no code implementations • 11 Oct 2024 • Aldo Gael Carranza, Susan Athey
We consider the problem of using observational bandit feedback data from multiple heterogeneous data sources to learn a personalized decision policy that robustly generalizes across diverse target settings.
no code implementations • 15 Sep 2024 • Keyon Vafa, Susan Athey, David M. Blei
Classical methods for decomposing the wage gap employ simple predictive models of wages which condition on a small set of simple summaries of labor history.
no code implementations • 25 Jun 2024 • Susan Athey, Herman Brunborg, Tianyu Du, Ayush Kanodia, Keyon Vafa
We demonstrate the value of fine-tuning and further show that by adding more career data from a different population, fine-tuning smaller LLMs surpasses the performance of fine-tuning larger models.
no code implementations • 7 May 2024 • Sanath Kumar Krishnamurthy, Susan Athey, Emma Brunskill
However, for a class of estimate-estimand-error tuples, nontrivial high probability upper bounds on the maximum error often require class complexity as input -- limiting the practicality of such methods and often resulting in loose bounds.
no code implementations • 30 Apr 2024 • Susan Athey, Emil Palikot
Further, learners in the treatment group were 6\% more likely to report new employment within a year, with an 8\% increase in jobs related to their certificates.
no code implementations • 16 Oct 2023 • Keshav Agrawal, Susan Athey, Ayush Kanodia, Emil Palikot
In addition, we observed a 6\% increase in retention within the treatment group.
no code implementations • 12 Oct 2023 • Susan Athey, Niall Keleher, Jann Spiess
Nonetheless, targeting on low baseline outcomes is common in practice, for example because the relationship between individual characteristics and treatment effects is often difficult or impossible to estimate with historical data.
no code implementations • 13 Jul 2023 • Susan Athey, Lisa K. Simon, Oskar N. Skans, Johan Vikstrom, Yaroslav Yakymovych
Using generalized random forests and rich Swedish administrative data, we show that the earnings effects of job displacement due to establishment closures are extremely heterogeneous across and within (observable) worker types, establishments, and markets.
no code implementations • 21 May 2023 • Aldo Gael Carranza, Susan Athey
We consider the problem of learning personalized decision policies from observational bandit feedback data across multiple heterogeneous data sources.
1 code implementation • 4 Apr 2023 • Tianyu Du, Ayush Kanodia, Susan Athey
$\texttt{torch-choice}$ provides a $\texttt{ChoiceDataset}$ data structure to manage databases flexibly and memory-efficiently.
1 code implementation • 27 Jan 2023 • Damian Clarke, Daniel Pailañir, Susan Athey, Guido Imbens
In this paper, we describe a computational implementation of the Synthetic difference-in-differences (SDID) estimator of Arkhangelsky et al. (2021) for Stata.
no code implementations • 22 Nov 2022 • Susan Athey, Undral Byambadalai, Vitor Hadad, Sanath Kumar Krishnamurthy, Weiwen Leung, Joseph Jay Williams
We design and implement an adaptive experiment (a ``contextual bandit'') to learn a targeted treatment assignment policy, where the goal is to use a participant's survey responses to determine which charity to expose them to in a donation solicitation.
no code implementations • 18 Nov 2022 • Susan Athey, Emil Palikot
We estimate that Mentoring increases the probability of finding a technology job within four months from 29% to 42% and Challenges from 20% to 29%, and the treatment effects do not attenuate over 12 months.
no code implementations • 2 Sep 2022 • Susan Athey, Dean Karlan, Emil Palikot, Yuan Yuan
Online platforms often face challenges being both fair (i. e., non-discriminatory) and efficient (i. e., maximizing revenue).
no code implementations • 30 Aug 2022 • Keshav Agrawal, Susan Athey, Ayush Kanodia, Emil Palikot
We study the impact of personalized content recommendations on the usage of an educational app for children.
no code implementations • 21 Jun 2022 • Susan Athey, Kristen Grabarz, Michael Luca, Nils Wernerfelt
During the course of the COVID-19 pandemic, a common strategy for public health organizations around the world has been to launch interventions via advertising campaigns on social media.
no code implementations • 30 Mar 2022 • Aldo Gael Carranza, Sanath Kumar Krishnamurthy, Susan Athey
Contextual bandit algorithms often estimate reward models to inform decision-making.
1 code implementation • 16 Feb 2022 • Keyon Vafa, Emil Palikot, Tianyu Du, Ayush Kanodia, Susan Athey, David M. Blei
We fit CAREER to a dataset of 24 million job sequences from resumes, and adjust it on small longitudinal survey datasets.
1 code implementation • 25 Jul 2021 • Ruoxuan Xiong, Allison Koenecke, Michael Powell, Zhu Shen, Joshua T. Vogelstein, Susan Athey
We are interested in estimating the effect of a treatment applied to individuals at multiple sites, where data is stored locally for each site.
no code implementations • 11 Jun 2021 • Sanath Kumar Krishnamurthy, Adrienne Margaret Propp, Susan Athey
Our algorithm is based on a novel misspecification test, and our analysis demonstrates the benefits of using model selection for reward estimation.
1 code implementation • 3 Jun 2021 • Ruohan Zhan, Vitor Hadad, David A. Hirshberg, Susan Athey
In particular, when the pattern of treatment assignment in the collected data looks little like the pattern generated by the policy to be evaluated, the importance weights used in DR estimators explode, leading to excessive variance.
1 code implementation • 5 May 2021 • Ruohan Zhan, Zhimei Ren, Susan Athey, Zhengyuan Zhou
Learning optimal policies from historical data enables personalization in a wide variety of applications including healthcare, digital recommendations, and online education.
no code implementations • 26 Feb 2021 • Sanath Kumar Krishnamurthy, Vitor Hadad, Susan Athey
Computationally efficient contextual bandits are often based on estimating a predictive model of rewards given contexts and arms using past data.
no code implementations • 25 Oct 2020 • Sanath Kumar Krishnamurthy, Vitor Hadad, Susan Athey
When realizability does not hold, our algorithm ensures the same guarantees on regret achieved by realizability-based algorithms under realizability, up to an additive term that accounts for the misspecification error.
no code implementations • 21 Apr 2020 • Allison Koenecke, Michael Powell, Ruoxuan Xiong, Zhu Shen, Nicole Fischer, Sakibul Huq, Adham M. Khalafallah, Marco Trevisan, Pär Sparen, Juan J Carrero, Akihiko Nishimura, Brian Caffo, Elizabeth A. Stuart, Renyuan Bai, Verena Staedtke, David L. Thomas, Nickolas Papadopoulos, Kenneth W. Kinzler, Bert Vogelstein, Shibin Zhou, Chetan Bettegowda, Maximilian F. Konig, Brett Mensh, Joshua T. Vogelstein, Susan Athey
Here, we conducted retrospective analyses in two cohorts of patients with acute respiratory distress (ARD, n=18, 547) and three cohorts with pneumonia (n=400, 907).
no code implementations • 23 Feb 2020 • Sanath Kumar Krishnamurthy, Susan Athey
We consider a variant of the contextual bandit problem.
no code implementations • 31 Jan 2020 • Kun Kuang, Ruoxuan Xiong, Peng Cui, Susan Athey, Bo Li
Then, these weights are used in the weighted regression to improve the accuracy of estimation on the effect of each variable, thus help to improve the stability of prediction across unknown test data.
1 code implementation • 9 Nov 2019 • Ruoxuan Xiong, Susan Athey, Mohsen Bayati, Guido Imbens
Next, we study an adaptive experimental design problem, where both the decision to continue the experiment and treatment assignment decisions are updated after each period's data is collected.
1 code implementation • 7 Nov 2019 • Vitor Hadad, David A. Hirshberg, Ruohan Zhan, Stefan Wager, Susan Athey
In this context, typical estimators that use inverse propensity weighting to eliminate sampling bias can be problematic: their distributions become skewed and heavy-tailed as the propensity scores decay to zero.
2 code implementations • 5 Sep 2019 • Susan Athey, Guido Imbens, Jonas Metzger, Evan Munro
We discuss the use of Wasserstein Generative Adversarial Networks (WGANs) as a method for systematically generating artificial data that mimic closely any given real data set without the researcher having many degrees of freedom.
Econometrics Methodology
3 code implementations • 26 Aug 2019 • Jonathan Johannemann, Vitor Hadad, Susan Athey, Stefan Wager
Many learning algorithms require categorical data to be transformed into real vectors before it can be used as input.
1 code implementation • 6 Jun 2019 • Rob Donnelly, Francisco R. Ruiz, David Blei, Susan Athey
One source of the improvement is the ability of the model to accurately estimate heterogeneity in preferences (by pooling information across categories); another source of improvement is its ability to estimate the preferences of consumers who have rarely or never made a purchase in a given category in the training data.
no code implementations • 24 Mar 2019 • Susan Athey, Guido Imbens
We discuss the relevance of the recent Machine Learning (ML) literature for economics and econometrics.
no code implementations • 24 Mar 2019 • Susan Athey, Mohsen Bayati, Guido Imbens, Zhaonan Qu
This paper studies a panel data setting where the goal is to estimate causal effects of an intervention by predicting the counterfactual values of outcomes for treated units, had they not received the treatment.
2 code implementations • 20 Feb 2019 • Susan Athey, Stefan Wager
We apply causal forests to a dataset derived from the National Study of Learning Mindsets, and consider resulting practical and conceptual challenges.
Methodology
4 code implementations • 24 Dec 2018 • Dmitry Arkhangelsky, Susan Athey, David A. Hirshberg, Guido W. Imbens, Stefan Wager
We present a new estimator for causal effects with panel data that builds on insights behind the widely used difference in differences and synthetic control methods.
Methodology
no code implementations • 15 Dec 2018 • Maria Dimakopoulou, Zhengyuan Zhou, Susan Athey, Guido Imbens
Contextual bandit algorithms are sensitive to the estimation method of the outcome model as well as the exploration method used, particularly in the presence of rich heterogeneity or complex outcome models, which can lead to difficult estimation problems along the path of learning.
no code implementations • NeurIPS 2018 • Zhengyuan Zhou, Panayotis Mertikopoulos, Susan Athey, Nicholas Bambos, Peter W. Glynn, Yinyu Ye
We consider a game-theoretical multi-agent learning problem where the feedback information can be lost during the learning process and rewards are given by a broad class of games known as variationally stable games.
1 code implementation • 10 Oct 2018 • Zhengyuan Zhou, Susan Athey, Stefan Wager
In many settings, a decision-maker wishes to learn a rule, or policy, that maps from observable characteristics of an individual to an action.
no code implementations • 15 Aug 2018 • Susan Athey, Guido Imbens
In this paper we study estimation of and inference for average treatment effects in a setting with panel data.
3 code implementations • 30 Jul 2018 • Rina Friedberg, Julie Tibshirani, Susan Athey, Stefan Wager
Random forests are a powerful method for non-parametric regression, but are limited in their ability to fit smooth signals, and can show poor predictive performance in the presence of strong, smooth effects.
no code implementations • 16 Jun 2018 • Kun Kuang, Ruoxuan Xiong, Peng Cui, Susan Athey, Bo Li
In this paper, we propose a novel Deep Global Balancing Regression (DGBR) algorithm to jointly optimize a deep auto-encoder model for feature selection and a global balancing model for stable prediction across unknown environments.
no code implementations • 22 Jan 2018 • Susan Athey, David Blei, Robert Donnelly, Francisco Ruiz, Tobias Schmidt
The data is used to identify users' approximate typical morning location, as well as their choices of lunchtime restaurants.
1 code implementation • NeurIPS 2017 • Liping Liu, Francisco Ruiz, Susan Athey, David Blei
Embedding models consider the probability of a target observation (a word or an item) conditioned on the elements in the context (other words or items).
no code implementations • 19 Nov 2017 • Maria Dimakopoulou, Zhengyuan Zhou, Susan Athey, Guido Imbens
We develop parametric and non-parametric contextual bandits that integrate balancing methods from the causal inference literature in their estimation to make it less prone to problems of estimation bias.
2 code implementations • 9 Nov 2017 • Francisco J. R. Ruiz, Susan Athey, David M. Blei
We develop SHOPPER, a sequential probabilistic model of shopping data.
2 code implementations • 27 Oct 2017 • Susan Athey, Mohsen Bayati, Nikolay Doudchenko, Guido Imbens, Khashayar Khosravi
In this paper we study methods for estimating causal effects in settings with panel data, where some units are exposed to a treatment during some periods and the goal is estimating counterfactual (untreated) outcomes for the treated unit/period combinations.
Statistics Theory Econometrics Statistics Theory
1 code implementation • NeurIPS 2017 • Maja Rudolph, Francisco Ruiz, Susan Athey, David Blei
Here we develop structured exponential family embeddings (S-EFE), a method for discovering embeddings that vary across related groups of data.
no code implementations • 6 Jun 2017 • Alberto Abadie, Susan Athey, Guido W. Imbens, Jeffrey M. Wooldridge
We derive standard errors that account for design-based uncertainty instead of, or in addition to, sampling-based uncertainty.
Statistics Theory Econometrics Statistics Theory
1 code implementation • 9 Feb 2017 • Susan Athey, Stefan Wager
In many areas, practitioners seek to use observational data to learn a treatment assignment policy that satisfies application-specific constraints, such as budget, fairness, simplicity, or other functional form constraints.
5 code implementations • 5 Oct 2016 • Susan Athey, Julie Tibshirani, Stefan Wager
We propose generalized random forests, a method for non-parametric statistical estimation based on random forests (Breiman, 2001) that can be used to fit any quantity of interest identified as the solution to a set of local moment equations.
1 code implementation • 25 Apr 2016 • Susan Athey, Guido W. Imbens, Stefan Wager
There are many settings where researchers are interested in estimating average treatment effects and are willing to rely on the unconfoundedness assumption, which requires that the treatment assignment be as good as random conditional on pre-treatment variables.
Methodology Econometrics Statistics Theory Statistics Theory
no code implementations • 30 Mar 2016 • Susan Athey, Raj Chetty, Guido Imbens, Hyunseung Kang
We focus primarily on a setting with two samples, an experimental sample containing data about the treatment indicator and the surrogates and an observational sample containing information about the surrogates and the primary outcome.
6 code implementations • 14 Oct 2015 • Stefan Wager, Susan Athey
Many scientific and engineering challenges -- ranging from personalized medicine to customized marketing recommendations -- require an understanding of treatment effect heterogeneity.
1 code implementation • 5 Apr 2015 • Susan Athey, Guido Imbens
The challenge is that the "ground truth" for a causal effect is not observed for any individual unit: we observe the unit with the treatment, or without the treatment, but not both at the same time.