Search Results for author: Victor Veitch

Found 25 papers, 18 papers with code

Transforming and Combining Rewards for Aligning Large Language Models

no code implementations1 Feb 2024 ZiHao Wang, Chirag Nagpal, Jonathan Berant, Jacob Eisenstein, Alex D'Amour, Sanmi Koyejo, Victor Veitch

A common approach for aligning language models to human preferences is to first learn a reward model from preference data, and then use this reward model to update the language model.

Language Modelling

The Linear Representation Hypothesis and the Geometry of Large Language Models

1 code implementation7 Nov 2023 Kiho Park, Yo Joong Choe, Victor Veitch

Using this causal inner product, we show how to unify all notions of linear representation.

counterfactual Sentence

Causal Context Connects Counterfactual Fairness to Robust Prediction and Group Fairness

1 code implementation NeurIPS 2023 Jacy Reese Anthis, Victor Veitch

This is an intuitive standard, as reflected in the U. S. legal system, but its use is limited because counterfactuals cannot be directly observed in real-world data.

counterfactual Fairness

Concept Algebra for (Score-Based) Text-Controlled Generative Models

1 code implementation NeurIPS 2023 ZiHao Wang, Lin Gui, Jeffrey Negrea, Victor Veitch

This suggests these models have internal representations that encode concepts in a `disentangled' manner.

Efficient Conditionally Invariant Representation Learning

1 code implementation16 Dec 2022 Roman Pogodin, Namrata Deka, Yazhe Li, Danica J. Sutherland, Victor Veitch, Arthur Gretton

The procedure requires just a single ridge regression from $Y$ to kernelized features of $Z$, which can be done in advance.

Fairness regression +1

Causal Estimation for Text Data with (Apparent) Overlap Violations

no code implementations30 Sep 2022 Lin Gui, Victor Veitch

To estimate a causal effect from observational data, we need to adjust for confounding aspects of the text that affect both the treatment and outcome -- e. g., the topic or writing level of the text.

Attribute Causal Identification +2

The Causal Structure of Domain Invariant Supervised Representation Learning

no code implementations15 Aug 2022 ZiHao Wang, Victor Veitch

Then, we characterize the causal structures that are compatible with this notion of invariance. With this in hand, we find conditions under which method-specific invariance notions correspond to real-world invariant structure, and we clarify the relationship between invariant structure and robustness to domain shifts.

Data Augmentation Representation Learning

Invariant and Transportable Representations for Anti-Causal Domain Shifts

1 code implementation4 Jul 2022 Yibo Jiang, Victor Veitch

In this paper, we study representation learning under a particular notion of domain shift that both respects causal invariance and that naturally handles the "anti-causal" structure.

Representation Learning

Using Embeddings for Causal Estimation of Peer Influence in Social Networks

2 code implementations17 May 2022 Irina Cristali, Victor Veitch

The main aim is to perform this adjustment nonparametrically, without functional form assumptions on either the process that generated the network or the treatment assignment and outcome processes.

Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Tests

no code implementations NeurIPS 2021 Victor Veitch, Alexander D'Amour, Steve Yadlowsky, Jacob Eisenstein

We introduce counterfactual invariance as a formalization of the requirement that changing irrelevant parts of the input shouldn't change model predictions.

Causal Inference counterfactual +2

Counterfactual Invariance to Spurious Correlations in Text Classification

no code implementations NeurIPS 2021 Victor Veitch, Alexander D'Amour, Steve Yadlowsky, Jacob Eisenstein

We introduce counterfactual invariance as a formalization of the requirement that changing irrelevant parts of the input shouldn't change model predictions.

Causal Inference counterfactual +2

Invariant Representation Learning for Treatment Effect Estimation

1 code implementation24 Nov 2020 Claudia Shi, Victor Veitch, David Blei

To address this challenge, practitioners collect and adjust for the covariates, hoping that they adequately correct for confounding.

Causal Identification Causal Inference +2

Causal Effects of Linguistic Properties

1 code implementation NAACL 2021 Reid Pryzant, Dallas Card, Dan Jurafsky, Victor Veitch, Dhanya Sridhar

Second, in practice, we only have access to noisy proxies for the linguistic properties of interest -- e. g., predictions from classifiers and lexicons.

Language Modelling

Valid Causal Inference with (Some) Invalid Instruments

no code implementations19 Jun 2020 Jason Hartford, Victor Veitch, Dhanya Sridhar, Kevin Leyton-Brown

The technique is simple to apply and is "black-box" in the sense that it may be used with any instrumental variable estimator as long as the treatment effect is identified for each valid instrument independently.

Causal Inference valid

Sense and Sensitivity Analysis: Simple Post-Hoc Analysis of Bias Due to Unobserved Confounding

2 code implementations NeurIPS 2020 Victor Veitch, Anisha Zaveri

The purpose of this paper is to develop \emph{Austen plots}, a sensitivity analysis tool to aid such judgments by making it easier to reason about potential bias induced by unobserved confounding.

Causal Inference

Adapting Neural Networks for the Estimation of Treatment Effects

5 code implementations NeurIPS 2019 Claudia Shi, David M. Blei, Victor Veitch

We propose two adaptations based on insights from the statistical literature on the estimation of treatment effects.

Causal Inference

Adapting Text Embeddings for Causal Inference

4 code implementations29 May 2019 Victor Veitch, Dhanya Sridhar, David M. Blei

To address this challenge, we develop causally sufficient embeddings, low-dimensional document representations that preserve sufficient information for causal identification and allow for efficient estimation of causal effects.

Causal Identification Causal Inference +4

The Holdout Randomization Test for Feature Selection in Black Box Models

3 code implementations1 Nov 2018 Wesley Tansey, Victor Veitch, Haoran Zhang, Raul Rabadan, David M. Blei

We propose the holdout randomization test (HRT), an approach to feature selection using black box predictive models.

Methodology

Empirical Risk Minimization and Stochastic Gradient Descent for Relational Data

1 code implementation27 Jun 2018 Victor Veitch, Morgane Austern, Wenda Zhou, David M. Blei, Peter Orbanz

We solve this problem using recent ideas from graph sampling theory to (i) define an empirical risk for relational data and (ii) obtain stochastic gradients for this empirical risk that are automatically unbiased.

Graph Sampling Node Classification

Exchangeable modelling of relational data: checking sparsity, train-test splitting, and sparse exchangeable Poisson matrix factorization

1 code implementation6 Dec 2017 Victor Veitch, Ekansh Sharma, Zacharie Naulet, Daniel M. Roy

A variety of machine learning tasks---e. g., matrix factorization, topic modelling, and feature allocation---can be viewed as learning the parameters of a probability distribution over bipartite graphs.

Test Variational Inference

Bootstrap estimators for the tail-index and for the count statistics of graphex processes

1 code implementation5 Dec 2017 Zacharie Naulet, Ekansh Sharma, Victor Veitch, Daniel M. Roy

Graphex processes resolve some pathologies in traditional random graph models, notably, providing models that are both projective and allow sparsity.

Statistics Theory Statistics Theory Primary 62F10, secondary 60G55, 60G70

Cannot find the paper you are looking for? You can Submit a new open access paper.