Ting-Chih Chen Email & Phone Number

Blacksburg, Virginia, United States

- Developed a multimodal claim generation method and a multiclass classification on news articles.- Built up an evidence summarization model to assist fact‑checkers in distinguishing the truthfulness of claims.- Built up an instrumental video dataset and conducted research on inductive and abductive reasoning.- Employed SOTA multimodal models to extract valuable insights from the dataset and induce knowledge from LLMs, resulting in an impressive Bert score F1 of 86.5%.Summary: (1) Submitted 2 top-tier papers (ACL, EMNLP). (2) Worked on detecting disinformation in the meta news data. (3) Built up a large multimodalities dataset for the fact-checking task. (4) Researched in reasoning the cross-modalities information.Contributions: (1) Created a new data format that combines knowledge graphs with heterogeneous data. (2) Studied a method for claim generation using AMR graphs and a traversal algorithm. (3) Trained a model that can summarize information from multi‑modalities information. (4) Explored methods to enhance the performance of the summarization model through RLHF and PPO to fine‑tune the model with Llama 2.

Blacksburg, Virginia, United States

Fall 2022: CS3114 Data Structures and AlgorithmsSpring 2023: CS5024 Ethics and Professionalism in Computer Science

Changhua County, Taiwan, Taiwan

- Developed the cryptocurrency and smart contract.- Conducted extensive research on the integration of blockchain technology with industrial technology.- Analyzed the challenges and opportunities associated with implementing blockchain solutions in industrial settings.

Taichung City, Taiwan

- Built up a boat image dataset for research and explored YOLOv3 extensively in object detection task, achieving an 82% mAP.Summary: At the National Center for High-Performance Computing, I worked on two projects during my research internship. The first project involved creating a government database that consolidates heterogeneous data from both commercial and military ports. We aimed to identify unknown boats to ensure security. We achieved this by training the YOLO model to detect boats and determine if they belong to us or not. My second project focused on the smart city initiative. We aimed to integrate an object detection model with traffic lights, enabling them to automatically detect cars and control the lights accordingly. This approach aims to make traffic lights more environmentally friendly.

Master'S Degree, Computer Science, 3.7/4.0

MS thesis: Multimodal Multi-Document Evidence Summarization For Fact-Checking

Bachelor'S Degree, Computer Science, 3.4/4.0

- Undergraduate senior project: Data Distribution Strategy With High Computational Efficiency In Spark - Deployed a distribution system.