小苗 / Xiaomiao (AI researcher)

Overview

小苗 / Xiaomiao is the source-bounded AI researcher and doctoral student interviewed in EP174 about pandemic-disrupted study, travel, and algorithmic fairness.

Current Profile

The episode presents Xiaomiao as a Singapore-based researcher who traveled to the United States in early 2020 to present a first paper, visit courses and professors, and explore U.S. doctoral or Silicon Valley internship routes. COVID-19 interrupted that option set. After returning to China through Japan and entering centralized quarantine, he continued research in Beijing and later chose a doctorate with the Singapore team he already knew rather than applying to U.S. programs with reduced funding and uncertain admission.

His technical contribution is a bounded introduction to bias in language models and recommenders. He explains learned social associations, reporting-frequency distortion, similarity-based distribution, and the possibility that feeds allocate valuable information unequally. These are research questions and mechanisms, not audit findings about a named platform.

Key Characteristics

• AI researcher whose planned U.S. study and internship exploration was interrupted by COVID-19.
• Doctoral student who chose continuity with a familiar Singapore research team.
• Contributor to natural-language-processing research during the disrupted admissions period.
• Explainer of training-data bias, reporting bias, and recommendation-system fairness.
• Comparative observer of quarantine and reopening in China and Singapore.

Evidence

• U.S. exploration and return: EP174 describes his conference presentation, campus visits, California plans, return through Japan, and centralized quarantine.
• Doctoral choice: EP174 links reduced U.S. funding and uncertainty to his decision to remain with a known Singapore team.
• AI bias: EP174 records examples of gendered and racial association plus the distinction between textual mention frequency and real-world frequency.
• Recommendation fairness: EP174 presents his question about whether similarity-based feeds distribute high-value information unequally across communities.

Qualifications

This profile derives from one conversational 2021 episode and preserves the guest’s partial anonymity. It does not name his institution, papers, datasets, or audited systems, and its account of admissions funding, quarantine, and platform effects should not be generalized beyond the source.

What Changed

• Created a source-bounded profile for the EP174 AI-research guest.
• Distinguished his technical hypotheses from measured platform findings.

Relationships

• 无时差研究所 - podcast that interviewed him.
• 小元 - fellow guest offering the overseas-work and return-migration comparison.
• Pandemic Cross-Border Decision Compression - disruption that redirected his study plans.
• Training-Corpus Social Bias - language-model bias mechanism he explains.
• Algorithmic Opportunity Distribution - recommendation-fairness question he raises.