arXiv subjects

Jenny Huang

Publications and source records attributed to Jenny Huang.

5 recordsLinked to original sources

arXiv2026

LLM-Based User Personas for Recommendations at Scale

Large Language Models (LLMs) offer unprecedented potential for enhancing recommendation systems through their world knowledge and reasoning capabilities. However, existing approaches often rely on structured IDs or offline processing, limiting semantic richness, real-time adaptability, and user-facing interpretability. In this paper, we introduce a novel framework that enables real-time generation of LLM-based user interest personas for a large-scale commercial video recommendation platform. Our method generates natural-language user interest personas that address the exploitation-exploration trade-off by combining the summarization of existing interests with novel topics, directly during serving. To overcome the computational challenges of online LLM inference at a billion-user scale, we design a cost-efficient architecture leveraging knowledge distillation, asynchronous inference, and input optimization via semantically clustered video representations. Extensive offline evaluations, user studies, and live A/B tests demonstrate significant improvements in viewer value. This work bridges the gap between high-level semantic understanding and industrial-scale recommendation, paving the way for more dynamic, explainable, and satisfying personalized experiences.

cs.IR↗

arXiv2023

Fitting a stochastic model of intensive care occupancy to noisy hospitalization time series during the COVID-19 pandemic

Intensive care occupancy is an important indicator of health care stress that has been used to guide policy decisions during the COVID-19 pandemic. Toward reliable decision-making as a pandemic progresses, estimating the rates at which patients are admitted to and discharged from hospitals and intensive care units (ICUs) is crucial. Since individual-level hospital data are rarely available to modelers in each geographic locality of interest, it is important to develop tools for inferring these rates from publicly available daily numbers of hospital and ICU beds occupied. We develop such an estimation approach based on an immigration-death process that models fluctuations of ICU occupancy. Our flexible framework allows for immigration and death rates to depend on covariates, such as hospital bed occupancy and daily SARS-CoV-2 test positivity rate, which may drive changes in hospital ICU operations. We demonstrate via simulation studies that the proposed method performs well on noisy time series data and apply our statistical framework to hospitalization data from the University of California, Irvine (UCI) Health and Orange County, California. By introducing a likelihood-based framework where immigration and death rates can vary with covariates, we find, through rigorous model selection, that hospitalization and positivity rates are crucial covariates for modeling ICU stay dynamics and validate our per-patient ICU stay estimates using anonymized patient-level UCI hospital data.

stat.ME↗

arXiv2022

Detecting Changes in the Transmission Rate of a Stochastic Epidemic Model

Throughout the course of an epidemic, the rate at which disease spreads varies with behavioral changes, the emergence of new disease variants, and the introduction of mitigation policies. Estimating such changes in transmission rates can help us better model and predict the dynamics of an epidemic, and provide insight into the efficacy of control and intervention strategies. We present a method for likelihood-based estimation of parameters in the stochastic SIR model under a time-inhomogeneous transmission rate comprised of piecewise constant components. In doing so, our method simultaneously learns change points in the transmission rate via a Markov chain Monte Carlo algorithm. The method targets the exact model posterior in a difficult missing data setting given only partially observed case counts over time. We validate performance on simulated data before applying our approach to data from an Ebola outbreak in Western Africa and COVID-19 outbreak on a university campus.

stat.ME↗

arXiv2019

Garbage In, Garbage Out? Do Machine Learning Application Papers in Social Computing Report Where Human-Labeled Training Data Comes From?

Many machine learning projects for new application areas involve teams of humans who label data for a particular purpose, from hiring crowdworkers to the paper's authors labeling the data themselves. Such a task is quite similar to (or a form of) structured content analysis, which is a longstanding methodology in the social sciences and humanities, with many established best practices. In this paper, we investigate to what extent a sample of machine learning application papers in social computing --- specifically papers from ArXiv and traditional publications performing an ML classification task on Twitter data --- give specific details about whether such best practices were followed. Our team conducted multiple rounds of structured content analysis of each paper, making determinations such as: Does the paper report who the labelers were, what their qualifications were, whether they independently labeled the same items, whether inter-rater reliability metrics were disclosed, what level of training and/or instructions were given to labelers, whether compensation for crowdworkers is disclosed, and if the training data is publicly available. We find a wide divergence in whether such practices were followed and documented. Much of machine learning research and education focuses on what is done once a "gold standard" of training data is available, but we discuss issues around the equally-important aspect of whether such data is reliable in the first place.

cs.CY↗

arXiv2017

Individuals, Institutions, and Innovation in the Debates of the French Revolution

The French Revolution brought principles of "liberty, equality, and brotherhood" to bear on the day-to-day challenges of governing what was then the largest country in Europe. Its experiments provided a model for future revolutions and democracies across the globe, but this first modern revolution had no model to follow. Using reconstructed transcripts of debates held in the Revolution's first parliament, we present a quantitative analysis of how this system managed innovation. We use information theory to track the creation, transmission, and destruction of patterns of word-use across over 40,000 speeches and more than one thousand speakers. The parliament as a whole was biased toward the adoption of new patterns, but speakers' individual qualities could break these overall trends. Speakers on the left innovated at higher rates while speakers on the right acted, often successfully, to preserve prior patterns. Key players such as Robespierre (on the left) and Abbé Maury (on the right) played information-processing roles emblematic of their politics. Newly-created organizational functions---such as the Assembly's President and committee chairs---had significant effects on debate outcomes, and a distinct transition appears mid-way through the parliament when committees, external to the debate process, gain new powers to "propose and dispose" to the body as a whole. Taken together, these quantitative results align with existing qualitative interpretations but also reveal crucial information-processing dynamics that have hitherto been overlooked. Great orators had the public's attention, but deputies (mostly on the political left) who mastered the committee system gained new powers to shape revolutionary legislation.

physics.soc-ph↗