Pradeep Ravikumar
YOU?
Author Swipe
View article: Contextures: Representations from Contexts
Contextures: Representations from Contexts Open
Despite the empirical success of foundation models, we do not have a systematic characterization of the representations that these models learn. In this paper, we establish the contexture theory. It shows that a large class of representati…
View article: LogiCity: Advancing Neuro-Symbolic AI with Abstract Urban Simulation
LogiCity: Advancing Neuro-Symbolic AI with Abstract Urban Simulation Open
Recent years have witnessed the rapid development of Neuro-Symbolic (NeSy) AI systems, which integrate symbolic reasoning into deep neural networks. However, most of the existing benchmarks for NeSy AI fail to provide long-horizon reasonin…
View article: Identifying General Mechanism Shifts in Linear Causal Representations
Identifying General Mechanism Shifts in Linear Causal Representations Open
We consider the linear causal representation learning setting where we observe a linear mixing of $d$ unknown latent factors, which follow a linear structural causal model. Recent work has shown that it is possible to recover the latent fa…
View article: Markov Equivalence and Consistency in Differentiable Structure Learning
Markov Equivalence and Consistency in Differentiable Structure Learning Open
Existing approaches to differentiable structure learning of directed acyclic graphs (DAGs) rely on strong identifiability assumptions in order to guarantee that global minimizers of the acyclicity-constrained optimization problem identifie…
View article: Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers
Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers Open
Large Language Models (LLMs) have the capacity to store and recall facts. Through experimentation with open-source models, we observe that this ability to retrieve facts can be easily manipulated by changing contexts, even without altering…
View article: On the Origins of Linear Representations in Large Language Models
On the Origins of Linear Representations in Large Language Models Open
Recent works have argued that high-level semantic concepts are encoded "linearly" in the representation space of large language models. In this work, we study the origins of such linear representations. To that end, we introduce a simple l…
View article: Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models
Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models Open
To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build h…
View article: Spectrally Transformed Kernel Regression
Spectrally Transformed Kernel Regression Open
Unlabeled data is a key component of modern machine learning. In general, the role of unlabeled data is to impose a form of smoothness, usually from the similarity information encoded in a base kernel, such as the $ε$-neighbor kernel or th…
View article: An Interventional Perspective on Identifiability in Gaussian LTI Systems with Independent Component Analysis
An Interventional Perspective on Identifiability in Gaussian LTI Systems with Independent Component Analysis Open
We investigate the relationship between system identification and intervention design in dynamical systems. While previous research demonstrated how identifiable representation learning methods, such as Independent Component Analysis (ICA)…
View article: Responsible AI (RAI) Games and Ensembles
Responsible AI (RAI) Games and Ensembles Open
Several recent works have studied the societal effects of AI; these include issues such as fairness, robustness, and safety. In many of these objectives, a learner seeks to minimize its worst-case loss over a set of predefined distribution…
View article: Sample based Explanations via Generalized Representers
Sample based Explanations via Generalized Representers Open
We propose a general class of sample based explanations of machine learning models, which we term generalized representers. To measure the effect of a training sample on a model's test prediction, generalized representers use two component…
View article: Identifying Representations for Intervention Extrapolation
Identifying Representations for Intervention Extrapolation Open
The premise of identifiable and causal representation learning is to improve the current representation learning paradigm in terms of generalizability or robustness. Despite recent progress in questions of identifiability, more theoretical…
View article: Individual Fairness Under Uncertainty
Individual Fairness Under Uncertainty Open
Algorithmic fairness, the research field of making machine learning (ML) algorithms fair, is an established area in ML. As ML technologies expand their application domains, including ones with high societal impact, it becomes essential to …
View article: iSCAN: Identifying Causal Mechanism Shifts among Nonlinear Additive Noise Models
iSCAN: Identifying Causal Mechanism Shifts among Nonlinear Additive Noise Models Open
Structural causal models (SCMs) are widely used in various disciplines to represent causal relationships among variables in complex systems. Unfortunately, the underlying causal structure is often unknown, and estimating it from data remai…
View article: Global Optimality in Bivariate Gradient-based DAG Learning
Global Optimality in Bivariate Gradient-based DAG Learning Open
Recently, a new class of non-convex optimization problems motivated by the statistical problem of learning an acyclic directed graphical model from data has attracted significant interest. While existing work uses standard first-order opti…
View article: Learning Linear Causal Representations from Interventions under General Nonlinear Mixing
Learning Linear Causal Representations from Interventions under General Nonlinear Mixing Open
We study the problem of learning causal representations from unknown, latent interventions in a general setting, where the latent distribution is Gaussian but the mixing function is completely general. We prove strong identifiability resul…
View article: Understanding Augmentation-based Self-Supervised Representation Learning via RKHS Approximation and Regression
Understanding Augmentation-based Self-Supervised Representation Learning via RKHS Approximation and Regression Open
Data augmentation is critical to the empirical success of modern self-supervised representation learning, such as contrastive learning and masked language modeling. However, a theoretical understanding of the exact role of augmentation rem…
View article: Representer Point Selection for Explaining Regularized High-dimensional Models
Representer Point Selection for Explaining Regularized High-dimensional Models Open
We introduce a novel class of sample-based explanations we term high-dimensional representers, that can be used to explain the predictions of a regularized high-dimensional model in terms of importance weights for each of the training samp…
View article: Optimizing NOTEARS Objectives via Topological Swaps
Optimizing NOTEARS Objectives via Topological Swaps Open
Recently, an intriguing class of non-convex optimization problems has emerged in the context of learning directed acyclic graphs (DAGs). These problems involve minimizing a given loss or score function, subject to a non-convex continuous c…
View article: Learning with Explanation Constraints
Learning with Explanation Constraints Open
As larger deep learning models are hard to interpret, there has been a recent focus on generating explanations of these black-box models. In contrast, we may have apriori explanations of how models should behave. In this paper, we formaliz…
View article: Individual Fairness under Uncertainty
Individual Fairness under Uncertainty Open
Algorithmic fairness, the research field of making machine learning (ML) algorithms fair, is an established area in ML. As ML technologies expand their application domains, including ones with high societal impact, it becomes essential to …
View article: Nash Equilibria and Pitfalls of Adversarial Training in Adversarial Robustness Games
Nash Equilibria and Pitfalls of Adversarial Training in Adversarial Robustness Games Open
Adversarial training is a standard technique for training adversarially robust models. In this paper, we study adversarial training as an alternating best-response strategy in a 2-player zero-sum game. We prove that even in a simple scenar…
View article: Label Propagation with Weak Supervision
Label Propagation with Weak Supervision Open
Semi-supervised learning and weakly supervised learning are important paradigms that aim to reduce the growing demand for labeled data in current machine learning applications. In this paper, we introduce a novel analysis of the classical …
View article: DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterization
DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterization Open
The combinatorial problem of learning directed acyclic graphs (DAGs) from data was recently framed as a purely continuous optimization problem by leveraging a differentiable acyclicity characterization of DAGs based on the trace of a matri…
View article: Concept Gradient: Concept-based Interpretation Without Linear Assumption
Concept Gradient: Concept-based Interpretation Without Linear Assumption Open
Concept-based interpretations of black-box models are often more intuitive for humans to understand. The most widely adopted approach for concept-based interpretation is Concept Activation Vector (CAV). CAV relies on learning a linear rela…
View article: Identifiability of deep generative models without auxiliary information
Identifiability of deep generative models without auxiliary information Open
We prove identifiability of a broad class of deep latent variable models that (a) have universal approximation capabilities and (b) are the decoders of variational autoencoders that are commonly used in practice. Unlike existing work, our …
View article: Building Robust Ensembles via Margin Boosting
Building Robust Ensembles via Margin Boosting Open
In the context of adversarial robustness, a single model does not usually have enough power to defend against all possible adversarial attacks, and as a result, has sub-optimal robustness. Consequently, an emerging line of work has focused…
View article: Faith-Shap: The Faithful Shapley Interaction Index
Faith-Shap: The Faithful Shapley Interaction Index Open
Shapley values, which were originally designed to assign attributions to individual players in coalition games, have become a commonly used approach in explainable machine learning to provide attributions to input features for black-box ma…
View article: Human-Centered Concept Explanations for Neural Networks
Human-Centered Concept Explanations for Neural Networks Open
Understanding complex machine learning models such as deep neural networks with explanations is crucial in various applications. Many explanations stem from the model perspective, and may not necessarily effectively communicate why the mod…
View article: Threading the Needle of On and Off-Manifold Value Functions for Shapley Explanations
Threading the Needle of On and Off-Manifold Value Functions for Shapley Explanations Open
A popular explainable AI (XAI) approach to quantify feature importance of a given model is via Shapley values. These Shapley values arose in cooperative games, and hence a critical ingredient to compute these in an XAI context is a so-call…