Shih-Chieh Dai (戴士捷)

SaLT Lab, the University of Utah

Advisor: Dr. Guanhong Tao

Penn State University & NLPSA, Academia Sinica

Advisor: Dr. Aiping Xiong & Lun-Wei Ku

Artificial Intelligence and Human-Centered Computing Lab, UT-Austin

Advisor: Dr. Matt Lease

Natural Language Processing and Sentiment Analysis Lab, Academia Sinica

Advisor: Dr. Lun-Wei Ku

User oriented computing Lab, NCCU

Advisor: Dr. Yi-Ling Lin

Generating Counterfactual Explanations for Fake Claims

Automated fact-checking systems have been proposed that quickly provide veracity prediction at scale to mitigate the negative influ- ence of fake news on people and on public opinion. However, most studies focus on veracity classifiers of those systems, which merely predict the truthfulness of news articles. We posit that effective fact checking also relies on people’s understanding of the predictions. We propose elucidating fact-checking predictions using counterfactual explanations to help people understand why a specific piece of news was identified as fake. In this work, generating counterfactual explanations for fake news involves three steps: asking good questions, finding contradictions, and reasoning appropriately. We frame this research question as contradicted entailment reasoning through question answering (QA).

Automating Response Evaluation: A Tool for Automatic Coding Responses to Open-ended Questions

The open-ended question is a widely used method in human-subject surveys. Open-ended questions provide opportunities for discovering human subjects’ spontaneous responses to the questions. However, it is labor-intensive and time- consuming to analyze the responses. In light of the challenge, I aim to leverage NLP methods to create a tool for assisting the analysis of open-ended questions. (UT Master's Capstone Project)

Detecting Political Ideological Leaning

This project aims to design a human-in-the-loop classification using active learning and rationales. The concept here is to build a framework that predicts the political ideology leans by the rationale provided by humans or extracted by the machine learning model.
My parts in this project are running the ERASER benchmark Bert2Bert model, trying active learning with uncertainty-based query strategy and random sampling, and implementing the random sampling baseline using Hugging Face (UT Master's Independent Study)

Escape from An Echo Chamber

An echo chamber effect refers to the phenomena that online users revealed selective exposure and ideological segregation on political issues. Prior studies indicate the connection between the spread of misinformation and online echo chambers. In this paper, to help users escape from an echo chamber, we propose a novel news-analysis platform that provides a panoramic view of stances towards a particular event from different news media sources. Moreover, to help users better recognize the stances of news sources which published these news articles, we adopt a news stance classification model to categorize their stances into “agree”, “disagree”, “discuss”, or “unrelated” to a relevant claim for specified events with political stances.

All the Wiser: Fake News Intervention Using User Reading Preferences

To address the increasingly significant issue of fake news, we develop a news reading platform in which we propose an implicit approach to reduce people’s belief in fake news. Specifically, we leverage reinforcement learning to learn an intervention module on top of a recommender system (RS) such that the module is activated to replace RS to recommend news toward the verification once users touch the fake news. To examine the effect of the proposed method, we conduct a comprehensive evaluation with 89 human subjects and check the effective rate of change in belief but without their other limitations. Moreover, 84% participants indicate the proposed platform can help them defeat fake news.

Exam Keeper: Detecting Questions with Easy-to-Find Answers

We present Exam Keeper, a tool to measure the availability of answers to exam questions for ESL students. Exam Keeper targets two major sources of answers: the web, and apps. ESL teachers can use it to estimate which questions are easily answered by information on the web or by using automatic question answering systems, which should help teachers avoid such questions on their exams or homework to prevent students from misusing technology.

NLPSA, Academia Sinica

Manage GPU servers in the lab and maintain web server, file server (NAS), and database (10 servers and 22 GPUs)

Fintech Research Center, NCCU

Developed a conformance tool using Flask, Python, Swagger and Docker for the third party providers to test interfaces and data endpoints against the Taiwan Open Banking API standard

Informaship

Developed website [Link] using HTML, JavaScript, CSS for Taiwan International Fisheries and Seafood Show (TIFSS)

Department of Computer Science, NCCU