|
Date: 4 October 2026 (Sunday) |
|
|
General public can also watch a live broadcast of the lecture through the YouTube Channel of the Hong Kong Public Libraries (https://www.youtube.com/channel/UCJUipdvSACZjO1ldoO3YQBA) |
|
|
Brief Introduction Can artificial intelligence help us read old books and archives? This lecture introduces a project that uses AI to recognize Manchu, an important language of Qing dynasty history that few people can read today. Through this example, the talk explains how new technology can help preserve endangered scripts, make rare historical materials easier to access, and bring the past closer to the public. Topic 2: Online Prices This lecture will share the analyze of an extensive set of high-frequency data from a major online shopping website in Hong Kong in order to understand how sellers set prices across a broad range of products. Consistent with earlier work on online markets, the research team observes that sellers tend to keep their prices relatively stable, yet the gap between the posted price and the underlying “regular” price is larger than previously reported, suggesting that temporary sales are more pronounced. This lecture will discuss how sellers react to short-term, intra-week fluctuations in demand—particularly those generated by platform-wide, members-only subsidies—by moving prices in the opposite direction and by practicing price discrimination. In contrast, sellers display muted price responses to larger, low-frequency demand surges that arise from consumption voucher disbursements or pandemic-induced lockdowns. This lecture will also discuss suggestive evidence that, in response to fiscal stimulus, sellers are more likely to issue promotional coupons, especially volume discounts, that encourage larger purchases. |