Towards Hybrid Architectures: Integrating Large Language Models in Informative Chatbots

2023 | International Conference on Business Informatics | Citations: 0

Authors: Arz von Straussenburg, Arnold F.; Wolters, Anna

Abstract: Informative chatbots embedded in an organization-specific context can provide a ... Expand

Abstract: Informative chatbots embedded in an organization-specific context can provide a reliable, interactive, and engaging source of information for users. However, traditional chatbot techniques have limitations in processing and understanding user input and generating human-like responses. On the other hand, the latest implementations of large language models show promising results in these domains but have limitations in providing accurate and up-to-date facts from domain-specific knowledge bases. With the advent of popular chatbots like Chat-GPT, they are increasingly becoming part of organizations' digital infrastructure. In this research-in-progress paper, we argue that the strengths and weaknesses of traditional techniques and large language models are complementary. Therefore, we propose a hybrid chatbot architecture that utilizes inter-agent communication to compensate for disadvantages while enhancing the chatbot's abilities, as perceived by the user. This approach will form the basis for development and evaluation using Design Science Research (DSR) as part of our research. Collapse

Topics: information retrieval user experience

Methods: chatbot design artifact design science NLP method large language model

Informational and emotional appeals of cover image in crowdfunding platforms and the moderating role of campaign outputs

2023 | Decision Support Systems | Citations: 0

Authors: Chen, Shiqin; Wang, Hongwei; Fang, Yulin; Wang, Wei

Abstract: The first-stage decision is critical in the decision process in crowdfunding, wh ... Expand

Abstract: The first-stage decision is critical in the decision process in crowdfunding, where the cover image plays an important role in forming first impressions and persuading click funders. This study explores how emotional visual cues (human face) and informational visual cues (visual text) in the cover image affect funders’ decisions and how the effects differ across campaigns. The results demonstrate the great persuasive power of human faces and visual text. We further illustrate the effects of joyful expressions and image-text congruency as well as the non-significant influence of the length and polarity of visual text. Moreover, emotional visual cues are more effective in intangible product campaigns, while informational visual cues are more effective in tangible product campaigns, and both are important in social service campaigns. Valuable insight is provided to creators about the design of the cover image. Collapse

Topics: crowdsourcing crowdfunding decision making application programming interface missing data

Methods: robustness check sentiment analysis descriptive statistic deep learning theoretical contribution

Theories: processing fluency theory

Extraction of Forward-looking Financial Information for Stock Price Prediction from Annual Reports Using NLP Techniques

2023 | HICSS | Citations: 0

Authors: Glodd, Alexander; Hristova, Diana

Abstract: Annual reports are one of the most important sources of information for financia ... Expand

Abstract: Annual reports are one of the most important sources of information for financial decisions. They contain forward-looking statements (FLS), which describe future trends and expectations. Thus, several studies deal with the automated identification of FLS, where the latest ones involve a combination of a rule-based approach and machine learning classification. In this paper, we extend this research with state-of-the-art NLP methods. We use DistilBERT for FLS identification and determine their sentiment with FinBERT. The result is processed by a Random Forest model for stock price growth prediction of different periods. Our evaluation shows that DestilBERT achieves higher accuracies on FLS identification than existing methods. For short-term stock price rate prediction, the extracted FLS information together with historical stock data outperforms the sole use of historical stock data. For mid-term prediction, using FLS alone with DestilBERT shows the best result. Finally, in the long-term, FLS provide no benefit. Collapse

Topics: Python decision making

Methods: BERT support vector machine natural language processing DistilBERT literature study

Champions for Social Good: How Can We Discover Social Sentiment and Attitude-Driven Patterns in Prosocial Communication?

2023 | Journal of the Association for Information Systems | Citations: 0

Authors: Mukkamala, Raghava; Kauffman, Robert; Henriksen, Helle

Abstract: The UN High Commissioner on Refugees (UNHCR) is pursuing a social media strateg ... Expand

Abstract: The UN High Commissioner on Refugees (UNHCR) is pursuing a social media strategy to inform people about displaced populations and refugee emergencies. It is actively engaging public figures to increase awareness through its prosocial communications and improve social informedness and support for policy changes in its services. We studied the Twitter communications of UNHCR social media champions and investigated their role as high-profile influencers. In this study, we offer a design science research and data analytics framework and propositions based on the social informedness theory we propose in this paper to assess communication about UNHCR's mission. Two variables-refugee-emergency and champion type-relate to the informedness of UNHCR champions' followers. Based on a Twitter sentiment and attitude corpus, we applied a five-step design science analytics framework involving machine learning and natural language processing to test how the emergency type and champion type impact social communication patterns. Positive and neutral sentiment dominated the tweets of the champions and their followers for most refugee-emergency types. High participation-intensity champions emphasized high-intensity emergencies with tweet patterns reflecting dominant positive or neutral sentiment and sharing/liking attitudes. However, we found that sports figures effects were limited in spreading UNHCR's message, despite their millions of followers. We demonstrate the power of data science for prosocial policy based on refugee crisis awareness and instantiate our methods and knowledge contributions in a research framework that derives knowledge, decisions, and actions from behavioral, design, and economics of information systems perspectives. Collapse

Topics: Twitter social media analytical information system big data application programming interface

Methods: design science BERT natural language processing unsupervised learning artificial neural network

Theories: Lewin's field theory

Getting Personal: A Deep Learning Artifact for Text-Based Measurement of Personality

2023 | Information Systems Research | Citations: 0

Authors: Yang, Kai; Lau, Raymond Y. K.; Abbasi, Ahmed

Abstract: Analysts, managers, and policymakers are interested in predictive analytics capa ... Expand

Abstract: Analysts, managers, and policymakers are interested in predictive analytics capable of offering better foresight. It is generally accepted that in forecasting scenarios involving organizational policies or consumer decision making, personal characteristics, including personality, may be an important predictor of downstream outcomes. The inclusion of personality features in forecasting models has been hindered by the fact that traditional measurement mechanisms are often infeasible. Text-based personality detection has garnered attention because of the public availability of digital textual traces. However, the text machine learning space has bifurcated into two branches: feature-based methods relying on manually crafted human intuition, or deep learning language models that leverage big data and compute, the main commonality being that neither branch generates accurate personality assessments, thereby making personality measures infeasible for downstream forecasting applications. In this study, we propose DeepPerson, a design artifact for text-based personality detection that bridges these two branches by leveraging concepts from relevant psycholinguistic theories in conjunction with advanced deep learning strategies. DeepPerson incorporates novel transfer learning and hierarchical attention network methods that use psychological concepts and data augmentation in conjunction with person-level linguistic information. We evaluate the utility of the proposed artifact using an extensive design evaluation on three personality data sets in comparison with state-of-the-art methods proposed in academia and industry. DeepPerson can improve detection of personality dimensions by 10–20 percentage points relative to the best comparison methods. Using case studies in the finance and health domains, we show that more accurate text-based personality detection can translate into significant improvements in downstream applications such as forecasting future firm performance or predicting pandemic infection rates. Our findings have important implications for research at the intersection of design and data science, and practical implications for managers focused on enabling, producing, or consuming predictive analytics.History: Olivia Sheng, Senior Editor; Huimin Zhao, Associate Editor.Funding: This work was supported by Oracle for Research (NLP for the Greater Good) and the U.S. National Science Foundation’s Division of Information and Intelligent Systems [Grants BDS-1636933 and IIS-1816504]. The work was also partly supported by the Research Grants Council of the Hong Kong Special Administrative Region [Project: CityU 11507219] and the City University of Hong Kong SRG [Project: 7005780].Supplemental Material: The online appendices are available at https://doi.org/10.1287/isre.2022.1111. Collapse

Topics: personality Twitter social media pandemic Google+

Methods: machine learning BERT design science deep learning computational algorithm

Theories: big five model

Comparative Analysis of Classical and Deep Learning-based Natural Language Processing for Prioritizing Customer Complaints

2022 | HICSS | Citations: 0

Authors: Blümel, Jan; Zaki, Mohamed

Abstract: Recent advancements in natural language processing have been shown to be very ef ... Expand

Abstract: Recent advancements in natural language processing have been shown to be very effective for different text mining tasks and thus have provided the opportunity to enhance service research. To improve the customer service experience, this paper compares several natural language processing approaches in order to automatically prioritize incoming customer complaints for service agents. This can help companies to reduce customers’ friction and enable effective resource allocations. Our paper uses stateof-the-art feature engineering techniques (e.g., term frequency, TF-IDF and Word2Vec) to identify key words that could enable machine to prioritize complainers. We experimented with many classical machine learning classification algorithms, such as Random Forests, Support Vector Machines, Decision Trees and Logistic Regression, as well as with deep learning-based classifiers, such as convolutional neural networks, bidirectional long short-term memory, and the pre-trained language model BERT to compare the model performance. Our findings show that the pre-trained language model BERT and TFIDF in combination with Logistic Regression yields the highest macro averaged F1-score across the multiple classes and is therefore most capable of predicting the priority group of incoming customer complaints. Collapse

Topics: Wikipedia customer service database system Google+ application programming interface

Methods: natural language processing deep learning machine learning term frequency–inverse document frequency word2vec

Transformer-based Summarization and Sentiment Analysis of SEC 10-K Annual Reports for Company Performance Prediction

2022 | HICSS | Citations: 0

Authors: Hsieh, Hsin-Ting; Hristova, Diana

Abstract: Annual reports published by companies contain important insights regarding their ... Expand

Abstract: Annual reports published by companies contain important insights regarding their performance and are often analyzed in a manual, subjective manner. We address this point by combining the streams of research on text summarization and topic modelling with the one on sentiment analysis. Our approach consists of the steps of text summarization using BERTSUMEXT, topic modelling with LDA, sentiment analysis with FinBERT, and performance prediction with Decision Trees and Random Forest. The result provides decision makers with an interpretable and condensed representation of the content of annual reports, together with its relationship to future company performance. We evaluate our approach on 10-K reports, demonstrating both its interpretability for analysts and explanatory power regarding future company performance. Collapse

Topics: organizational productivity data breach cloud computing Python explainable artificial intelligence

Methods: sentiment analysis topic model machine learning BERT decision tree classification

ASPECT-BASED SENTIMENT ANALYSIS FOR UNIVERSITY TEACHING ANALYTICS

2022 | European Conference On Information Systems | Citations: 0

Authors: Schurig, Tim; Zambach, Sine; Mukkamala, Raghava Rao; Petry, Malte

Abstract: Aspect-based sentiment analysis (ABSA) is a natural language processing method ... Expand

Abstract: Aspect-based sentiment analysis (ABSA) is a natural language processing method to analyze sentiments from large amounts of unstructured text in a much more fine-grained manner at the aspect level. In this research work, we apply it to analyze open text replies from surveys regarding online teaching. Like most other educational institutions, Copenhagen Business School (CBS) had to shift to online teaching from one day to the next. Using ABSA, we investigated the impact of this forced online learning experiment on teaching quality in the spring semester of 2020. Our findings reveal that students disliked online teaching due to insufficient information and unadjusted teaching methods. However, students liked its flexibility and possibility to learn at an individual pace. We show that ABSA can extract valuable information in an easily interpretable manner to support teaching and learning processes. Finally, our findings show that ABSA is a valuable tool to analyze unstructured text quantitatively. Collapse

Topics: online learning system Python decision making

Methods: survey literature study qualitative interview natural language processing NLTK

Exploring Machine-based Idea Landscapes – The Impact of Granularity

2022 | HICSS | Citations: 0

Authors: Wahl, Julian; Stroehle, Thomas; Füller, Johann; Hutter, Katja

Abstract: Effective exploration of a landscape full of crowdsourced ideas depends on the r ... Expand

Abstract: Effective exploration of a landscape full of crowdsourced ideas depends on the right search strategy, as well as the level of granularity in the representation. To categorize similar ideas on different granularity levels modern natural language processing methods and clustering algorithms can be usefully applied. However, the value of machine-based categorizations is dependent on their comprehensibility and coherence with human similarity perceptions. We find that machine-based and human similarity allocations are more likely to converge when comparing ideas across more distant solution clusters than within closely related ones. Our exploratory study contributes to research on the navigability of idea landscapes, by pointing out the impact of granularity on the exploration of crowdsourced knowledge. For practitioners, we provide insights on how to organize the search for the best possible solutions and control the cognitive demand of searchers. Collapse

Topics: crowdsourcing Wikipedia user-generated content programming language knowledge base

Methods: hierarchical clustering term frequency–inverse document frequency NLP method cluster analysis natural language processing

From semantics to pragmatics: where IS can lead in Natural Language Processing (NLP) research

2021 | European Journal of Information Systems | Citations: 0

Authors: Li, Yan; Thomas, Manoj a; Liu, Dapeng

Abstract: Natural Language Processing (NLP) is now widely integrated into web and mobile a ... Expand

Abstract: Natural Language Processing (NLP) is now widely integrated into web and mobile applications, enabling natural interactions between humans and computers. Although there is a large body of NLP studies published in Information Systems (IS), a comprehensive review of how NLP research is conceptualised and realised in the context of IS has not been conducted. To assess the current state of NLP research in IS, we use a variety of techniques to analyse a literature corpus comprising 356 NLP research articles published in IS journals between 2004 and 2018. Our analysis indicates the need to move from semantics to pragmatics. More importantly, our findings unpack the challenges and assumptions underlying current research trends in NLP. We argue that overcoming these challenges will require a renewed disciplinary IS focus. By proposing a roadmap of NLP research in IS, we draw attention to three NLP research perspectives and present future directions that IS researchers are uniquely positioned to address. Collapse

Topics: database system knowledge representation social media analytics social media information retrieval

Methods: natural language processing qualitative content analysis literature study computational algorithm literature sample