no code implementations • CL (ACL) 2021 • Lifeng Jin, Lane Schwartz, Finale Doshi-Velez, Timothy Miller, William Schuler
Abstract This article describes a simple PCFG induction model with a fixed category domain that predicts a large majority of attested constituent boundaries, and predicts labels consistent with nearly half of attested constituent labels on a standard evaluation data set of child-directed speech.
no code implementations • 13 Mar 2024 • Yaniv Yacoby, Weiwei Pan, Finale Doshi-Velez
It approximates the posterior of the true model a priori; fixing this posterior approximation, we then maximize the lower bound relative to only the generative model.
no code implementations • 26 Feb 2024 • Anna L. Trella, Kelly W. Zhang, Inbal Nahum-Shani, Vivek Shetty, Iris Yan, Finale Doshi-Velez, Susan A. Murphy
This paper proposes algorithm fidelity as a critical requirement for deploying online RL algorithms in clinical trials.
1 code implementation • 20 Feb 2024 • Marton Havasi, Sonali Parbhoo, Finale Doshi-Velez
Interpretability methods that utilise local surrogate models (e. g. LIME) are very good at describing the behaviour of the predictive model at a point of interest, but they are not guaranteed to extrapolate to the local region surrounding the point.
1 code implementation • 5 Feb 2024 • Anna L. Trella, Walter Dempsey, Finale Doshi-Velez, Susan A. Murphy
We consider the stochastic multi-armed bandit problem with non-stationary rewards.
no code implementations • 29 Jan 2024 • Yidou Weng, Finale Doshi-Velez
This paper proposes a model learning Semi-parametric rela- tionships in an Expert Bayesian Network (SEBN) with linear parameter and structure constraints.
no code implementations • 26 Jan 2024 • Eura Nofshin, Siddharth Swaroop, Weiwei Pan, Susan Murphy, Finale Doshi-Velez
Many important behavior changes are frictionful; they require individuals to expend effort over a long period with little immediate gratification.
no code implementations • 15 Dec 2023 • Lauren H. Cooke, Harvey Klyne, Edwin Zhang, Cassidy Laidlaw, Milind Tambe, Finale Doshi-Velez
Inverse reinforcement learning (IRL) is computationally challenging, with common approaches requiring the solution of multiple reinforcement learning (RL) sub-problems.
1 code implementation • 20 Sep 2023 • Jose Roberto Tello Ayala, Akl C. Fahed, Weiwei Pan, Eugene V. Pomerantsev, Patrick T. Ellinor, Anthony Philippakis, Finale Doshi-Velez
The adoption of machine learning in healthcare calls for model transparency and explainability.
no code implementations • 1 Sep 2023 • Varshini Subhash, Anna Bialas, Weiwei Pan, Finale Doshi-Velez
We believe this new geometric perspective on the underlying mechanism driving universal attacks could help us gain deeper insight into the internal workings and failure modes of LLMs, thus enabling their mitigation.
no code implementations • 9 Aug 2023 • Leo Benac, Sonali Parbhoo, Finale Doshi-Velez
Offline Reinforcement learning is commonly used for sequential decision-making in domains such as healthcare and education, where the rewards are known and the transition dynamics $T$ must be estimated on the basis of batch data.
no code implementations • 28 Jul 2023 • Charumathi Badrinath, Weiwei Pan, Finale Doshi-Velez
A common way to explore text corpora is through low-dimensional projections of the documents, where one hopes that thematically similar documents will be clustered together in the projected space.
no code implementations • 13 Jul 2023 • Yiqing Xu, Finale Doshi-Velez, David Hsu
Inverse reinforcement learning (IRL) algorithms often rely on (forward) reinforcement learning or planning over a given time horizon to compute an approximately optimal policy for a hypothesized reward function and then match this policy with expert demonstrations.
no code implementations • 22 Jun 2023 • Xudong Shen, Hannah Brown, Jiashu Tao, Martin Strobel, Yao Tong, Akshay Narayan, Harold Soh, Finale Doshi-Velez
There is increasing attention being given to how to regulate AI systems.
no code implementations • 20 Jun 2023 • Sarah Rathnam, Sonali Parbhoo, Weiwei Pan, Susan A. Murphy, Finale Doshi-Velez
We demonstrate that planning under a lower discount factor produces an identical optimal policy to planning using any prior on the transition matrix that has the same distribution for all states and actions.
no code implementations • 12 Jun 2023 • Siddharth Swaroop, Zana Buçinca, Krzysztof Z. Gajos, Finale Doshi-Velez
The precise benefit can depend on both the user and task.
2 code implementations • 2 May 2023 • Shengpu Tang, Maggie Makar, Michael W. Sjoding, Finale Doshi-Velez, Jenna Wiens
We study the theoretical properties of our approach, identifying scenarios where it is guaranteed to lead to zero bias when used to approximate the Q-function.
no code implementations • 6 Apr 2023 • Abhishek Sharma, Sonali Parbhoo, Omer Gottesman, Finale Doshi-Velez
Decision-focused (DF) model-based reinforcement learning has recently been introduced as a powerful algorithm that can focus on learning the MDP dynamics that are most relevant for obtaining high returns.
no code implementations • 1 Dec 2022 • Eura Shin, Siddharth Swaroop, Weiwei Pan, Susan Murphy, Finale Doshi-Velez
Mobile health (mHealth) technologies empower patients to adopt/maintain healthy behaviors in their daily lives, by providing interventions (e. g. push notifications) tailored to the user's needs.
no code implementations • 16 Nov 2022 • Jiayu Yao, Yaniv Yacoby, Beau Coker, Weiwei Pan, Finale Doshi-Velez
Comparing Bayesian neural networks (BNNs) with different widths is challenging because, as the width increases, multiple model properties change simultaneously, and, inference in the finite-width case is intractable.
no code implementations • 14 Nov 2022 • Sanjana Narayanan, Isaac Lage, Finale Doshi-Velez
We find that complete explanations are generally more effective when they are the same size or smaller than a contrastive explanation of the same policy, and no worse when they are larger.
no code implementations • 10 Nov 2022 • Zixi Chen, Varshini Subhash, Marton Havasi, Weiwei Pan, Finale Doshi-Velez
In this work, we survey properties defined in interpretable machine learning papers, synthesize them based on what they actually measure, and describe the trade-offs between different formulations of these properties.
no code implementations • 27 Oct 2022 • Michael L. Littman, Ifeoma Ajunwa, Guy Berger, Craig Boutilier, Morgan Currie, Finale Doshi-Velez, Gillian Hadfield, Michael C. Horowitz, Charles Isbell, Hiroaki Kitano, Karen Levy, Terah Lyons, Melanie Mitchell, Julie Shah, Steven Sloman, Shannon Vallor, Toby Walsh
In September 2021, the "One Hundred Year Study on Artificial Intelligence" project (AI100) issued the second report of its planned long-term periodic assessment of artificial intelligence (AI) and its impact on society.
no code implementations • 18 Sep 2022 • Harvineet Singh, Shalmali Joshi, Finale Doshi-Velez, Himabindu Lakkaraju
When deployment environments are expected to undergo changes (that is, dataset shifts), it is important for OPE methods to perform robust evaluation of the policies amidst such changes.
1 code implementation • 15 Aug 2022 • Anna L. Trella, Kelly W. Zhang, Inbal Nahum-Shani, Vivek Shetty, Finale Doshi-Velez, Susan A. Murphy
Dental disease is one of the most common chronic diseases despite being largely preventable.
no code implementations • 2 Aug 2022 • Mark Penrod, Harrison Termotto, Varshini Reddy, Jiayu Yao, Finale Doshi-Velez, Weiwei Pan
For responsible decision making in safety-critical settings, machine learning models must effectively detect and process edge-case data.
no code implementations • 30 Jul 2022 • Kelly W. Zhang, Omer Gottesman, Finale Doshi-Velez
In the reinforcement learning literature, there are many algorithms developed for either Contextual Bandit (CB) or Markov Decision Processes (MDP) environments.
no code implementations • 13 Jul 2022 • Jiayu Yao, Sonali Parbhoo, Weiwei Pan, Finale Doshi-Velez
We develop a Reinforcement Learning (RL) framework for improving an existing behavior policy via sparse, user-interpretable changes.
no code implementations • 22 Jun 2022 • Q. Vera Liao, Yunfeng Zhang, Ronny Luss, Finale Doshi-Velez, Amit Dhurandhar
We argue that one way to close the gap is to develop evaluation methods that account for different user requirements in these usage contexts.
1 code implementation • 8 Jun 2022 • Anna L. Trella, Kelly W. Zhang, Inbal Nahum-Shani, Vivek Shetty, Finale Doshi-Velez, Susan A. Murphy
Online reinforcement learning (RL) algorithms are increasingly used to personalize digital interventions in the fields of mobile health and online education.
no code implementations • spnlp (ACL) 2022 • Jeffrey Chiu, Rajat Mittal, Neehal Tumma, Abhishek Sharma, Finale Doshi-Velez
Topic models are some of the most popular ways to represent textual data in an interpret-able manner.
1 code implementation • 23 Feb 2022 • Beau Coker, Wessel P. Bruinsma, David R. Burt, Weiwei Pan, Finale Doshi-Velez
Finally, we show that the optimal approximate posterior need not tend to the prior if the activation function is not odd, showing that our statements cannot be generalized arbitrarily.
no code implementations • 20 Jan 2022 • Sonali Parbhoo, Shalmali Joshi, Finale Doshi-Velez
A precise description of the causal estimand highlights which OPE estimands are identifiable from observational data under the stated generative assumptions.
no code implementations • 28 Nov 2021 • Ramtin Keramati, Omer Gottesman, Leo Anthony Celi, Finale Doshi-Velez, Emma Brunskill
Off-policy policy evaluation methods for sequential decision making can be used to help identify if a proposed decision policy is better than a current baseline policy.
no code implementations • 25 Oct 2021 • Abhishek Sharma, Catherine Zeng, Sanjana Narayanan, Sonali Parbhoo, Finale Doshi-Velez
Probabilistic models help us encode latent structures that both model the data and are ideally also useful for specific downstream tasks.
no code implementations • 22 Sep 2021 • Nari Johnson, Sonali Parbhoo, Andrew Slavin Ross, Finale Doshi-Velez
Machine learning models that utilize patient data across time (rather than just the most recent measurements) have increased performance for many risk stratification tasks in the intensive care unit.
no code implementations • 16 Sep 2021 • Sarah Rathnam, Susan A. Murphy, Finale Doshi-Velez
In batch reinforcement learning, there can be poorly explored state-action pairs resulting in poorly learned, inaccurate models and poorly performing associated policies.
1 code implementation • 13 Sep 2021 • Simon P. Shen, Yecheng Jason Ma, Omer Gottesman, Finale Doshi-Velez
Importance sampling-based estimators for off-policy evaluation (OPE) are valued for their simplicity, unbiasedness, and reliance on relatively few assumptions.
no code implementations • 13 Sep 2021 • Shalmali Joshi, Sonali Parbhoo, Finale Doshi-Velez
Our deferral policy is adaptive to the non-stationarity in the dynamics.
no code implementations • 21 Jul 2021 • Eura Shin, Pedja Klasnja, Susan Murphy, Finale Doshi-Velez
Motivated by the need for efficient and personalized learning in mobile health, we investigate the problem of online kernel selection for Gaussian Process regression in the multi-task setting.
1 code implementation • 24 Jun 2021 • Anita Mahinpei, Justin Clark, Isaac Lage, Finale Doshi-Velez, Weiwei Pan
Machine learning models that incorporate concept learning as an intermediate step in their decision making process can match the performance of black-box predictive models while retaining the ability to explain outcomes in human understandable terms.
no code implementations • 13 Jun 2021 • Beau Coker, Weiwei Pan, Finale Doshi-Velez
Variational inference enables approximate posterior inference of the highly over-parameterized neural networks that are popular in modern machine learning.
no code implementations • NeurIPS 2021 • Kai Wang, Sanket Shah, Haipeng Chen, Andrew Perrault, Finale Doshi-Velez, Milind Tambe
In the predict-then-optimize framework, the objective is to train a predictive model, mapping from environment features to parameters of an optimization problem, which maximizes decision quality when the optimization is subsequently solved.
no code implementations • NeurIPS 2021 • Kai Wang, Sanket Shah, Haipeng Chen, Andrew Perrault, Finale Doshi-Velez, Milind Tambe
In the predict-then-optimize framework, the objective is to train a predictive model, mapping from environment features to parameters of an optimization problem, which maximizes decision quality when the optimization is subsequently solved.
no code implementations • 29 Mar 2021 • Harvineet Singh, Shalmali Joshi, Finale Doshi-Velez, Himabindu Lakkaraju
Most of the existing work focuses on optimizing for either adversarial shifts or interventional shifts.
1 code implementation • 9 Feb 2021 • Andrew Slavin Ross, Finale Doshi-Velez
In representation learning, there has been recent interest in developing algorithms to disentangle the ground-truth generative factors behind a dataset, and metrics to quantify how fully this occurs.
1 code implementation • 2 Feb 2021 • Andrew Slavin Ross, Nina Chen, Elisa Zhao Hang, Elena L. Glassman, Finale Doshi-Velez
On synthetic datasets, we find performance on this task much more reliably differentiates entangled and disentangled models than baseline approaches.
no code implementations • 13 Jan 2021 • Melanie F. Pradier, Javier Zazo, Sonali Parbhoo, Roy H. Perlis, Maurizio Zazzi, Finale Doshi-Velez
We propose Preferential MoE, a novel human-ML mixture-of-experts model that augments human expertise in decision making with a data-based classifier only when necessary for predictive performance.
no code implementations • 9 Jan 2021 • Kristine Zhang, Yuanheng Wang, Jianzhun Du, Brian Chu, Leo Anthony Celi, Ryan Kindle, Finale Doshi-Velez
Many batch RL health applications first discretize time into fixed intervals.
no code implementations • 10 Dec 2020 • Elisa Bertino, Finale Doshi-Velez, Maria Gini, Daniel Lopresti, David Parkes
There is a vital need for research in "AI and Cooperation" that seeks to understand the ways in which systems of AIs and systems of AIs with people can engender cooperative behavior.
no code implementations • 4 Dec 2020 • Isaac Lage, Finale Doshi-Velez
These limitations are particularly acute for high-dimensional tabular features.
1 code implementation • NeurIPS 2020 • Wanqian Yang, Lars Lorch, Moritz A. Graule, Himabindu Lakkaraju, Finale Doshi-Velez
Domains where supervised models are deployed often come with task-specific constraints, such as prior expert knowledge on the ground-truth function, or desiderata like safety and fairness.
no code implementations • 14 Jul 2020 • Yaniv Yacoby, Weiwei Pan, Finale Doshi-Velez
Variational Auto-encoders (VAEs) are deep generative latent variable models that are widely used for a number of downstream tasks.
no code implementations • 12 Jul 2020 • Théo Guénais, Dimitris Vamvourellis, Yaniv Yacoby, Finale Doshi-Velez, Weiwei Pan
Traditional training of deep classifiers yields overconfident models that are not reliable under dataset shift.
1 code implementation • NeurIPS 2020 • Jianzhun Du, Joseph Futoma, Finale Doshi-Velez
We present two elegant solutions for modeling continuous-time dynamics, in a novel model-based reinforcement learning (RL) framework for semi-Markov decision processes (SMDPs), using neural ordinary differential equations (ODEs).
Model-based Reinforcement Learning reinforcement-learning +1
no code implementations • 21 Jun 2020 • Sujay Thakur, Cooper Lorsung, Yaniv Yacoby, Finale Doshi-Velez, Weiwei Pan
Neural Linear Models (NLM) are deep Bayesian models that produce predictive uncertainties by learning features from the data and then performing Bayesian linear regression over these features.
no code implementations • 11 Jun 2020 • Yash Nair, Finale Doshi-Velez
First, we derive sample complexity bounds for DPL, and then show that model-based learning from expert actions can, even with a finite model class, be impossible.
no code implementations • 8 May 2020 • MingYu Lu, Zachary Shahn, Daby Sow, Finale Doshi-Velez, Li-wei H. Lehman
The potential of Reinforcement Learning (RL) has been demonstrated through successful applications to games such as Go and Atari.
1 code implementation • 13 Apr 2020 • Jiayu Yao, Emma Brunskill, Weiwei Pan, Susan Murphy, Finale Doshi-Velez
However, when bandits are deployed in the context of a scientific study -- e. g. a clinical trial to test if a mobile health intervention is effective -- the aim is not only to personalize for an individual, but also to determine, with sufficient statistical power, whether or not the system's intervention is effective.
no code implementations • pproximateinference AABI Symposium 2019 • Yaniv Yacoby, Weiwei Pan, Finale Doshi-Velez
Recent work shows that traditional training methods tend to yield solutions that violate modeling desiderata: (1) the learned generative model captures the observed data distribution but does so while ignoring the latent codes, resulting in codes that do not represent the data (e. g. van den Oord et al. (2017); Kim et al. (2018)); (2) the aggregate of the learned latent codes does not match the prior p(z).
no code implementations • ICML 2020 • Omer Gottesman, Joseph Futoma, Yao Liu, Sonali Parbhoo, Leo Anthony Celi, Emma Brunskill, Finale Doshi-Velez
Off-policy evaluation in reinforcement learning offers the chance of using observational data to improve future outcomes in domains such as healthcare and education, but safe deployment in high stakes settings requires ways of assessing its validity.
1 code implementation • 13 Jan 2020 • Joseph Futoma, Michael C. Hughes, Finale Doshi-Velez
Many medical decision-making tasks can be framed as partially observed Markov decision processes (POMDPs).
no code implementations • 9 Jan 2020 • Joseph Futoma, Muhammad A. Masood, Finale Doshi-Velez
Hypotension in critical care settings is a life-threatening emergency that must be recognized and treated early.
no code implementations • 12 Dec 2019 • Beau Coker, Melanie F. Pradier, Finale Doshi-Velez
While Bayesian neural networks have many appealing characteristics, current priors do not easily allow users to specify basic properties such as expected lengthscale or amplitude variance.
no code implementations • 15 Nov 2019 • Jason Ren, Russell Kunes, Finale Doshi-Velez
Supervised topic models present an attractive option for incorporating EHR data as features into a prediction problem: given a patient's record, we estimate a set of latent factors that are predictive of the response variable.
1 code implementation • 4 Nov 2019 • Andrew Slavin Ross, Weiwei Pan, Leo Anthony Celi, Finale Doshi-Velez
Ensembles depend on diversity for improved performance.
no code implementations • 1 Nov 2019 • Yaniv Yacoby, Weiwei Pan, Finale Doshi-Velez
Bayesian Neural Networks with Latent Variables (BNN+LVs) capture predictive uncertainty by explicitly modeling model uncertainty (via priors on network weights) and environmental stochasticity (via a latent input noise variable).
no code implementations • pproximateinference AABI Symposium 2019 • Melanie F. Pradier, Michael C. Hughes, Finale Doshi-Velez
Variational inference based on chi-square divergence minimization (CHIVI) provides a way to approximate a model's posterior while obtaining an upper bound on the marginal likelihood.
1 code implementation • 12 Oct 2019 • Jason Ren, Russell Kunes, Finale Doshi-Velez
Supervised topic models are often sought to balance prediction quality and interpretability.
no code implementations • 14 Aug 2019 • Mike Wu, Sonali Parbhoo, Michael C. Hughes, Volker Roth, Finale Doshi-Velez
Moreover, for situations in which a single, global tree is a poor estimator, we introduce a regional tree regularizer that encourages the deep model to resemble a compact, axis-aligned decision tree in predefined, human-interpretable contexts.
no code implementations • 13 Aug 2019 • Mike Wu, Sonali Parbhoo, Michael Hughes, Ryan Kindle, Leo Celi, Maurizio Zazzi, Volker Roth, Finale Doshi-Velez
The lack of interpretability remains a barrier to the adoption of deep neural networks.
no code implementations • ACL 2019 • Lifeng Jin, Finale Doshi-Velez, Timothy Miller, Lane Schwartz, William Schuler
This paper describes a neural PCFG inducer which employs context embeddings (Peters et al., 2018) in a normalizing flow model (Dinh et al., 2015) to extend PCFG induction to use semantic and morphological information.
no code implementations • 24 Jun 2019 • Jiayu Yao, Weiwei Pan, Soumya Ghosh, Finale Doshi-Velez
Bayesian Neural Networks (BNNs) place priors over the parameters in a neural network.
no code implementations • 31 May 2019 • Muhammad A. Masood, Finale Doshi-Velez
Standard reinforcement learning methods aim to master one way of solving a task whereas there may exist multiple near-optimal policies.
no code implementations • 30 May 2019 • Niranjani Prasad, Barbara E. Engelhardt, Finale Doshi-Velez
A key impediment to reinforcement learning (RL) in real applications with limited, batch data is defining a reward function that reflects what we implicitly know about reasonable behaviour for a task and allows for robust off-policy evaluation.
1 code implementation • 30 May 2019 • Isaac Lage, Daphna Lifschitz, Finale Doshi-Velez, Ofra Amir
We introduce an imitation learning-based approach to policy summarization; we demonstrate through computational simulations that a mismatch between the model used to extract a summary and the model used to reconstruct the policy results in worse reconstruction quality; and we demonstrate through a human-subject study that people use different models to reconstruct policies in different contexts, and that matching the summary extraction model to these can improve performance.
no code implementations • 24 May 2019 • Omer Gottesman, Weiwei Pan, Finale Doshi-Velez
Tensor decomposition methods allow us to learn the parameters of latent variable models through decomposition of low-order moments of data.
1 code implementation • 15 May 2019 • Wanqian Yang, Lars Lorch, Moritz A. Graule, Srivatsan Srinivasan, Anirudh Suresh, Jiayu Yao, Melanie F. Pradier, Finale Doshi-Velez
Bayesian neural network (BNN) priors are defined in parameter space, making it hard to encode prior knowledge expressed in function space.
no code implementations • 14 May 2019 • Omer Gottesman, Yao Liu, Scott Sussex, Emma Brunskill, Finale Doshi-Velez
We consider a model-based approach to perform batch off-policy evaluation in reinforcement learning.
no code implementations • 24 Mar 2019 • Dong-hun Lee, Srivatsan Srinivasan, Finale Doshi-Velez
We introduce a novel apprenticeship learning algorithm to learn an expert's underlying reward structure in off-policy model-free \emph{batch} settings.
no code implementations • 31 Jan 2019 • Isaac Lage, Emily Chen, Jeffrey He, Menaka Narayanan, Been Kim, Sam Gershman, Finale Doshi-Velez
Recent years have seen a boom in interest in machine learning systems that can provide a human-understandable rationale for their predictions or decisions.
no code implementations • 15 Jan 2019 • Xuefeng Peng, Yi Ding, David Wihl, Omer Gottesman, Matthieu Komorowski, Li-wei H. Lehman, Andrew Ross, Aldo Faisal, Finale Doshi-Velez
On a large retrospective cohort, this mixture-based approach outperforms physician, kernel only, and DRL-only experts.
no code implementations • 16 Nov 2018 • Melanie F. Pradier, Weiwei Pan, Jiayu Yao, Soumya Ghosh, Finale Doshi-Velez
As machine learning systems get widely adopted for high-stake decisions, quantifying uncertainty over predictions becomes crucial.
1 code implementation • EMNLP 2018 • Lifeng Jin, Finale Doshi-Velez, Timothy Miller, William Schuler, Lane Schwartz
There have been several recent attempts to improve the accuracy of grammar induction systems by bounding the recursive complexity of the induction model (Ponvert et al., 2011; Noji and Johnson, 2016; Shain et al., 2016; Jin et al., 2018).
no code implementations • 3 Jul 2018 • Aniruddh Raghu, Omer Gottesman, Yao Liu, Matthieu Komorowski, Aldo Faisal, Finale Doshi-Velez, Emma Brunskill
In this work, we consider the problem of estimating a behaviour policy for use in Off-Policy Policy Evaluation (OPE) when the true behaviour policy is unknown.
2 code implementations • 22 Jun 2018 • Andrew Slavin Ross, Weiwei Pan, Finale Doshi-Velez
There has been growing interest in developing accurate models that can also be explained to humans.
2 code implementations • ICML 2018 • Soumya Ghosh, Jiayu Yao, Finale Doshi-Velez
Bayesian Neural Networks (BNNs) have recently received increasing attention for their ability to provide well-calibrated posterior uncertainties.
no code implementations • 31 May 2018 • Omer Gottesman, Fredrik Johansson, Joshua Meier, Jack Dent, Dong-hun Lee, Srivatsan Srinivasan, Linying Zhang, Yi Ding, David Wihl, Xuefeng Peng, Jiayu Yao, Isaac Lage, Christopher Mosch, Li-wei H. Lehman, Matthieu Komorowski, Aldo Faisal, Leo Anthony Celi, David Sontag, Finale Doshi-Velez
Much attention has been devoted recently to the development of machine learning algorithms with the goal of improving treatment policies in healthcare.
no code implementations • NeurIPS 2018 • Isaac Lage, Andrew Slavin Ross, Been Kim, Samuel J. Gershman, Finale Doshi-Velez
We often desire our models to be interpretable as well as accurate.
1 code implementation • NeurIPS 2018 • Yao Liu, Omer Gottesman, Aniruddh Raghu, Matthieu Komorowski, Aldo Faisal, Finale Doshi-Velez, Emma Brunskill
We study the problem of off-policy policy evaluation (OPPE) in RL.
no code implementations • 16 Mar 2018 • M. Arjumand Masood, Finale Doshi-Velez
Bayesian Non-negative Matrix Factorization (NMF) is a promising approach for understanding uncertainty and structure in matrix data.
1 code implementation • TACL 2018 • Lifeng Jin, Finale Doshi-Velez, Timothy Miller, William Schuler, Lane Schwartz
There has been recent interest in applying cognitively or empirically motivated bounds on recursion depth to limit the search space of grammar induction models (Ponvert et al., 2011; Noji and Johnson, 2016; Shain et al., 2016).
no code implementations • 2 Feb 2018 • Menaka Narayanan, Emily Chen, Jeffrey He, Been Kim, Sam Gershman, Finale Doshi-Velez
Recent years have seen a boom in interest in machine learning systems that can provide a human-understandable rationale for their predictions or decisions.
1 code implementation • NeurIPS 2017 • Taylor W. Killian, Samuel Daulton, George Konidaris, Finale Doshi-Velez
We introduce a new formulation of the Hidden Parameter Markov Decision Process (HiP-MDP), a framework for modeling families of related tasks using low-dimensional latent embeddings.
no code implementations • 1 Dec 2017 • Michael C. Hughes, Gabriel Hope, Leah Weiner, Thomas H. McCoy, Roy H. Perlis, Erik B. Sudderth, Finale Doshi-Velez
Supervisory signals can help topic models discover low-dimensional data representations that are more interpretable for clinical tasks.
1 code implementation • 26 Nov 2017 • Andrew Slavin Ross, Finale Doshi-Velez
Deep neural networks have proven remarkably effective at solving many classification problems, but have been criticized recently for two major weaknesses: the reasons behind their predictions are uninterpretable, and the predictions themselves can often be fooled by small adversarial perturbations.
2 code implementations • 16 Nov 2017 • Mike Wu, Michael C. Hughes, Sonali Parbhoo, Maurizio Zazzi, Volker Roth, Finale Doshi-Velez
The lack of interpretability remains a key barrier to the adoption of deep models in many applications.
no code implementations • 3 Nov 2017 • Finale Doshi-Velez, Mason Kortz, Ryan Budish, Chris Bavitz, Sam Gershman, David O'Brien, Kate Scott, Stuart Schieber, James Waldo, David Weinberger, Adrian Weller, Alexandra Wood
The ubiquity of systems using artificial intelligence or "AI" has brought increasing attention to how those systems should be regulated.
1 code implementation • ICML 2018 • Stefan Depeweg, José Miguel Hernández-Lobato, Finale Doshi-Velez, Steffen Udluft
Bayesian neural networks with latent variables are scalable and flexible probabilistic models: They account for uncertainty in the estimation of the network weights and, by making use of latent variables, can capture complex noise patterns in the data.
no code implementations • 18 Oct 2017 • Omer Gottesman, Weiwei Pan, Finale Doshi-Velez
Tensor decomposition methods are popular tools for learning latent variables given only lower-order moments of the data.
no code implementations • 8 Sep 2017 • Kexin Yi, Finale Doshi-Velez
We propose a new framework for Hamiltonian Monte Carlo (HMC) on truncated probability distributions with smooth underlying density functions.
no code implementations • 23 Jul 2017 • Michael C. Hughes, Leah Weiner, Gabriel Hope, Thomas H. McCoy Jr., Roy H. Perlis, Erik B. Sudderth, Finale Doshi-Velez
Supervisory signals have the potential to make low-dimensional data representations, like those learned by mixture and topic models, more interpretable and useful.
no code implementations • 26 Jun 2017 • Stefan Depeweg, José Miguel Hernández-Lobato, Finale Doshi-Velez, Steffen Udluft
Bayesian neural networks (BNNs) with latent variables are probabilistic models which can automatically identify complex stochastic patterns in the data.
1 code implementation • 20 Jun 2017 • Taylor Killian, Samuel Daulton, George Konidaris, Finale Doshi-Velez
We introduce a new formulation of the Hidden Parameter Markov Decision Process (HiP-MDP), a framework for modeling families of related tasks using low-dimensional latent embeddings.
1 code implementation • 29 May 2017 • Soumya Ghosh, Finale Doshi-Velez
Bayesian Neural Networks (BNNs) have recently received increasing attention for their ability to provide well-calibrated posterior uncertainties.
1 code implementation • 10 Mar 2017 • Andrew Slavin Ross, Michael C. Hughes, Finale Doshi-Velez
Neural networks are among the most accurate supervised learning methods in use today, but their opacity makes them difficult to trust in critical applications, especially when conditions in training differ from those in test.
no code implementations • 28 Feb 2017 • Finale Doshi-Velez, Been Kim
As machine learning systems become ubiquitous, there has been a surge of interest in interpretable machine learning: systems that provide explanation for their outputs.
BIG-bench Machine Learning Interpretable Machine Learning +1
no code implementations • 12 Jan 2017 • Angela Fan, Finale Doshi-Velez, Luke Miratrix
In this work, we first show how the standard topic quality measures of coherence and pointwise mutual information act counter-intuitively in the presence of common but irrelevant words, making it difficult to even quantitatively identify situations in which topics may be dominated by stopwords.
no code implementations • 6 Dec 2016 • Michael C. Hughes, Huseyin Melih Elibol, Thomas McCoy, Roy Perlis, Finale Doshi-Velez
Supervised topic models can help clinical researchers find interpretable cooccurence patterns in count data that are relevant for diagnostics.
no code implementations • 1 Dec 2016 • Taylor Killian, George Konidaris, Finale Doshi-Velez
Due to physiological variation, patients diagnosed with the same condition may exhibit divergent, but related, responses to the same treatments.
no code implementations • COLING 2016 • Cory Shain, William Bryce, Lifeng Jin, Victoria Krakovna, Finale Doshi-Velez, Timothy Miller, William Schuler, Lane Schwartz
This paper presents a new memory-bounded left-corner parsing model for unsupervised raw-text syntax induction, using unsupervised hierarchical hidden Markov models (UHHMM).
no code implementations • 18 Nov 2016 • Viktoriya Krakovna, Finale Doshi-Velez
As deep neural networks continue to revolutionize various application domains, there is increasing interest in making these powerful models more understandable and interpretable, and narrowing down the causes of good and bad predictions.
no code implementations • 27 Oct 2016 • M. Arjumand Masood, Finale Doshi-Velez
Non-negative Matrix Factorization (NMF) is a popular tool for data exploration.
no code implementations • 20 Jun 2016 • Arjumand Masood, Weiwei Pan, Finale Doshi-Velez
and (2) How independent are the samples (as MCMC procedures produce correlated samples)?
no code implementations • 16 Jun 2016 • Viktoriya Krakovna, Finale Doshi-Velez
As deep neural networks continue to revolutionize various application domains, there is increasing interest in making these powerful models more understandable and interpretable, and narrowing down the causes of good and bad predictions.
2 code implementations • 23 May 2016 • Stefan Depeweg, José Miguel Hernández-Lobato, Finale Doshi-Velez, Steffen Udluft
We present an algorithm for model-based reinforcement learning that combines Bayesian neural networks (BNNs) with random roll-outs and stochastic optimization for policy learning.
Model-based Reinforcement Learning reinforcement-learning +2
no code implementations • 29 Mar 2016 • Dustin Tran, Minjae Kim, Finale Doshi-Velez
Method of moment estimators exhibit appealing statistical properties, such as asymptotic unbiasedness, for nonconvex problems.
no code implementations • NeurIPS 2015 • Been Kim, Julie A. Shah, Finale Doshi-Velez
We present the Mind the Gap Model (MGM), an approach for interpretable feature extraction and selection.
no code implementations • 28 Apr 2015 • Tong Wang, Cynthia Rudin, Finale Doshi-Velez, Yimin Liu, Erica Klampfl, Perry MacNeille
In both cases, there are prior parameters that the user can set to encourage the model to have a desired size and shape, to conform with a domain-specific definition of interpretability.
no code implementations • 16 Oct 2014 • Finale Doshi-Velez, Byron Wallace, Ryan Adams
In our model, topics are summarized by a few latent concept-words from the underlying graph that explain the observed words.
no code implementations • 15 Aug 2013 • Finale Doshi-Velez, George Konidaris
Control applications often feature tasks with similar, but not identical, dynamics.
no code implementations • NeurIPS 2010 • Finale Doshi-Velez, David Wingate, Nicholas Roy, Joshua B. Tenenbaum
We consider reinforcement learning in partially observable domains where the agent can query an expert for demonstrations.
no code implementations • NeurIPS 2009 • Finale Doshi-Velez, Shakir Mohamed, Zoubin Ghahramani, David A. Knowles
Nonparametric Bayesian models provide a framework for flexible probabilistic modelling of complex datasets.
no code implementations • NeurIPS 2009 • Finale Doshi-Velez
The Partially Observable Markov Decision Process (POMDP) framework has proven useful in planning domains that require balancing actions that increase an agents knowledge and actions that increase an agents reward.