People Feature

Linda Hong Cheng

MLitt in Sociology

Linda Hong Cheng is a BBC-featured Clarendon Scholar, AI founder, and DPhil researcher at Nuffield College, affiliated with the Leverhulme Centre for Demographic Science (LCDS) and International Max Planck Research School for Population, Health and Data Science (IMPRS-PHDS). She is also the Founder & CEO of Lychee Labs, an Oxford-born AI lab building physics-informed industrial intelligence for mission-critical manufacturing - minimizing waste and maximizing yield for battery, advanced chemical, precision device, and semiconductor production.

Linda’s research interests broadly encapsulate novel applications of artificial intelligence, physics-informed world models, applied machine learning, natural language processing, agent-based modeling, and hybrid AI systems to analyzing complex systems under constraint. Across academia and industry, her work asks: how can AI model, diagnose, and intervene in high-dimensional systems shaped by hidden feedback loops, nonlinear dynamics, institutional power, physical constraints, and emergent behavior? Her research spans colonial-patriarchal gender disparities, social inequalities, social demographic trends, contentious politics, digital inclusion, industrial process drift, physical-world foundation models, and adaptive intelligence in mission-critical production environments.

Her DPhil research, affiliated with the Gates Foundation-funded Digital Gender Gaps project, establishes “digital gender circularity”: the symbiotic relationships between increasing digital gender equality, offline gender equality, and global sustainable development. Rather than treating digital inequality as a narrow technology-access problem, this work conceptualizes it as a complex adaptive system shaped by platform infrastructures, institutional power, political economy, social norms, uneven global development, and recursive feedback between digital and offline inequality.

Most recently, Linda’s AI research has expanded into post-AGI physical and social systems through two ICLR Post-AGI Science and Society Workshop papers. In “Post-AGI Physical World Foundation Models: Event-Conditional Dynamics, Topology Transfer, and Risk-Limiting Guarantees,” she develops a framework for physical-world foundation models that reason across event-conditioned dynamics, industrial topology, time-series signals, uncertainty, and deploy-or-abstain decisions in high-stakes physical environments. In “Will AGI Accelerate Algorithmic Patriarchy? Digital Gender Circularity in the Generative AI Era,” she extends digital gender circularity into the generative AI era, analyzing how AI access, skills, legitimacy, labor-market conversion, and governance capacity may either reduce inequality or accelerate algorithmic patriarchy.

Linda's academic work is the intellectual bedrock of Lychee Labs - which applies physics-informed and hybrid AI to advanced manufacturing - focusing on battery and industrial production systems, particularly the relationship between upstream process drift, physical mechanisms, downstream defect risk, yield outcomes, and engineer-actionable root-cause analysis.

Linda was among the early AI (ML/NLP) researchers. She was the youngest invited author for the Oxford Handbook of the Sociology of Machine Learning, where she pioneered decolonial AI + NLP applications and established Chinese Computational Sociology - a new subfield situated at the intersections of sociology, computational methods, machine learning, and China studies. Her computational research (applied ML & NLP) modeling gender bias and sociopolitical unrest is published in Mobilization: An International Quarterly - the world’s premier social movements journal.

Previously, Linda founded Mung!, the world’s first AI-driven AgeTech startup serving older consumers (BBC-featured), and Girlpane (BBC Sounds & China Daily-featured, with sold-out exhibits in London and Oxford), an ArtTech collective democratizing access for women and marginalized artists to the global art market.

Linda is a global public speaker - sharing critical AI & tech insights on the stages of World Economic Forum (Stanford-MIT AI Summit), Google, IBM, Flutter, Oxford Internet Institute, Oxford AI Forum, Columbia University, Pride in Tech, the hallowed halls of centuries-old Oxford colleges, and more.

Prior to Oxford, Linda completed her MA in Applied Machine Learning (Regional Studies: East Asia - Computational Sociology) at Columbia University, where she was the only student in her cohort fully funded by the Weatherhead East Asia Institute FLAS Fellowship. Her Master’s thesis, published in Mobilization: An International Quarterly, uses applied machine learning, NLP, novel dictionary methods, feature engineering, and statistical modeling on novel big data from Weibo to analyze gender bias in media and government attention to protest events in Mainland China. Of particular interest in this work is how patriarchy is made, unmade, reified, and transgressed by the actions of individual actors, government bureaucracy, media institutions, and digital platforms.

Linda’s undergraduate senior honors thesis, winner of the prestigious Chancellor’s Best Honors Thesis Prize, uses survey and interview methods to model contentious politics in post-1978 reform China - analyzing the complex, oft-contradictory motivations behind the 1989 Tiananmen Square student protesters’ choice of “Nothing To My Name” as their protest anthem. Through micro-exploratory profiles of civilians, protesters, and state actors, this work weaves a larger political-economic tapestry of China’s turbulent post-1978 reform era, ultimately culminating in the explosive Tiananmen protests. Linda’s research critically intervenes in mainstream Western-dominated narratives of the Tiananmen protests, conceptualizing the protests as a site through which students attempted to negotiate their relationship, and loss thereof, with the state: the loss of “everything.”

Previously, Linda taught:

Politics of Social Movements, which analyzes contentious politics and state-society relations through an anti-colonial lesbian/queer feminist framework;

Political Activism in China, which critically analyzes watershed political movements in contemporary China;

Sociology of Social Change, which interrogates neo-colonialism through exploring varied vehicles of social transformation in post-colonial societies throughout the world.

Publications

Cheng, Linda. “Post-AGI Physical World Foundation Models: Event-Conditional Dynamics, Topology Transfer, and Risk-Limiting Guarantees.” International Conference on Learning Representations: Post-AGI Workshop. April 2026.

Cheng, Linda. “Will AGI Accelerate Algorithmic Patriarchy? Digital Gender Circularity in the Generative AI Era.” International Conference on Learning Representations: Post-AGI Workshop. April 2026.

Cheng, Linda, and Yao Lu (2024). “Chapter 5. Chinese Computational Sociology: Decolonial Applications of Machine Learning and Natural Language Processing Methods in Chinese-Language Contexts.” In Oxford Handbook of the Sociology of Machine Learning; ML as a methodological toolbox, edited by Juan Pablo Pardo-Guerra and Christian Borch.

  • As the youngest scholar invited to contribute, I establish ‘Chinese computational sociology’: A new, cutting-edge subfield situated at the intersections of AI, applied machine learning / computational methods, sociology, and China studies. We review existing literature in this growing subfield and elucidate the innovations, challenges, and future trajectories of this exciting new area.

Cheng, Linda, Yao Lu, and Han Zhang (2024). “Patriarchal Erasure and Manufactured Passivity: Gender Bias in Government and Media Attention to Protests in China.” Mobilization: An International Quarterly 29(2).

  • Revised from my MA thesis, which earned an A+ (highest mark possible). 
  • Applies advanced machine learning, natural language processing (NLP), and statistical methods to analyze gender-biased stratification of government and media attention to protests in China, using big social media data from Weibo. Presents a theory of the triadic gender-media-protest relationship.

Cheng, Linda. “Informal Labor Stratification in Pandemic-Era Asia: Considering Gender, Class, and Nativity.” Asian Pacific Affairs Council Journal (Invited Contributor), May 2023.

Cheng, Linda. “Resilience Among Gender Marginalized People in China During COVID-19 中國性邊緣者於新冠疫情中的韌性.” 12th Consortium of African and Asian Studies (CAAS) Proceedings (Invited Contributor), August 2022.

Cheng, Linda. “Accounting for the Gender-Protest-Media Triad Using Quantitative and Computational Methods.” The Reed (Invited Contributor): 16-20, July 2022.

Cheng, Linda. “Gender Dynamics of Protest and Visibility in China: Biased Erasure and Manufactured Passivity.” SynThesis Abstract Journal, Columbia University. April 2021.

Cheng, Linda. “Breach of Trust as Fuel for Protest: Tiananmen Demonstrations and the Erosion of State-Society Relations in 1980s China.” UNC-CH Senior Honors Theses. May 2020.

  • Winner of the prestigious Chancellor’s Best Senior Honors Thesis Award, awarded to one student every year.
  • Employs statistical methods, large-scale surveys, ethnographic interviews, and policy/media analysis to analyze how post-1978 reform social changes contributed to the 1989 Tiananmen Square Protests; proposes a theory of paternalist state-society relations within China. Presented this research at multiple university-level conferences.

Cheng, Linda. “Women’s Liberation in China: Necessity or Afterthought?JOURney 2: 69-73. March 2019.