Ming Qian

HCI International, Virtual (July 2023)

Translators search for information to resolve various types of uncertainties they face such as confirming the source of original texts, gaining proper understanding, and verifying whether the selected keywords are spelled correctly, commonly used, or matched properly between source and target languages. Under the constraints of tighter time schedules and stricter cost-effectiveness requirements imposed by the machine translation (MT) plus human post-editing (PE) business model, translators strive to achieve the goals of information seeking with enhanced efficiency and accuracy. This study investigates four information seeking strategies: (1) top-ranked results or featured snippets returned by large-scale web search engines; (2) abductive reasoning based on search result counts returned by large-scale web search engines; (3) direct answers provided by ChatGPT—a long-form question-answering conversational AI; (4) A novel conversational search engine (Perplexity.ai) combining OpenAI’s GPT language modeling technology and a large Internet search engine—Microsoft Bing. Human users should focus on forming effective human queries to develop an effective collaboration between human and large-scale web search engines, long-form question-answering conversational AI systems, and conversational search engine. While top-ranked search results and count-based abductive reasoning are effective strategies, emerging technologies such as long-form question-answering conversational AI and conversational search engine provide accurate and comprehensive answers to specific queries, and additional data and resources such as reference links and related questions.

For More Information

To learn more or request a copy of a paper (if available), contact Ming Qian.

(Please include your name, address, organization, and the paper reference. Requests without this information will not be honored.)