9 citations · 12 across the 10 of their papers we have counts for
11 papers
Cross-Entropy Games and Frost Training
Arthur Renard, Franck Gabriel, Valentin Hartmann +1
We present Frost Training, a method for improving Monte Carlo-based policy optimization for a large family of LLM-as-a-judge tasks called Cross-Entropy Games. The key idea is to ex…
TiME: Tiny Monolingual Encoders for Efficient NLP Pipelines
David Schulmeister, Valentin Hartmann, Lars Klein +1
Today, a lot of research on language models is focused on large, general-purpose models. However, many NLP pipelines only require models with a well-defined, small set of capabilit…
Neural Redshift: Random Networks are not Random Functions
Damien Teney, Armand Nicolicioiu, Valentin Hartmann +1
Our understanding of the generalization capabilities of neural networks (NNs) is still incomplete. Prevailing explanations are based on implicit biases of gradient descent (GD) but…
SoK: Memorization in General-Purpose Large Language Models
Valentin Hartmann, Anshuman Suri, Vincent Bindschaedler +3
Large Language Models (LLMs) are advancing at a remarkable pace, with myriad applications under development. Unlike most earlier machine learning models, they are no longer built f…
Language Model Decoding as Likelihood-Utility Alignment
Martin Josifoski, Maxime Peyrard, Frano Rajic +8
A critical component of a successful language generation pipeline is the decoding algorithm. However, the general principles that should guide the choice of a decoding algorithm re…
DiPPS: Differentially Private Propensity Scores for Bias Correction
Liangwei Chen, Valentin Hartmann, Robert West
In surveys, it is typically up to the individuals to decide if they want to participate or not, which leads to participation bias: the individuals willing to share their data might…
On the Choice of Databases in Differential Privacy Composition
Valentin Hartmann, Vincent Bindschaedler, Robert West
Differential privacy (DP) is a widely applied paradigm for releasing data while maintaining user privacy. Its success is to a large part due to its composition property that guaran…
Distribution inference risks: Identifying and mitigating sources of leakage
Valentin Hartmann, Léo Meynent, Maxime Peyrard +3
A large body of work shows that machine learning (ML) models can leak sensitive or confidential information about their training data. Recently, leakage due to distribution inferen…
Privacy accounting conomics: Improving differential privacy composition via a posteriori bounds
Valentin Hartmann, Vincent Bindschaedler, Alexander Bentkamp +1
Differential privacy (DP) is a widely used notion for reasoning about privacy when publishing aggregate data. In this paper, we observe that certain DP mechanisms are amenable to a…
Privacy-Preserving Classification with Secret Vector Machines
Valentin Hartmann, Konark Modi, Josep M. Pujol +1
Today, large amounts of valuable data are distributed among millions of user-held devices, such as personal computers, phones, or Internet-of-things devices. Many companies collect…
Secure Summation via Subset Sums: A New Primitive for Privacy-Preserving Distributed Machine Learning
Valentin Hartmann, Robert West
For population studies or for the training of complex machine learning models, it is often required to gather data from different actors. In these applications, summation is an imp…