arXiv subjects

Judy Hanwen Shen

Publications and source records attributed to Judy Hanwen Shen.

At least 19 recordsLinked to original sources

arXiv2026

AI Organizations are More Effective but Less Aligned than Individual Agents

AI is increasingly deployed in multi-agent systems; however, most research considers only the behavior of individual models. We experimentally show that multi-agent "AI organizations" are simultaneously more effective at achieving business goals, but less aligned, than individual AI agents. We examine 12 tasks across two practical settings: an AI consultancy providing solutions to business problems and an AI software team developing software products. Across all settings, AI Organizations composed of aligned models produce solutions with higher utility but greater misalignment compared to a single aligned model. Our work demonstrates the importance of considering interacting systems of AI agents when doing both capabilities and safety research.

cs.AI↗

arXiv2026

Investigating Data Interventions for Subgroup Fairness: An ICU Case Study

In high-stakes settings where machine learning models are used to automate decision-making about individuals, the presence of algorithmic bias can exacerbate systemic harm to certain subgroups of people. These biases often stem from the underlying training data. In practice, interventions to "fix the data" depend on the actual additional data sources available -- where many are less than ideal. In these cases, the effects of data scaling on subgroup performance become volatile, as the improvements from increased sample size are counteracted by the introduction of distribution shifts in the training set. In this paper, we investigate the limitations of combining data sources to improve subgroup performance within the context of healthcare. Clinical models are commonly trained on datasets comprised of patient electronic health record (EHR) data from different hospitals or admission departments. Across two such datasets, the eICU Collaborative Research Database and the MIMIC-IV dataset, we find that data addition can both help and hurt model fairness and performance, and many intuitive strategies for data selection are unreliable. We compare model-based post-hoc calibration and data-centric addition strategies to find that the combination of both is important to improve subgroup performance. Our work questions the traditional dogma of "better data" for overcoming fairness challenges by comparing and combining data- and model-based approaches.

cs.LG↗

arXiv2026

Algorithms with Calibrated Machine Learning Predictions

The field of algorithms with predictions incorporates machine learning advice in the design of online algorithms to improve real-world performance. A central consideration is the extent to which predictions can be trusted -- while existing approaches often require users to specify an aggregate trust level, modern machine learning models can provide estimates of prediction-level uncertainty. In this paper, we propose calibration as a principled and practical tool to bridge this gap, demonstrating the benefits of calibrated advice through two case studies: the ski rental and online job scheduling problems. For ski rental, we design an algorithm that achieves near-optimal prediction-dependent performance and prove that, in high-variance settings, calibrated advice offers more effective guidance than alternative methods for uncertainty quantification. For job scheduling, we demonstrate that using a calibrated predictor leads to significant performance improvements over existing methods. Evaluations on real-world data validate our theoretical findings, highlighting the practical impact of calibration for algorithms with predictions.

stat.ML↗

arXiv2026

How AI Impacts Skill Formation

AI assistance produces significant productivity gains across professional domains, particularly for novice workers. Yet how this assistance affects the development of skills required to effectively supervise AI remains unclear. Novice workers who rely heavily on AI to complete unfamiliar tasks may compromise their own skill acquisition in the process. We conduct randomized experiments to study how developers gained mastery of a new asynchronous programming library with and without the assistance of AI. We find that AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average. Participants who fully delegated coding tasks showed some productivity improvements, but at the cost of learning the library. We identify six distinct AI interaction patterns, three of which involve cognitive engagement and preserve learning outcomes even when participants receive AI assistance. Our findings suggest that AI-enhanced productivity is not a shortcut to competence and AI assistance should be carefully adopted into workflows to preserve skill formation -- particularly in safety-critical domains.

cs.CY↗

arXiv2026

The Limits of AI Data Transparency Policy: Three Disclosure Fallacies

Data transparency has emerged as a rallying cry for addressing concerns about AI: data quality, privacy, and copyright chief among them. Yet while these calls are crucial for accountability, current transparency policies often fall short of their intended aims. Similar to nutrition facts for food, policies aimed at nutrition facts for AI currently suffer from a limited consideration of research on effective disclosures. We offer an institutional perspective and identify three common fallacies in policy implementations of data disclosures for AI. First, many data transparency proposals exhibit a specification gap between the stated goals of data transparency and the actual disclosures necessary to achieve such goals. Second, reform attempts exhibit an enforcement gap between required disclosures on paper and enforcement to ensure compliance in fact. Third, policy proposals manifest an impact gap between disclosed information and meaningful changes in developer practices and public understanding. Informed by the social science on transparency, our analysis identifies affirmative paths for transparency that are effective rather than merely symbolic.

cs.CY↗

arXiv2025

Disclosure and Evaluation as Fairness Interventions for General-Purpose AI

Despite conflicting definitions and conceptions of fairness, AI fairness researchers broadly agree that fairness is context-specific. However, when faced with general-purpose AI, which by definition serves a range of contexts, how should we think about fairness? We argue that while we cannot be prescriptive about what constitutes fair outcomes, we can specify the processes that different stakeholders should follow in service of fairness. Specifically, we consider the obligations of two major groups: system providers and system deployers. While system providers are natural candidates for regulatory attention, the current state of AI understanding offers limited insight into how upstream factors translate into downstream fairness impacts. Thus, we recommend that providers invest in evaluative research studying how model development decisions influence fairness and disclose whom they are serving their models to, or at the very least, reveal sufficient information for external researchers to conduct such research. On the other hand, system deployers are closer to real-world contexts and can leverage their proximity to end users to address fairness harms in different ways. Here, we argue they should responsibly disclose information about users and personalization and conduct rigorous evaluations across different levels of fairness. Overall, instead of focusing on enforcing fairness outcomes, we prioritize intentional information-gathering by system providers and deployers that can facilitate later context-aware action. This allows us to be specific and concrete about the processes even while the contexts remain unknown. Ultimately, this approach can sharpen how we distribute fairness responsibilities and inform more fluid, context-sensitive interventions as AI continues to advance.

cs.CY↗

arXiv2025

Societal Impacts Research Requires Benchmarks for Creative Composition Tasks

Foundation models that are capable of automating cognitive tasks represent a pivotal technological shift, yet their societal implications remain unclear. These systems promise exciting advances, yet they also risk flooding our information ecosystem with formulaic, homogeneous, and potentially misleading synthetic content. Developing benchmarks grounded in real use cases where these risks are most significant is therefore critical. Through a thematic analysis using 2 million language model user prompts, we identify creative composition tasks as a prevalent usage category where users seek help with personal tasks that require everyday creativity. Our fine-grained analysis identifies mismatches between current benchmarks and usage patterns among these tasks. Crucially, we argue that the same use cases that currently lack thorough evaluations can lead to negative downstream impacts. This position paper argues that benchmarks focused on creative composition tasks is a necessary step towards understanding the societal harms of AI-generated content. We call for greater transparency in usage patterns to inform the development of new benchmarks that can effectively measure both the progress and the impacts of models with creative capabilities.

cs.CY↗

arXiv2025

Two Tickets are Better than One: Fair and Accurate Hiring Under Strategic LLM Manipulations

In an era of increasingly capable foundation models, job seekers are turning to generative AI tools to enhance their application materials. However, unequal access to and knowledge about generative AI tools can harm both employers and candidates by reducing the accuracy of hiring decisions and giving some candidates an unfair advantage. To address these challenges, we introduce a new variant of the strategic classification framework tailored to manipulations performed using large language models, accommodating varying levels of manipulations and stochastic outcomes. We propose a ``two-ticket'' scheme, where the hiring algorithm applies an additional manipulation to each submitted resume and considers this manipulated version together with the original submitted resume. We establish theoretical guarantees for this scheme, showing improvements for both the fairness and accuracy of hiring decisions when the true positive rate is maximized subject to a no false positives constraint. We further generalize this approach to an $n$-ticket scheme and prove that hiring outcomes converge to a fixed, group-independent decision, eliminating disparities arising from differential LLM access. Finally, we empirically validate our framework and the performance of our two-ticket scheme on real resumes using an open-source resume screening tool.

cs.LG↗

arXiv2024

Towards Data-Centric RLHF: Simple Metrics for Preference Dataset Comparison

The goal of aligning language models to human preferences requires data that reveal these preferences. Ideally, time and money can be spent carefully collecting and tailoring bespoke preference data to each downstream application. However, in practice, a select few publicly available preference datasets are often used to train reward models for reinforcement learning from human feedback (RLHF). While new preference datasets are being introduced with increasing frequency, there are currently no existing efforts to measure and compare these datasets. In this paper, we systematically study preference datasets through three perspectives: scale, label noise, and information content. We propose specific metrics for each of these perspectives and uncover different axes of comparison for a better understanding of preference datasets. Our work is a first step towards a data-centric approach to alignment by providing perspectives that aid in training efficiency and iterative data collection for RLHF.

cs.AI↗

arXiv2024

Dissenting Explanations: Leveraging Disagreement to Reduce Model Overreliance

While explainability is a desirable characteristic of increasingly complex black-box models, modern explanation methods have been shown to be inconsistent and contradictory. The semantics of explanations is not always fully understood - to what extent do explanations "explain" a decision and to what extent do they merely advocate for a decision? Can we help humans gain insights from explanations accompanying correct predictions and not over-rely on incorrect predictions advocated for by explanations? With this perspective in mind, we introduce the notion of dissenting explanations: conflicting predictions with accompanying explanations. We first explore the advantage of dissenting explanations in the setting of model multiplicity, where multiple models with similar performance may have different predictions. In such cases, providing dissenting explanations could be done by invoking the explanations of disagreeing models. Through a pilot study, we demonstrate that dissenting explanations reduce overreliance on model predictions, without reducing overall accuracy. Motivated by the utility of dissenting explanations we present both global and local methods for their generation.

cs.AI↗

arXiv2024

The Data Addition Dilemma

In many machine learning for healthcare tasks, standard datasets are constructed by amassing data across many, often fundamentally dissimilar, sources. But when does adding more data help, and when does it hinder progress on desired model outcomes in real-world settings? We identify this situation as the \textit{Data Addition Dilemma}, demonstrating that adding training data in this multi-source scaling context can at times result in reduced overall accuracy, uncertain fairness outcomes, and reduced worst-subgroup performance. We find that this possibly arises from an empirically observed trade-off between model performance improvements due to data scaling and model deterioration from distribution shift. We thus establish baseline strategies for navigating this dilemma, introducing distribution shift heuristics to guide decision-making on which data sources to add in data scaling, in order to yield the expected model performance improvements. We conclude with a discussion of the required considerations for data collection and suggestions for studying data composition and scale in the age of increasingly larger models.

cs.LG↗

arXiv2024

Multigroup Robustness

To address the shortcomings of real-world datasets, robust learning algorithms have been designed to overcome arbitrary and indiscriminate data corruption. However, practical processes of gathering data may lead to patterns of data corruption that are localized to specific partitions of the training dataset. Motivated by critical applications where the learned model is deployed to make predictions about people from a rich collection of overlapping subpopulations, we initiate the study of multigroup robust algorithms whose robustness guarantees for each subpopulation only degrade with the amount of data corruption inside that subpopulation. When the data corruption is not distributed uniformly over subpopulations, our algorithms provide more meaningful robustness guarantees than standard guarantees that are oblivious to how the data corruption and the affected subpopulations are related. Our techniques establish a new connection between multigroup fairness and robustness.

cs.LG↗

arXiv2023

Unifying Corroborative and Contributive Attributions in Large Language Models

As businesses, products, and services spring up around large language models, the trustworthiness of these models hinges on the verifiability of their outputs. However, methods for explaining language model outputs largely fall across two distinct fields of study which both use the term "attribution" to refer to entirely separate techniques: citation generation and training data attribution. In many modern applications, such as legal document generation and medical question answering, both types of attributions are important. In this work, we argue for and present a unified framework of large language model attributions. We show how existing methods of different types of attribution fall under the unified framework. We also use the framework to discuss real-world use cases where one or both types of attributions are required. We believe that this unified framework will guide the use case driven development of systems that leverage both types of attribution, as well as the standardization of their evaluation.

cs.CL↗

arXiv2023

Unlocking Accuracy and Fairness in Differentially Private Image Classification

Privacy-preserving machine learning aims to train models on private data without leaking sensitive information. Differential privacy (DP) is considered the gold standard framework for privacy-preserving training, as it provides formal privacy guarantees. However, compared to their non-private counterparts, models trained with DP often have significantly reduced accuracy. Private classifiers are also believed to exhibit larger performance disparities across subpopulations, raising fairness concerns. The poor performance of classifiers trained with DP has prevented the widespread adoption of privacy preserving machine learning in industry. Here we show that pre-trained foundation models fine-tuned with DP can achieve similar accuracy to non-private classifiers, even in the presence of significant distribution shifts between pre-training data and downstream tasks. We achieve private accuracies within a few percent of the non-private state of the art across four datasets, including two medical imaging benchmarks. Furthermore, our private medical classifiers do not exhibit larger performance disparities across demographic groups than non-private models. This milestone to make DP training a practical and reliable technology has the potential to widely enable machine learning practitioners to train safely on sensitive datasets while protecting individuals' privacy.

cs.LG↗

arXiv2023

Bidding Strategies for Proportional Representation in Advertisement Campaigns

Many companies rely on advertising platforms such as Google, Facebook, or Instagram to recruit a large and diverse applicant pool for job openings. Prior works have shown that equitable bidding may not result in equitable outcomes due to heterogeneous levels of competition for different types of individuals. Suggestions have been made to address this problem via revisions to the advertising platform. However, it may be challenging to convince platforms to undergo a costly re-vamp of their system, and in addition it might not offer the flexibility necessary to capture the many types of fairness notions and other constraints that advertisers would like to ensure. Instead, we consider alterations that make no change to the platform mechanism and instead change the bidding strategies used by advertisers. We compare two natural fairness objectives: one in which the advertisers must treat groups equally when bidding in order to achieve a yield with group-parity guarantees, and another in which the bids are not constrained and only the yield must satisfy parity constraints. We show that requiring parity with respect to both bids and yield can result in an arbitrarily large decrease in efficiency compared to requiring equal yield proportions alone. We find that autobidding is a natural way to realize this latter objective and show how existing work in this area can be extended to provide efficient bidding strategies that provide high utility while satisfying group parity constraints as well as deterministic and randomized rounding techniques to uphold these guarantees. Finally, we demonstrate the effectiveness of our proposed solutions on data adapted from a real-world employment dataset.

cs.CY↗

arXiv2022

Leximax Approximations and Representative Cohort Selection

Finding a representative cohort from a broad pool of candidates is a goal that arises in many contexts such as choosing governing committees and consumer panels. While there are many ways to define the degree to which a cohort represents a population, a very appealing solution concept is lexicographic maximality (leximax) which offers a natural (pareto-optimal like) interpretation that the utility of no population can be increased without decreasing the utility of a population that is already worse off. However, finding a leximax solution can be highly dependent on small variations in the utility of certain groups. In this work, we explore new notions of approximate leximax solutions with three distinct motivations: better algorithmic efficiency, exploiting significant utility improvements, and robustness to noise. Among other definitional contributions, we give a new notion of an approximate leximax that satisfies a similarly appealing semantic interpretation and relate it to algorithmically-feasible approximate leximax notions. When group utilities are linear over cohort candidates, we give an efficient polynomial-time algorithm for finding a leximax distribution over cohort candidates in the exact as well as in the approximate setting. Furthermore, we show that finding an integer solution to leximax cohort selection with linear utilities is NP-Hard.

cs.DS↗

arXiv2022

Longitudinal Complex Dynamics of Labour Markets Reveal Increasing Polarisation

In this paper we conduct a longitudinal analysis of the structure of labour markets in the US over 7 decades of technological, economic and policy change. We make use of network science, natural language processing and machine learning to uncover structural changes in the labour market over time. We find a steady rate of both disappearance of jobs and a shift in the required work tasks, despite much technological and economic change over this time period. Machine learning is used to classify jobs as being predominantly cognitive or physical based on the textual description of the workplace tasks. We also measure increasing polarisation between these two classes of jobs, linked by the similarity of tasks, over time that could constrain workers wishing to move to different jobs.

cs.CY↗

arXiv2022

Differentially Private Set Union

We study the basic operation of set union in the global model of differential privacy. In this problem, we are given a universe $U$ of items, possibly of infinite size, and a database $D$ of users. Each user $i$ contributes a subset $W_i \subseteq U$ of items. We want an ($ε$,$δ$)-differentially private algorithm which outputs a subset $S \subset \cup_i W_i$ such that the size of $S$ is as large as possible. The problem arises in countless real world applications; it is particularly ubiquitous in natural language processing (NLP) applications as vocabulary extraction. For example, discovering words, sentences, $n$-grams etc., from private text data belonging to users is an instance of the set union problem. Known algorithms for this problem proceed by collecting a subset of items from each user, taking the union of such subsets, and disclosing the items whose noisy counts fall above a certain threshold. Crucially, in the above process, the contribution of each individual user is always independent of the items held by other users, resulting in a wasteful aggregation process, where some item counts happen to be way above the threshold. We deviate from the above paradigm by allowing users to contribute their items in a $\textit{dependent fashion}$, guided by a $\textit{policy}$. In this new setting ensuring privacy is significantly delicate. We prove that any policy which has certain $\textit{contractive}$ properties would result in a differentially private algorithm. We design two new algorithms, one using Laplace noise and other Gaussian noise, as specific instances of policies satisfying the contractive properties. Our experiments show that the new algorithms significantly outperform previously known mechanisms for the problem.

cs.CR↗