Search Results for author: Kostis Gourgoulias

Found 7 papers, 1 papers with code

DeepClean: Machine Unlearning on the Cheap by Resetting Privacy Sensitive Weights using the Fisher Diagonal

no code implementations17 Nov 2023 Jiaeli Shi, Najah Ghalyan, Kostis Gourgoulias, John Buford, Sean Moran

By leveraging the FIM diagonal, our approach provides an interpretable, lightweight, and efficient solution for machine unlearning with practical privacy benefits.

Estimating Class Separability of Datasets Using Persistent Homology with Application to LLM Fine-Tuning

no code implementations24 May 2023 Najah Ghalyan, Kostis Gourgoulias, Yash Satsangi, Sean Moran, Maxime Labonne, Joseph Sabelja

This paper proposes a method to estimate the class separability of an unlabeled text dataset by inspecting the topological characteristics of sentence-transformer embeddings of the text.

Language Modelling Transductive Learning

Masking schemes for universal marginalisers

no code implementations pproximateinference AABI Symposium 2019 Divya Gautam, Maria Lomeli, Kostis Gourgoulias, Daniel H. Thompson, Saurabh Johri

We consider the effect of structure-agnostic and structure-dependent masking schemes when training a universal marginaliser (arXiv:1711. 00695) in order to learn conditional distributions of the form $P(x_i |\mathbf x_{\mathbf b})$, where $x_i$ is a given random variable and $\mathbf x_{\mathbf b}$ is some arbitrary subset of all random variables of the generative model of interest.


Universal Marginaliser for Deep Amortised Inference for Probabilistic Programs

no code implementations16 Oct 2019 Robert Walecki, Kostis Gourgoulias, Adam Baker, Chris Hart, Chris Lucas, Max Zwiessele, Albert Buchard, Maria Lomeli, Yura Perov, Saurabh Johri

Probabilistic programming languages (PPLs) are powerful modelling tools which allow to formalise our knowledge about the world and reason about its inherent uncertainty.

Probabilistic Programming

Universal Marginalizer for Amortised Inference and Embedding of Generative Models

no code implementations12 Nov 2018 Robert Walecki, Albert Buchard, Kostis Gourgoulias, Chris Hart, Maria Lomeli, A. K. W. Navarro, Max Zwiessele, Yura Perov, Saurabh Johri

Probabilistic graphical models are powerful tools which allow us to formalise our knowledge about the world and reason about its inherent uncertainty.


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