to serve it. Research indicates that the price of brand-name drugs can drop up to 80% after the commercialization of a new generic which has the same action and can potentially replace them. The global generic drug market worth is expected to increase from $311.8 billion in 2021 to $442.3 billion in 2026. Excipients represent a market value of $4 billion, accounting for 0.5% of the total pharmaceutical market. The global market of AI was estimated at 43.1 billion in 2020 and is predicted to reach $228.3 billion by 2026 with a 32.7 % CAGR. On the other hand, the revenues of the AI Health market are projected to grow from $6.9 billion in 2021 to $67.4 billion in 2027 reaching $120.2 billion by 2028 with a CAGR of 45.3%.The choice of excipients in drug development is a critical and time-consuming process. Currently, excipients are chosen based on the route of administration, physicochemical characteristics, place of action, and the type of release of the active ingredient. The process involves many quality control tests on the drug such as fragility, dissolution, disintegration, dosage uniformity, and stability, which are repeated when the excipient changes. This laborious and time-consuming process considers a massive number of existing excipients categorized into different functional groups used for different purposes.This paper addresses this challenge and introduces an approach to resolve it using Artificial Intelligence for social innovation in the formulation development industry. Specifically, the paper presents an Expert system (ES) based software architecture to facilitate assess and utilize drug-excipient relationship data scattered in various forms of documentation and/or scientific literature. The inference engine of the ES operates with rule base and case-based reasoning powered by Machine Reading Comprehension (MRC) and Natural Language Processing (NLP) technologies that populate and enrich the knowledge base. The MRC and NLP technologies interpret existing drug formulations and propose potential new drug formulations, based on its physicochemical characteristics.According to research results, the time to introduce a generic drug can be reduced by 30% if there is an indicative formulation to start the process. The eight months gained can be used to market the product. This is a significant amount of time that reduces research and development costs, reduces the time to market, and increases productivity and operations efficiency. The research conducted is based on an extensive literature review, primary research with surveys and interviews but also with the analysis of several case studies to indicate the need for the proposed technology and support the system architecture design. Furthermore, the paper presents the pre and post-condition for adopting such technology, highlights research limitations, and identifies areas of further research to be conducted for the optimization of the technology and its contribution to the global economy and society.
Evangelos Markopoulos, Chrystalla Protopapa
Open Access
Article
Conference Proceedings
Image Caption Generation of Arts: Review and Outlook
Image captioning extract image features and automatically describe the content of an image in words. Recently image captioning has broken through the application of natural images and is widely used in the arts. It can be applied to art retrieval and management, and it can also automatically provide artistic introductions for the visually impaired. This paper reviews related research in image captioning of artworks, and divides image captioning into three types, including template-based approach, retrieval-based approach, and generative approach. Furthermore, mainstream generative approaches include Encoder-decoder, Transformer, New generation framework, etc. Finally, this paper summarizes the evaluation metrics for image captioning, and looks forward to the application and future development of art image captioning.
Baoying Zheng, Fang Liu
Open Access
Article
Conference Proceedings
Automated Visual Story Synthesis with Character Trait Control
Visual storytelling is an art form that has been utilized for centuries to communicate stories, convey messages, and evoke emotions. The images and text must be used in harmony to create a compelling narrative experience. With the rise of text-to-image generation models such as Stable Diffusion, it is becoming more promising to investigate methods of automatically creating illustrations for stories. However, these diffusion models are usually developed to generate a single image, resulting in a lack of consistency be- tween figures and objects across different illustrations of the same story, which is especially important in stories with human characters.This work introduces a novel technique for creating consistent human figures in visual stories. This is achieved in two steps. The first step is to collect human portraits with various identifying characteristics, such as gender and age, that describe the character. The second step is to use this collection to train DreamBooth to generate a unique token ID for each type of character. These IDs can then be used to replace the names of the story characters in the image-generation process. By combining these two steps, we can create controlled human figures for various visual storytelling contexts.
Yuetian Chen, Bowen Shi, Peiru Liu, Ruohua Li, Mei Si
Open Access
Article
Conference Proceedings
Towards a Proper Evaluation of Automated Conversational Systems
Efficient evaluation of dialogue agents is a major problem in conversational AI, with current research still relying largely on human studies for method validation. Recently, there has been a trend toward the use of automatic self-play and bot-bot evaluation as an approximation for human ratings of conversational systems. Such methods promise to alleviate the time and financial costs associated with human evaluation, and current proposed methods show moderate to strong correlation with human judgements. In this study, we further investigate the fitness of end-to-end self-play and bot-bot interaction for dialogue system evaluation. Specifically, we perform a human study to confirm self-play evaluations of a recently proposed agent that implements a GPT-2 based response generator on the Persuasion For Good charity solicitation task. This agent leverages Progression Function (PF) models to predict the evolving acceptability of an ongoing dialogue and uses dialogue rollouts to proactively simulate how candidate responses may impact the future success of the conversation. The agent was evaluated in an automatic self-play setting, using automatic metrics to estimate sentiment and intent to donate in each simulated dialogue. This evaluation indicated that sentiment and intent to donate were higher (p < 0.05) across dialogues involving the progression-aware agents with rollouts, compared to a baseline agent with no rollout-based planning mechanism. To validate the use of self-play in this setting, we follow up by conducting a human evaluation of this same agent on a range of factors including convincingness, aggression, competence, confidence, friendliness, and task utility on the same Persuasion For Good solicitation task. Results show that human users agree with previously reported automatic self-play results with respect to agent sentiment, specifically showing improvement in friendliness and confidence in the experimental condition; however, we also discover that for the same agent, humans reported a lower desire to use it in the future compared to the baseline. We perform a qualitative sentiment analysis of participant feedback to explore possible reasons for this, and discuss implications for self-play and bot-bot interaction as a general framework for evaluating conversational systems.
Abraham Sanders, Mara Schwartz, Albert Chang, Shannon Briggs, Jonas Braasch, Dakuo Wang, Mei Si, Tomek Strzalkowski
Open Access
Article
Conference Proceedings
Does Imageable Language Make Your Tweets More Persuasive?
Imageability is a psycholinguistic property of words that indicates how quickly and easily a word evokes a mental image or other sensory experience. Highly imageable words are easier to read and comprehend, and, as a result, their use in communications, such as social media, makes messages more memorable, and, potentially, more impactful and influential. In this paper, we explore the relationship between the imageability of messages in social media and their influence on the target audience. We focus on messages surrounding important public events and approximate the influence of a message by the number of retweets the message receives. First, we propose novel ways to determine an imageability score for a text, utilizing combinations of word-level imageability scores from the MRCPD+ lexicon, as well as word embeddings, image caption data, and word frequency data. Next, we compare these new imageability score functions to a variety of simple baseline functions in correlation between tweet imageability and number of retweets in the domain of the 2017 French Presidential Elections. We find that the imageability score of messages is correlated with the number of retweets in general, and also when normalized for topic and novelty; thus, imageable language is potentially more influential. We consider grouping tweets into imageability score ranges, and find that tweets within higher ranges of imageability scores receive more retweets on average compared to tweets within lower ranges. Lastly, we manually annotate a small number of tweets for imageability and show that our imageability score functions agree well with the human annotators when the agreement between human raters is high.
Andy Bernhardt, Tomek Strzalkowski, Ning Sa, Ankita Bhaumik, Gregorios Katsios
Open Access
Article
Conference Proceedings
Emotional Analysis of Candidates During Online Interviews
The recent empirical findings from the related fields including psychology, behavioral sciences, and neuroscience indicate that both emotion and cognition are influential during the decision making processes and so on the final behavioral outcome. On the other hand, emotions are mostly reflected by facial expressions that could be accepted as a vital means of communication and critical for social cognition. This has been known as the facial activation coding in the related academic literature. There have been several different AI-based systems that produce analysis of facial expressions with respect to 7 basic emotions including happy, sad, angry, disgust, fear, surprise, and neutral through the photos captured by camera-based systems. The system we have designed is composed of the following stages: (1) face verification, (2) facial emotion analysis and reporting, (3) emotion recognition from speech. The users upload their online video in which the participants tell about themselves within 3 minutes duration. In this study, several classification methods were applied for model development processes, and the candidates' emotional analysis in online interviews was focused on, and inferences about the situation were attempted using the related face images and sounds. In terms of the face verification system obtained as a result of the model used, 98% success was achieved. The main target of this paper is related to the analysis of facial expressions. The distances between facial landmarks are made up of the starting and ending points of these points. 'Face frames' were obtained while the study was being conducted by extracting human faces from the video using the VideoCapture and Haar Cascade functions in the OpenCV library in the Python programming language with the image taken in the recorded video. The videos consist of 24 frames for 1000 milliseconds. During the whole video, the participant's emotion analysis with respect to facial expressions is provided for the durations of 500 milliseconds. Since there are more than one face in the video, face verification was done with the help of different algorithms: VGG-Face, Facenet, OpenFace, DeepFace, DeepID, Dlib and ArcFace. Emotion analysis via facial landmarks was performed on all photographs of the participant during the interview. DeepFace algorithm was used to analyze face frames through study that recognizes faces using convolutional neural networks, then analyzes age, gender, race, and emotions. The study classified emotions as basic emotions. Emotion analysis was performed on all of the photographs obtained as a result of the verification, and the average mood analysis was carried out throughout the interview, and the data with the highest values on the basis of emotion were also recorded and the probability values have been extracted for further analyses. Besides the local analyses, there have also been global outputs with respect to the whole video session. The main target has been to introduce different potential features to the feature matrix that could be correlated with the other variables and labels tagged by the HR expert.
Alperen Sayar, Tuna Çakar, Tunahan Bozkan, Seyit Ertuğrul, Mert Güvençli
Open Access
Article
Conference Proceedings
Emotion Recognition from Speech via the Use of Different Audio Features, Machine Learning and Deep Learning Algorithms
Speech has been accepted as one of the basic, efficient and powerful communication methods. At the beginning of the 20th century, electroacoustic analysis was used for determining emotions in psychology. In academics, Speech Emotion Recognition (SER) has become one of the most studied and investigated research areas. This research program aims to determine the emotional state of the speaker based on speech signals. Significant studies have been undertaken during the last two decades to identify emotions from speech by using machine learning. However, it is still a challenging task because emotions rotate from one to another and there are environmental factors which have significant effects on emotions. Furthermore, sound consists of numerous parameters and there are various anatomical characteristics to take into consideration. Determining an appropriate audio feature set for emotion recognition is still a critical decision point for an emotion recognition system. The demand for voice technology in both art and human – machine interaction systems has recently been increased. Our voice conveys both linguistic and paralinguistic messages in the course of speaking. The paralinguistic part, for example, rhythm and pitch, provides emotional cues to the speaker. The speech emotion recognition topic examines the question ‘How is it said?’ and an algorithm detects the emotional state of the speaker from an audio record. Although a considerable number of the studies have been conducted for selecting and extracting an optimal set of features, appropriate attributes for automatic emotion recognition from audio are still under research. The main aim of this study is obtaining the most distinctive emotional audio features. For this purpose, time- based features, frequency-based features and spectral shape-based features are used for comparing recognition accuracies. Besides these features, a pre-trained model is used for obtaining input for emotion recognition. Machine learning models are developed for classifying emotions with Support Vector Machine, Multi-Layer Perceptron and Convolutional Neural Network algorithms. Three emotional databases in English and German are combined and a larger database is obtained for training and testing the models. Emotions namely, Happy, Calm, Angry, Boredom, Disgust, Fear, Neutral, Sad and Surprised are classified with these models. When the classification results are examined, it is concluded that the pre- trained representations make the most successful predictions. The weighted accuracy ratio is 91% for both Convolutional Neural Network and Multilayer Perceptron algorithms while this ratio is 87% for the Support Vector Machine algorithm. A hybrid model is being developed which contains both a pre-trained model and spectral shaped based features. Speech contains silent and noisy sections which increase the computational complexity. Time performance is the other major factor which should be a great deal of careful consideration. Although there are many advancements on SER, custom architectures are designed to fuse accuracy and time performance. Even further for a more realistic emotion estimation all physical gestures like voice, body parts of movement and facial expression can be obtained together as humans use them collectively to express themselves.
Alperen Sayar, Tuna Çakar, Tunahan Bozkan, Seyit Ertuğrul, Fatma Gümüş
Open Access
Article
Conference Proceedings
Evaluating the Effect of Time on Trust Calibration of Explainable Artificial Intelligence
Explainable Artificial Intelligence (XAI) has played a significant role in human-computer interaction. The cognitive resources it carries allow humans to understand the complex algorithm powering Artificial Intelligence (AI), virtually resolving the acceptance and adoption barrier from the lack of transparency. This resulted in more systems leveraging XAI and triggering interest and efforts to develop newer and more capable techniques. However, though the research stream is expanding, little is known about the extent of its effectiveness on end-users. Current works have only measured XAI effects on either moment time effect or compared it cross-sectionally on various types of users. Filling this out can improve the understanding of existing studies and provide practical limitations on its use for trust calibration. To address this gap, a multi-time research experiment was conducted with 103 participants to use and evaluate XAI in an image classification application for three days. Measurement that was considered is on perceived usefulness for its cognitive contribution, integral emotions for affective change, trust, and reliance, and was analyzed via covariance-based structural equation modelling. Results showed that time only moderates the path from cognitive to trust and reliance as well as trust to reliance, with its effect dampening through time. On the other hand, affective change has remained consistent in all interactions. This shows that if an AI system uses XAI over a longer time frame, prioritization should be on its affective properties (i.e., things that will trigger emotional change) rather than purely on its cognitive purpose to maximize the positive effect of XAI.
Ezekiel Bernardo, Rosemary Seva
Open Access
Article
Conference Proceedings
Towards Kenyan Sign Language Hand Gesture Recognition Dataset
Datasets for hand gesture recognition are now an important aspect of machine learning. Many datasets have been created for machine learning purposes. Some of the notable datasets include Modified National Institute of Standards and Technology (MNIST) dataset, Common Objects in Context (COCO) dataset, Canadian Institute For Advanced Research (CIFAR-10) dataset, LeNet-5, AlexNet, GoogLeNet, The American Sign Language Lexicon Video Dataset and 2D Static Hand Gesture Colour Image Dataset for ASL Gestures. However, there is no dataset for Kenya Sign language (KSL). This paper proposes the creation of a KSL hand gesture recognition dataset. The dataset is intended to be in two-fold. One for static hand gestures, and one for dynamic hand gestures. With respect to dynamic hand gestures short videos of the KSL alphabet a to z and numbers 0 to 10 will be considered. Likewise, for the static gestures KSL alphabet a to z will be considered. It is anticipated that this dataset will be vital in creation of sign language hand gesture recognition systems not only for Kenya sign language but of other sign languages as well. This will be possible because of learning transfer ability when implementing sign language systems using neural network models.
Casam Nyaga, Ruth Wario
Open Access
Article
Conference Proceedings
The evolution of artificial intelligence adoption in industry
Artificial intelligence (AI) in the fourth industrial revolution is a key building block and is becoming more significant as digitization increases.AI implementation in enterprises is increasingly focused on the technological and economic aspects, disregarding the human factors. In this context, the implementation and success of AI technologies depend on employee acceptance. Low employee adoption can lead to poorer performance as well as dissatisfaction. To ensure the expected added value through AI, it is necessary for companies to increase AI acceptance. People see AI as a machine with human intelligence that surpasses employees' capabilities and acts autonomously. Moreover, workers therefore fear that AI will replace humans and that they will lose their jobs in this way. This aspect leads to a distrust of the new technology. This results in a negative attitude towards AI. Since the research field of AI acceptance and its influencing factors have not been sufficiently investigated so far, the aim of this study is to analyze the development of AI acceptance in the industrial environment.In order to achieve the goal of this study, the systematic literature review according to Tranfield et al. (2003) is chosen as the research method, as it draws on previous results and in this way the development of acceptance can be investigated. After discussing the relevance of the topic and the resulting problem, an explanation of the terms that are considered important for the understanding of this study follows. Thereupon the systematic literature research is planned, in which different search terms and databases are determined.In order to analyze the development of the individual aspects, these were then compared with the factors from existing technology acceptance models from earlier years. This provides the insight that the workers without AI experience tend to reject the AI technologies due to the fear of consequences and other factors, therefore, an increase in AI understanding through improved expertise is required. In addition, this work shows that insufficient infrastructure in enterprises slows down AI adoption, which is one of the main problems. Based on the results, a model is established for this purpose, which is compared with the technology acceptance models and the Unified Theory of Acceptance and Use of Technology model to show the similarities and differences of the factors of technology acceptance.
Matthias Vogel, Giuseppe Strina, Christophe Said, Tobias Schmallenbach
Open Access
Article
Conference Proceedings
Application of Educational Context Data using Artificial Intelligence Methods
Today the web generates a large amount of data, the same ones that come from social networks, online platforms, communities, cloud computing, etc., but one type of data has not been recognized for its relevance and that is data from Learning Management Systems like Moodle in the educational context. Considering this context, this research will apply some Artificial Intelligence methods and techniques such as the TSA methodology, Text mining, and Sentiment Analysis to assess the data about the opinion of the students, converting them into stable information structures that allow their reflection and analysis. The work carried out focuses on determining the level of user satisfaction, in this case, the students, of the virtual learning platforms. The results obtained show that applying Artificial Intelligence allows obtaining relevant information that helps to undertake improvement actions by authorities and managers in the educational context based on the opinion of the students, detecting important problems in online learning during these times of COVID-19 we are just past.
Myriam Peñafiel, Maria Vasquez, Diego Vasquez
Open Access
Article
Conference Proceedings
An ontology-based modeling and CBR method for cable process planning
Cables exist in a large number of complex electronic devices, the quality of cable process design has a direct impact on the service quality and efficiency of the equipment. Cable process planning is a complex, time-consuming, and typically knowledge-intensive task that involves product information, process routing, parameters, and material selection, depending on the experience and knowledge of the process designers heavily. However, the unstructured and tacit nature of the knowledge makes it difficult to reuse. To implement knowledge-based intelligent cable process reasoning, and increase the design quality of the cable process plan while lowering the cost, it is critical to managing the cable process knowledge systematically and effectively. This paper proposes an ontology-based modeling method for cable product and process knowledge, in this approach, (1) A cable knowledge model containing cable product description and process plan is built; (2) Case-Based Reasoning (CBR) is employed to reuse knowledge from the previous case with the maximum similarity to realize rapid cable process planning, reducing the time and cost of process planning. In addition, a customized control cable process planning is taken as an example to verify the feasibility and effectiveness of the proposed method.
Chengjin Qiu, Xiaojun Liu, Changbiao Zhu, Feng Xiao
Open Access
Article
Conference Proceedings
A study on the current state of development of data-driven intelligent design and its impact on design paradigm
With the rise of big data and artificial intelligence, intelligent design platforms with data technology-driven and application scenarios have gradually become a common focus of design academia and industry, during which various intelligent design platforms have emerged, bringing profound changes to the design paradigm. This paper firstly, by collecting and analyzing related literature and cases, we elaborate the concept of data-driven intelligent design and sort out the research and application status of intelligent design tools based on the design process; then we analyze the impact of intelligent design on the design paradigm from three perspectives: design process, design object and designer, combined with the application and development status of intelligent design tools. We found that the development and application of intelligent design tools have made considerable progress, but at this stage the design process still requires the participation of human designers, so human-machine collaborative intelligence will be one of the long-term issues in the development of intelligent design tools; secondly, the application and development of intelligent design tools, while empowering the design process, also poses new challenges to the functions of designers and the adaptation of human-machine relationships.
Hu Zhao, Yuan Xiang
Open Access
Article
Conference Proceedings
Lowering the risk of bias in AI applications
Data is not free of biases, and AI systems that are based on the data are not either. What can be done to try the best, to minimize the risk of building systems that perpetuate the biases that exist in society and in data? In our paper we explore the possibilities along the User Centered Design Process and in Design Thinking, to lower the risk of keeping imbalances or gaps in data and models. But looking at the design process is not enough: Decision makers, development team and design team, respectively their composition and awareness towards risks of discrimination and their decisions in involving potential users and non-users, collecting data and testing the application also play a major role in trying to implement systems with the least biases possible.
Jj Link, Helena Dadakou, Anne Elisabeth Krüger
Open Access
Article
Conference Proceedings
'"Human Swarms” of novice sports fans beat professional handicappers when forecasting NFL football games
The biological phenomenon of Swarm Intelligence (SI) enables social species to converge on group decisions by interacting in real-time systems. Studied in schools of fish, bee swarms, and bird flocks, biologists have shown for decades that SI can greatly amplify group intelligence in natural systems. Artificial Swarm Intelligence (ASI) is a computer-mediated technique developed in 2015 to enable networked human groups to form real-time systems that can deliberate and converge on decisions, predictions, estimations, and prioritizations. A unique combination of real-time HCI methods and AI algorithms, ASI technology (also called “Human Swarming” or “Swarm AI”) has been shown in many studies to amplify group intelligence in forecasting tasks, often enabling small groups of non-professionals to exceed expert level performance. In the current study, small groups of approximately 24 amateur sports fans used an online platform called Swarm to collaboratively make weekly predictions (against the spread) of every football game in four consecutive NFL seasons (2019 - 2022) for a total of 1027 forecasted games. Approximately 5 games per week (as forecast by the human swarm) were identified as “predictable” using statistical heuristics. Performance was compared against the Vegas betting markets and measured against accepted performance benchmarks for professional handicappers. It is well known that professional bettors rarely achieve more than 55% accuracy against the Vegas spread and that top experts in the world rarely exceed 58% accuracy. In this study the amateur sports fans achieved 62.5% accuracy against the spread when connected as real-time “swarms.” A statistical analysis of this result (across 4 NFL seasons) found that swarms outperformed the 55% accuracy benchmark for human experts with significance (p=0.002). These results confirmed for the first time that groups of amateurs, when connected in real-time using ASI, can consistently generate forecasts that exceeded expert level performance with a high degree of statistical certainty.Keywords: Swarm Intelligence, Artificial Swarm Intelligence, Collective Intelligence, Wisdom of Crowds, Hyperswarms,
Hans Schumann, Louis Rosenberg, Gregg Willcox
Open Access
Article
Conference Proceedings
Interactive design of water purification products based on modern urban life
In order to explore innovative interaction methods from the technical level of AI, and further improve the use experience of interactive products, the author proposes the interactive design of AI urban modern life products. The author takes the artificial intelligence technology as the center, and applies the technical means to the product interaction design. After investigation and analysis of the technical means of its application, it summarizes how artificial intelligence drives the development of product interaction design. In addition, it further analyzes the application thinking and performance in the whole design process in combination with specific design cases. The results show that: The people aged 25-30 and 35-49 are undoubtedly the main consumers and users because of their economic foundation and health awareness, the main buyers are men, but women pay more attention to them, it can be seen that women have a strong degree of health awareness and sense of responsibility for their families. According to Maslow's needs theory, human needs are divided into five aspects: Physiological needs, security needs, social needs, respect needs and self-realization needs. At present, water purification products only reach the level of safety requirements, because of the design concept and technical limitations of traditional water purification products, the upgrading of products is slow, it is not comprehensive to simply emphasize the research and development of water purification technology. In the era of consumption upgrading, many water purification products ignore the social needs, respect needs and higher needs of consumers in the competitive environment, that is, the human-computer interaction mode, emotional experience of products and the sense of achievement of product use. The author puts forward the redefinition of multi-dimensional product design concepts such as traditional product interaction design methods, interactive interfaces and information architecture, and envisages the future development direction.
Sijia Wang
Open Access
Article
Conference Proceedings
An Analysis of the Spatial Distribution of the Crime in Urban Villages--Taking Dianqian Village,Xiamen as an Example
Urban villages are the legacy of China's rapid urban development, those is characterized by high mobility and complex structure, which is endangering the personal safety of residents in urban villages and the surrounding people, affecting the harmony and stability of the communities in urban villages, and not conducive to social and economic development. By taking the distribution of Robbery, Grab and Theft cases as the data source, combining with big data POI information obtained from Gaode, which is the one of the large map service companies in China. Methods such as kernel density analysis, standard deviation ellipsometry and spatial syntax theory are applied in the study, respectively, in order to quantitatively analyze the relationship between the spatial configuration and the environment of crime distribution within urban villages. The results show that: the accessibility, global integration, local integration and connectivity affect the occurrence of Robbery, Grab and Theft and the escape routes of criminals in the village. Different types of POI points and the occurrence of Robbery, Grab and Theft are correlated. This study helps to identify and improve the environmental factors that induce crime, and provides some references on security for the future renovation and construction of public spaces in urban villages in southern Fujian, China.
Yiting Wu
Open Access
Article
Conference Proceedings
Understanding Deepfakes: A Comprehensive Analysis of Creation, Generation, and Detection
This paper provides a comprehensive analysis of deepfakes, focusing on their creation, generation, and detection. Deepfakes are realistic fabricated videos, images, or audios generated using artificial intelligence algorithms. While initially seen as a source of entertainment and commercial applications, the negative social consequences of deepfakes have become apparent. They are misused for creating adult content, blackmailing individuals, and spreading misinformation, leading to a decline in trust and potential societal implications. The paper also discusses the importance of legislation in regulating the use of deepfakes and explores techniques for their detection, including machine learning and natural language processing. Understanding deepfakes is essential to address their ethical and legal implications in today's digital landscape.
Sami Alanazi, Seemal Asif
Open Access
Article
Conference Proceedings