Liu Leqi
YOU?
Author Swipe
View article: SFT-TA: Supervised Fine-Tuned Agents in Multi-Agent LLMs for Automated Inductive Thematic Analysis
SFT-TA: Supervised Fine-Tuned Agents in Multi-Agent LLMs for Automated Inductive Thematic Analysis Open
Thematic Analysis (TA) is a widely used qualitative method that provides a structured yet flexible framework for identifying and reporting patterns in clinical interview transcripts. However, manual thematic analysis is time-consuming and …
View article: Position: Thematic Analysis of Unstructured Clinical Transcripts with Large Language Models
Position: Thematic Analysis of Unstructured Clinical Transcripts with Large Language Models Open
This position paper examines how large language models (LLMs) can support thematic analysis of unstructured clinical transcripts, a widely used but resource-intensive method for uncovering patterns in patient and provider narratives. We co…
View article: Parasitism and Stripping of Desire: Lady Macbeth as the Host of Macbeths Tyrannical Personality
Parasitism and Stripping of Desire: Lady Macbeth as the Host of Macbeths Tyrannical Personality Open
Macbeth, as a representative of Shakespeares classic tyrant image, reflects the complex psychological contradictions of the character under moral dilemma. Traditional scholarship on Macbeth has long framed the witches as supernatural agent…
View article: Hypothesis Testing for Quantifying LLM-Human Misalignment in Multiple Choice Settings
Hypothesis Testing for Quantifying LLM-Human Misalignment in Multiple Choice Settings Open
As Large Language Models (LLMs) increasingly appear in social science research (e.g., economics and marketing), it becomes crucial to assess how well these models replicate human behavior. In this work, using hypothesis testing, we present…
View article: Deep mechanism design: Learning social and economic policies for human benefit
Deep mechanism design: Learning social and economic policies for human benefit Open
Human society is coordinated by mechanisms that control how prices are agreed, taxes are set, and electoral votes are tallied. The design of robust and effective mechanisms for human benefit is a core problem in the social, economic, and p…
View article: Exploring the best way for UAV visual localization under Low-altitude Multi-view Observation Condition: a Benchmark
Exploring the best way for UAV visual localization under Low-altitude Multi-view Observation Condition: a Benchmark Open
Absolute Visual Localization (AVL) enables Unmanned Aerial Vehicle (UAV) to determine its position in GNSS-denied environments by establishing geometric relationships between UAV images and geo-tagged reference maps. While many previous wo…
View article: A Common Pitfall of Margin-based Language Model Alignment: Gradient Entanglement
A Common Pitfall of Margin-based Language Model Alignment: Gradient Entanglement Open
Reinforcement Learning from Human Feedback (RLHF) has become the predominant approach for language model (LM) alignment. At its core, RLHF uses a margin-based loss for preference optimization, specifying ideal LM behavior only by the diffe…
View article: A Unified Causal Framework for Auditing Recommender Systems for Ethical Concerns
A Unified Causal Framework for Auditing Recommender Systems for Ethical Concerns Open
As recommender systems become widely deployed in different domains, they increasingly influence their users' beliefs and preferences. Auditing recommender systems is crucial as it not only ensures the continuous improvement of recommendati…
View article: Using deep reinforcement learning to promote sustainable human behaviour on a common pool resource problem
Using deep reinforcement learning to promote sustainable human behaviour on a common pool resource problem Open
A canonical social dilemma arises when finite resources are allocated to a group of people, who can choose to either reciprocate with interest, or keep the proceeds for themselves. What resource allocation mechanisms will encourage levels …
View article: Accounting for AI and Users Shaping One Another: The Role of Mathematical Models
Accounting for AI and Users Shaping One Another: The Role of Mathematical Models Open
As AI systems enter into a growing number of societal domains, these systems increasingly shape and are shaped by user preferences, opinions, and behaviors. However, the design of AI systems rarely accounts for how AI and users shape one a…
View article: Prompting Fairness: Integrating Causality to Debias Large Language Models
Prompting Fairness: Integrating Causality to Debias Large Language Models Open
Large language models (LLMs), despite their remarkable capabilities, are susceptible to generating biased and discriminatory responses. As LLMs increasingly influence high-stakes decision-making (e.g., hiring and healthcare), mitigating th…
View article: Personalized Language Modeling from Personalized Human Feedback
Personalized Language Modeling from Personalized Human Feedback Open
Personalized large language models (LLMs) are designed to tailor responses to individual user preferences. While Reinforcement Learning from Human Feedback (RLHF) is a commonly used framework for aligning LLMs with human preferences, vanil…
View article: A Taxonomy of Human and ML Strengths in Decision-Making to Investigate Human-ML Complementarity
A Taxonomy of Human and ML Strengths in Decision-Making to Investigate Human-ML Complementarity Open
Hybrid human-ML systems increasingly make consequential decisions in a wide range of domains. These systems are often introduced with the expectation that the combined human-ML system will achieve complementary performance, that is, the co…
View article: Development analysis and future of computer vision-based automobile license plate recognition technology and perception technology
Development analysis and future of computer vision-based automobile license plate recognition technology and perception technology Open
With the increasing number of automobiles, there is a growing demand for the recognition and perception of license plates. As a unique identifier for each vehicle, license plates can help traffic management departments with vehicle trackin…
View article: A Field Test of Bandit Algorithms for Recommendations: Understanding the Validity of Assumptions on Human Preferences in Multi-armed Bandits
A Field Test of Bandit Algorithms for Recommendations: Understanding the Validity of Assumptions on Human Preferences in Multi-armed Bandits Open
Personalized recommender systems suffuse modern life, shaping what media we read and what products we consume. Algorithms powering such systems tend to consist of supervised learning-based heuristics, such as latent factor models with a va…
View article: Off-Policy Risk Assessment in Markov Decision Processes
Off-Policy Risk Assessment in Markov Decision Processes Open
Addressing such diverse ends as safety alignment with human preferences, and the efficiency of learning, a growing line of reinforcement learning research focuses on risk functionals that depend on the entire distribution of returns. Recen…
View article: Ambient Air Pollution and Hospitalization for Acute Myocardial Infarction in Chongqing, China: A Time-Stratified Case Crossover Analysis
Ambient Air Pollution and Hospitalization for Acute Myocardial Infarction in Chongqing, China: A Time-Stratified Case Crossover Analysis Open
Previous studies have demonstrated that short-term exposure to ambient air pollution was associated with hospital admissions for cardiovascular diseases, but the evidence of its effects on acute myocardial infarction (AMI) in East Asian co…
View article: Modeling Attrition in Recommender Systems with Departing Bandits
Modeling Attrition in Recommender Systems with Departing Bandits Open
Traditionally, when recommender systems are formalized as multi-armed bandits, the policy of the recommender system influences the rewards accrued, but not the length of interaction. However, in real-world systems, dissatisfied users may d…
View article: Supervised Learning with General Risk Functionals
Supervised Learning with General Risk Functionals Open
Standard uniform convergence results bound the generalization gap of the expected loss over a hypothesis class. The emergence of risk-sensitive learning requires generalization guarantees for functionals of the loss distribution beyond the…
View article: Many Ways to Be Lonely: Fine-Grained Characterization of Loneliness and Its Potential Changes in COVID-19
Many Ways to Be Lonely: Fine-Grained Characterization of Loneliness and Its Potential Changes in COVID-19 Open
Loneliness has been associated with negative outcomes for physical and mental health. Understanding how people express and cope with various forms of loneliness is critical for early screening and targeted interventions to reduce lonelines…
View article: A Taxonomy of Human and ML Strengths in Decision-Making to Investigate Human-ML Complementarity
A Taxonomy of Human and ML Strengths in Decision-Making to Investigate Human-ML Complementarity Open
Hybrid human-ML systems increasingly make consequential decisions in a wide range of domains. These systems are often introduced with the expectation that the combined human-ML system will achieve complementary performance, that is, the co…
View article: Modeling Attrition in Recommender Systems with Departing Bandits
Modeling Attrition in Recommender Systems with Departing Bandits Open
Traditionally, when recommender systems are formalized as multi-armed bandits, the policy of the recommender system influences the rewards accrued, but not the length of interaction. However, in real-world systems, dissatisfied users may d…
View article: Many Ways to Be Lonely: Fine-Grained Characterization of Loneliness and Its Potential Changes in COVID-19
Many Ways to Be Lonely: Fine-Grained Characterization of Loneliness and Its Potential Changes in COVID-19 Open
Loneliness has been associated with negative outcomes for physical and mental health. Understanding how people express and cope with various forms of loneliness is critical for early screening and targeted interventions to reduce lonelines…
View article: Off-Policy Risk Assessment in Contextual Bandits
Off-Policy Risk Assessment in Contextual Bandits Open
Even when unable to run experiments, practitioners can evaluate prospective policies, using previously logged data. However, while the bandits literature has adopted a diverse set of objectives, most research on off-policy evaluation to da…
View article: When curation becomes creation
When curation becomes creation Open
Algorithms, microcontent, and the vanishing distinction between platforms and creators.
View article: Action-Sufficient State Representation Learning for Control with Structural Constraints
Action-Sufficient State Representation Learning for Control with Structural Constraints Open
Perceived signals in real-world scenarios are usually high-dimensional and noisy, and finding and using their representation that contains essential and sufficient information required by downstream decision-making tasks will help improve …
View article: Analyzing Personality through Social Media Profile Picture Choice
Analyzing Personality through Social Media Profile Picture Choice Open
The content of images users post to their social media is driven in part by personality. In this study, we analyze how Twitter profile images vary with the personality of the users posting them. In our main analysis, we use profile images …
View article: When Curation Becomes Creation: Algorithms, Microcontent, and the Vanishing Distinction between Platforms and Creators
When Curation Becomes Creation: Algorithms, Microcontent, and the Vanishing Distinction between Platforms and Creators Open
Ever since social activity on the Internet began migrating from the wilds of the open web to the walled gardens erected by so-called platforms, debates have raged about the responsibilities that these platforms ought to bear. And yet, desp…