Search Results for author:
Found 36 papers, 13 papers with code
Defending LLM Watermarking Against Spoofing Attacks with Contrastive Representation Learning
We identify two core challenges that make defending against spoofing difficult: (1) the need for watermarks to be both sensitive to semantic-distorting changes and insensitive to semantic-preserving edits, and (2) the contradiction between the need to detect global semantic shifts and the local, auto-regressive nature of most watermarking schemes.
Dataset Protection via Watermarked Canaries in Retrieval-Augmented LLMs
Retrieval-Augmented Generation (RAG) has become an effective method for enhancing large language models (LLMs) with up-to-date knowledge.
Distributional Information Embedding: A Framework for Multi-bit Watermarking
This paper introduces a novel problem, distributional information embedding, motivated by the practical demands of multi-bit watermarking for large language models (LLMs).
Image Watermarks are Removable Using Controllable Regeneration from Clean Noise
This scheme adds varying numbers of noise steps to the latent representation of the watermarked image, followed by a controlled denoising process starting from this noisy latent representation.
Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approach
Our approach focuses on maximizing detection performance while maintaining control over the worst-case Type-I error and text distortion.
SAUC: Sparsity-Aware Uncertainty Calibration for Spatiotemporal Prediction with Graph Neural Networks
Quantifying uncertainty is crucial for robust and reliable predictions.
Information-Theoretic Opacity-Enforcement in Markov Decision Processes
The goal of the observer is to infer some secret, represented by a random variable, from its partial observations, while the goal of the planning agent is to make the secret maximally opaque to the observer while achieving a satisfactory total return.
Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage?
This paper questions the effectiveness of a modern predictive uncertainty quantification approach, called \emph{evidential deep learning} (EDL), in which a single neural network model is trained to learn a meta distribution over the predictive distribution by minimizing a specific objective function.
Operator SVD with Neural Networks via Nested Low-Rank Approximation
Computing eigenvalue decomposition (EVD) of a given linear operator, or finding its leading eigenvalues and eigenfunctions, is a fundamental task in many machine learning and scientific computing problems.
Adaptive Text Watermark for Large Language Models
The advancement of Large Language Models (LLMs) has led to increasing concerns about the misuse of AI-generated text, and watermarking for LLM-generated text has emerged as a potential solution.
Class-wise Generalization Error: an Information-Theoretic Analysis
Existing generalization theories of supervised learning typically take a holistic approach and provide bounds for the expected generalization over the whole data distribution, which implicitly assumes that the model generalizes similarly for all the classes.
Gibbs-Based Information Criteria and the Over-Parameterized Regime
Double-descent refers to the unexpected drop in test loss of a learning algorithm beyond an interpolating threshold with over-parameterization, which is not predicted by information criteria in their classical forms due to the limitations in the standard asymptotic approach.
Feature Learning in Image Hierarchies using Functional Maximal Correlation
This paper proposes the Hierarchical Functional Maximal Correlation Algorithm (HFMCA), a hierarchical methodology that characterizes dependencies across two hierarchical levels in multiview systems.
A Bilateral Bound on the Mean-Square Error for Estimation in Model Mismatch
A bilateral (i. e., upper and lower) bound on the mean-square error under a general model mismatch is developed.
Reliable Gradient-free and Likelihood-free Prompt Tuning
Due to privacy or commercial constraints, large pre-trained language models (PLMs) are often offered as black-box APIs.
On the Generalization Error of Meta Learning for the Gibbs Algorithm
We analyze the generalization ability of joint-training meta learning algorithms via the Gibbs algorithm.
Group Fairness with Uncertainty in Sensitive Attributes
To overcome this limitation, we propose a bootstrap-based algorithm that achieves the target level of fairness despite the uncertainty in sensitive attributes.
Post-hoc Uncertainty Learning using a Dirichlet Meta-Model
It is known that neural networks have the problem of being over-confident when directly using the output label distribution to generate uncertainty measures.
How Does Pseudo-Labeling Affect the Generalization Error of the Semi-Supervised Gibbs Algorithm?
Our findings offer new insights that the generalization performance of SSL with pseudo-labeling is affected not only by the information between the output hypothesis and input training data but also by the information {\em shared} between the {\em labeled} and {\em pseudo-labeled} data samples.
Data-Driven Blind Synchronization and Interference Rejection for Digital Communication Signals
We study the potential of data-driven deep learning methods for separation of two communication signals from an observation of their mixture.
Exploiting Temporal Structures of Cyclostationary Signals for Data-Driven Single-Channel Source Separation
We study the problem of single-channel source separation (SCSS), and focus on cyclostationary signals, which are particularly suitable in a variety of application domains.
A Maximal Correlation Framework for Fair Machine Learning
As machine learning algorithms grow in popularity and diversify to many industries, ethical and legal concerns regarding their fairness have become increasingly relevant.
Tighter Expected Generalization Error Bounds via Convexity of Information Measures
Due to the convexity of the information measures, the proposed bounds in terms of Wasserstein distance and total variation distance are shown to be tighter than their counterparts based on individual samples in the literature.
On Balancing Bias and Variance in Unsupervised Multi-Source-Free Domain Adaptation
Due to privacy, storage, and other constraints, there is a growing need for unsupervised domain adaptation techniques in machine learning that do not require access to the data used to train a collection of source models.
Source-Free Domain Adaptation
Unsupervised Domain Adaptation
An Exact Characterization of the Generalization Error for the Gibbs Algorithm
Various approaches have been developed to upper bound the generalization error of a supervised learning algorithm.
Characterizing and Understanding the Generalization Error of Transfer Learning with Gibbs Algorithm
We provide an information-theoretic analysis of the generalization ability of Gibbs-based transfer learning algorithms by focusing on two popular transfer learning approaches, $\alpha$-weighted-ERM and two-stage-ERM.
Selective Regression Under Fairness Criteria
Selective regression allows abstention from prediction if the confidence to make an accurate prediction is not sufficient.
Characterizing the Generalization Error of Gibbs Algorithm with Symmetrized KL information
As a result, they may fail to characterize the exact generalization ability of a learning algorithm.
A Maximal Correlation Approach to Imposing Fairness in Machine Learning
As machine learning algorithms grow in popularity and diversify to many industries, ethical and legal concerns regarding their fairness have become increasingly relevant.
Adaptive Sequential Machine Learning
A framework previously introduced in [3] for solving a sequence of stochastic optimization problems with bounded changes in the minimizers is extended and applied to machine learning problems such as regression and classification.
Information-Theoretic Understanding of Population Risk Improvement with Model Compression
We show that model compression can improve the population risk of a pre-trained model, by studying the tradeoff between the decrease in the generalization error and the increase in the empirical risk with model compression.
Tightening Mutual Information Based Bounds on Generalization Error
The bound is derived under more general conditions on the loss function than in existing studies; nevertheless, it provides a tighter characterization of the generalization error.
Active Learning in Recommendation Systems with Multi-level User Preferences
While recommendation systems generally observe user behavior passively, there has been an increased interest in directly querying users to learn their specific preferences.
Model change detection with application to machine learning
The goal is to detect whether the change in the model is significant, i. e., whether the difference between the pre-change parameter and the post-change parameter $\|\theta-\theta'\|_2$ is larger than a pre-determined threshold $\rho$.
Active and Adaptive Sequential learning
Furthermore, an estimator of the change in the learning problems using the active learning samples is constructed, which provides an adaptive sample size selection rule that guarantees the excess risk is bounded for sufficiently large number of time steps.
Linear-Complexity Exponentially-Consistent Tests for Universal Outlying Sequence Detection
A sequence is considered as outlying if the observations therein are generated by a distribution different from those generating the observations in the majority of the sequences.
