cv

Primary web CV plus downloadable targeted PDF variants.

Contact Information

Name Yuan Tian
Professional Title Research and Projects
Email ytian24@stanford.edu
Phone (650)546-9848

Professional Summary

Researcher trained in linguistics with corpus, field, and behavioral research experience. Recent work focuses on AI-mediated interaction, multilingual communication, and human-AI differences in language use and interpretation.

Experience

  • 2026 - Present

    Stanford, California

    Research Assistant, SALT Lab
    Stanford University
    Investigate mitigation strategies for harmful AI companion interaction, especially emotional dependence, by testing conversational pushback/friction interventions.
    • Investigate mitigation strategies for harmful AI companion interaction, especially emotional dependence, by testing conversational pushback/friction interventions.
    • Built a theory-grounded codebook for risk constructs in multi-turn Character.AI conversations and aligned annotators to high agreement.
    • Design surveys and behavioral studies to identify interventions that reduce harm without sharply degrading engagement.
  • 2026 - Present

    Stanford, California

    Research Assistant, Social Interaction Lab
    Stanford University
    Compare communication strategies in human-human and human-LLM interaction to study how bilingual ambiguity is resolved under cognitive constraints.
    • Compare communication strategies in human-human and human-LLM interaction to study how bilingual ambiguity is resolved under cognitive constraints.
    • Design Mandarin behavioral experiments with a Codenames-style paradigm and collect interactive communication data for model comparison.
    • Develop computational analyses using Rational Speech Act and Resource Rationality frameworks to formalize strategic choice.
  • 2024 -

    Volunteer
    Beijing Stars and Rain Education Institute for Autism
    May 12, 2024
  • 2024 -

    Volunteer
    Peking University Library
    December 12, 2024
  • 2022 -

    Editorial Intern
    People's Literature Publishing House
    July 2022
  • 2022 - Present

    Editorial Team Leader
    Peking University Linguistics Journal
    September 2022 – Present

Education

  • 2025 - Present

    Stanford

    M.A. student
    Stanford University
    East Asian Languages and Cultures (Chinese Linguistics)
  • 2021 - 2025

    Beijing

    B.A. student
    Peking University
    Chinese linguistic & Sociology (Dual Bachelor's Degree)
    • From Languages to Information
    • Statistical Methods for Behavioral and Social Sciences
    • Introduction to Computing
    • Introduction to Psychology
    • Social Psychology
    • Cognitive Linguistics
    • Theoretical Linguistics
    • Chinese Phonetics
    • GPA: 3.98/4.0, Grade: 94.3/100, Rank: 1/120.
    • Scholarship: National Scholarship(1%), May4th Scholarship(1%), WeiMing Bachelor(1%), Academic Outstanding Scholarship(5%).
    • Honors: Merit Student Pacemaker (top 1%) for 3 consecutive years.
  • 2024 - 2024

    Online

    Summer School
    Leiden University
    Languages and Linguistics
    • Evolutionary linguistics
    • Introduction to syntax

Publications

Projects

  • SafeCompanion: Evaluating Pushback/Friction as an Intervention for Harmful AI Interaction

    Investigating mitigation strategies for high-risk AI companion interactions (e.g., emotional dependence) by evaluating user responses to varying levels of conversational friction.

    • Investigating mitigation strategies for high-risk AI companion interactions (e.g., emotional dependence) by evaluating user responses to varying levels of conversational friction.
    • Developing a theoretically grounded codebook to annotate risk-constructs in multi-turn Character.AI conversation datasets, establishing high inter-annotator agreement.
    • Designing and conducting user surveys and behavioral studies to identify Pareto-optimal interventions that reduce harm without significantly diminishing user engagement.
  • Bilingual Ambiguity Mental Models: Communication Strategies in Human vs. AI Interaction

    Comparing communication strategies in human-human versus human-LLM interactions to understand how cognitive constraints shape strategy adoption in resolving bilingual ambiguity.

    • Comparing communication strategies in human-human versus human-LLM interactions to understand how cognitive constraints shape strategy adoption in resolving bilingual ambiguity.
    • Designing and conducting behavioral experiments in Mandarin Chinese using a Codenames-style paradigm to elicit and evaluate interactive communication data.
    • Developing computational models utilizing Rational Speech Act (RSA) and Resource Rationality frameworks to formalize and analyze strategic decision-making processes.
  • AI Tutoring for Zhongkao Preparation

    Audited limitations in existing Chinese exam-prep apps, especially shallow explanation flows and weak support for students without reliable offline tutoring resources.

    • Audited limitations in existing Chinese exam-prep apps, especially shallow explanation flows and weak support for students without reliable offline tutoring resources.
    • Designed and prototyped an AI tutoring app for Zhongkao preparation around whiteboard-style walkthroughs, diagram-plus-text explanations, and stepwise problem solving grounded in learning-science research.
    • Positioned the product as a scalable educational-equity intervention to extend more structured and responsive exam-prep support to learners in under-resourced regions.
  • Pragmatics-Aware Evaluation of Degree Adverbs in Chinese LLM vs. Human Text

    Comparing Chinese degree adverb usage in LLM-generated vs. human text, motivated by Chinese word-class heterogeneity and pragmatics-driven POS ambiguity.

    • Comparing Chinese degree adverb usage in LLM-generated vs. human text, motivated by Chinese word-class heterogeneity and pragmatics-driven POS ambiguity.
    • Designing introspective prompts to probe edge cases (e.g., complex clause relations, modification scope) and elicit non-canonical constructions.
    • Preliminary findings indicate systematic “over-normalization” in model text (preference for canonical intensifiers and rigid patterns) and pragmatic mismatches in ambiguous-scope contexts.

Talks

The Grammaticalization of “Have”: A Corpus Study on Cantonese-English Bilinguals: Poster · TEAL-5 Conference, Tokyo · 2025-04-19
Attempting Principles for Lexical Item Determination: Based on a Semantic Field Study of Disease-Related Vocabulary in Ancient Chinese: Presentation · Conference on Corpus-based Ancient Chinese Studies, Peking University, Beijing · 2023-09-01
Attempting Principles for Lexical Item Determination: Disease-Related Vocabulary in Ancient Chinese: Workshop on Corpus-based Ancient Chinese Studies · Peking University · 2023-09-01

Skills

Data Analysis & Programming: R (proficient), Python (proficient), CLAN (familiar)
Other Tools: Powerpoint (proficient), Audacity (proficient), Eye tracker (proficient), Praat (familiar)

Languages

Mandarin : native
Jinyu Chinese : native
Classical Chinese : fluent
English : fluent
Cantonese : elementary proficiency
Spanish : elementary proficiency

Academic Interests

Focus Areas: human-centered NLP, multilingual and cross-cultural language use, pragmatics, semantics, social psychology

Honors and Awards

  •
    Peking University cohort rank: 1/120
  •
    National Scholarship
  •
    May 4th Scholarship
  •
    WeiMing Bachelor Scholarship