FairDropout: Using Example-Tied Dropout to Enhance Generalization of Minority Groups
G'eraldin Nanfack
Eugene Belilovsky
Deep learning models frequently exploit spurious features in training data to achieve low training error, often resulting in poor generaliza… (voir plus)tion when faced with shifted testing distributions. To address this issue, various methods from imbalanced learning, representation learning, and classifier recalibration have been proposed to enhance the robustness of deep neural networks against spurious correlations. In this paper, we observe that models trained with empirical risk minimization tend to generalize well for examples from the majority groups while memorizing instances from minority groups. Building on recent findings that show memorization can be localized to a limited number of neurons, we apply example-tied dropout as a method we term FairDropout, aimed at redirecting this memorization to specific neurons that we subsequently drop out during inference. We empirically evaluate FairDropout using the subpopulation benchmark suite encompassing vision, language, and healthcare tasks, demonstrating that it significantly reduces reliance on spurious correlations, and outperforms state-of-the-art methods.
2025-02-10
ArXiv (prépublication)
arxiv.org
arxiv.org
Membership Inference Risks in Quantized Models: A Theoretical and Empirical Study
Eric Aubinais
Philippe Formont
Pablo Piantanida
Elisabeth Gassiat
Quantizing machine learning models has demonstrated its effectiveness in lowering memory and inference costs while maintaining performance l… (voir plus)evels comparable to the original models. In this work, we investigate the impact of quantization procedures on the privacy of data-driven models, specifically focusing on their vulnerability to membership inference attacks. We derive an asymptotic theoretical analysis of Membership Inference Security (MIS), characterizing the privacy implications of quantized algorithm weights against the most powerful (and possibly unknown) attacks. Building on these theoretical insights, we propose a novel methodology to empirically assess and rank the privacy levels of various quantization procedures. Using synthetic datasets, we demonstrate the effectiveness of our approach in assessing the MIS of different quantizers. Furthermore, we explore the trade-off between privacy and performance using real-world data and models in the context of molecular modeling.
2025-02-10
ArXiv (prépublication)
doi.org
arxiv.org
Membership Inference Risks in Quantized Models: A Theoretical and Empirical Study
Eric Aubinais
Philippe Formont
Pablo Piantanida
Elisabeth Gassiat
Quantizing machine learning models has demonstrated its effectiveness in lowering memory and inference costs while maintaining performance l… (voir plus)evels comparable to the original models. In this work, we investigate the impact of quantization procedures on the privacy of data-driven models, specifically focusing on their vulnerability to membership inference attacks. We derive an asymptotic theoretical analysis of Membership Inference Security (MIS), characterizing the privacy implications of quantized algorithm weights against the most powerful (and possibly unknown) attacks. Building on these theoretical insights, we propose a novel methodology to empirically assess and rank the privacy levels of various quantization procedures. Using synthetic datasets, we demonstrate the effectiveness of our approach in assessing the MIS of different quantizers. Furthermore, we explore the trade-off between privacy and performance using real-world data and models in the context of molecular modeling.
2025-02-10
ArXiv (prépublication)
arxiv.org
arxiv.org
Outsourced diffusion sampling: Efficient posterior inference in latent spaces of generative models
Siddarth Venkatraman
Mohsin Hasan
Minsu Kim
Luca Scimeca
Marcin Sendera
Yoshua Bengio
Glen Berseth
Nikolay Malkin
Any well-behaved generative model over a variable …
2025-02-10
ArXiv (prépublication)
arxiv.org
arxiv.org
Overcoming Political Upheaval to Deliver Pediatric Surgical Care in Afghanistan: A Prospective Analysis of the First 1000 Procedures.
Dunya Moghul
Phillip J Hsu
Emma Bryce
Yalda Obaidy
Zane Hellman
Ajmal Sherzad
Dan Poenaru
Maija Cheung
BACKGROUND Pediatric surgical care is limited in Afghanistan. Few details are known about the state of pediatric surgery in Afghanistan. We … (voir plus)explore the impact of a newly installed pediatric operating room by a children's charity on the provision of pediatric surgery in Afghanistan. STUDY DESIGN Following the opening in March 2023 of the new KidsOR operating room at Ataturk Hospital in Kabul, Afghanistan, perioperative clinical data was prospectively collected until December 2023. All children (age 14 years) undergoing surgical procedures were included in a REDCap database, and descriptive analyses were performed. RESULTS 1,014 opera
2025-02-10
Journal of the American College of Surgeons (publié)
doi.org
A physics-based data-driven model for CO$_2$ gas diffusion electrodes to drive automated laboratories
Ivan Grega
F'elix Therrien
Abhishek Soni
Karry Ocean
Kevan Dettelbach
Ribwar Ahmadi
Mehrdad Mokhtari
C. Berlinguette
Yoshua Bengio
The electrochemical reduction of atmospheric CO…
2025-02-10
ArXiv (prépublication)
arxiv.org
arxiv.org
Using Image-based AI for insect monitoring and conservation - InsectAI COST Action
Tom August
Mario Balzan
Paul Bodesheim
Gunnar Brehm
Lisette Cantú-Salazar
Sílvia Castro
Joseph Chipperfield
Guillaume Ghisbain
Alba Gomez-Segura
Jérémie Goulnik
Quentin Groom
Laurens Hogeweg
Chantal Huijbers
Andreas Kamilaris
Karolis Kazlauskis
Wouter Koch
Dimitri Korsch
João Loureiro
Youri Martin
Angeliki Martinou … (voir 27 de plus)
Kent McFarland
Xavier Mestdagh
Denis Michez
Charlie Outhwaite
Luca Pegoraro
Nadja Pernat
Lars Pettersson
Pavel Pipek
Cristina Preda
David Rolnick
Tobias Roth
David Roy
Helen Roy
Veljo Runnel
Martina Sasic
Dmitry Schigel
Julie Sheard
Cecilie Svenningsen
Heliana Teixeira
Nicolas Titeux
Thomas Tscheulin
Elli Tzirkalli
Marijn van der Velde
Roel van Klink
Nicolas Vereecken
Sarah Vray
Toke Thomas Høye
2025-02-10
Research Ideas and Outcomes (publié)
doi.org
RadiSeq: a single- and bulk-cell whole-genome DNA sequencing simulator for radiation-damaged cell models
Felix Mathew
Luc Galarneau
John Kildea
Objective To build and validate a simulation framework to perform single-cell and bulk-cell whole genome sequencing simulation of radiation-… (voir plus)exposed Monte Carlo cell models to assist radiation genomics studies. Approach Sequencing the genomes of radiation-damaged cells can provide useful insight into radiation action for radiobiology research. However, carrying out post-irradiation sequencing experiments can often be challenging, expensive, and time-consuming. Although computational simulations have the potential to provide solutions to these experimental challenges, and aid in designing optimal experiments, the absence of tools currently limits such application. Monte Carlo toolkits exist to simulate radiation exposures of cell models but there are no tools to simulate single- and bulk-cell sequencing of cell models containing radiation-damaged DNA. Therefore, we aimed to develop a Monte Carlo simulation framework to address this gap by designing a tool capable of simulating sequencing processes for radiation-damaged cells. Main Results We developed RadiSeq – a multi-threaded whole-genome DNA sequencing simulator written in C++. RadiSeq can be used to simulate Illumina sequencing of radiation-damaged cell models produced by Monte Carlo simulations. RadiSeq has been validated through comparative analysis, where simulated data were matched against experimentally obtained data, demonstrating reasonable agreement between the two. Additionally, it comes with numerous features designed to closely resemble actual whole-genome sequencing. RadiSeq is also highly customizable with a single input parameter file. Significance RadiSeq enables the research community to perform complex simulations of radiation-exposed DNA sequencing, supporting the optimization, planning, and validation of costly and time-intensive radiation biology experiments. This framework provides a powerful tool for advancing radiation genomics research.
2025-02-09
bioRxiv (prépublication)
doi.org
Mol-MoE: Training Preference-Guided Routers for Molecule Generation
Diego Calanzone
Pierluca D'Oro
Pierre-Luc Bacon
Recent advances in language models have enabled framing molecule generation as sequence modeling. However, existing approaches often rely on… (voir plus) single-objective reinforcement learning, limiting their applicability to real-world drug design, where multiple competing properties must be optimized. Traditional multi-objective reinforcement learning (MORL) methods require costly retraining for each new objective combination, making rapid exploration of trade-offs impractical. To overcome these limitations, we introduce Mol-MoE, a mixture-of-experts (MoE) architecture that enables efficient test-time steering of molecule generation without retraining. Central to our approach is a preference-based router training objective that incentivizes the router to combine experts in a way that aligns with user-specified trade-offs. This provides improved flexibility in exploring the chemical property space at test time, facilitating rapid trade-off exploration. Benchmarking against state-of-the-art methods, we show that Mol-MoE achieves superior sample quality and steerability.
2025-02-08
ArXiv (prépublication)
arxiv.org
arxiv.org
Improving Patient Safety Culture in Conflict-Affected Zones: A Cross-Sectional Survey of North Kivu Surgical Personnel in the Democratic Republic of the Congo.
Jacques Fadhili Bake
Claude Kasereka Masumbuko
Zacharie Tsongo Kibendelwa
Georges Bushu Lubuto
Jean‐Claude Mafuta Kyembwa
Esaie Kasereka Nzala
Papy Waleyirwe Kakule
Clovis Bwami Akumbi
Jean Zanga Kitutu
Tresor Basubi Wakilongo
Theophile Kubuya Hangi
Wilson Katembo Kwiraviwe
Benjamin Musemakweli
Beate Tshikudju Bahati
Steve Kisembo Bakabona
Dan Poenaru
BACKGROUND Patient safety culture significantly impacts outcomes in surgery, where preventable errors can occur. This study assessed patient… (voir plus) safety culture and its determinants in operating rooms across North Kivu, a conflict-affected province in the eastern Democratic Republic of the Congo (DRC). METHODS A descriptive multicenter cross-sectional study was conducted from July to September 2024 in five urban and six rural hospitals. The French version of the Hospital Survey on Patient Safety Culture (HSOPSC) questionnaire was administered to 328 operating room healthcare professionals. RESULTS The response rate was 78% (256 completed surveys). Urban hospitals accounted for 55.5% of respondents, who were 73.4% male and 62.5% under the age of 40. The overall composite score for patient safety culture was 63.2%. Teamwork (81.1%) and management support for patient safety (77.7%) received the highest positive responses, whereas error reporting (39.9%) and patient safety event reporting (50%) scored lower. Half (49.6%) of the respondents rated patient safety as excellent or very good. There were no significant differences in overall mean composite scores between urban and rural hospitals (p = 0.677) and between medical and paramedical staff (p = 0.694). CONCLUSIONS The patient safety culture rating in North Kivu falls below international standards, highlighting an urgent need for improvement, particularly in error response and event reporting. Developing a tailored patient safety bundle for the region is essential to enhance overall health outcomes.
2025-02-07
World Journal of Surgery (publié)
doi.org
Agency Is Frame-Dependent
David Abel
Andre Barreto
Michael Bowling
Will Dabney
Shi Dong
Steven Hansen
A. Harutyunyan
Khimya Khetarpal
Clare Lyle
Razvan Pascanu
Georgios Piliouras
Doina Precup
Jonathan Richens
Mark Rowland
Tom Schaul
Satinder Singh
Agency is a system's capacity to steer outcomes toward a goal, and is a central topic of study across biology, philosophy, cognitive science… (voir plus), and artificial intelligence. Determining if a system exhibits agency is a notoriously difficult question: Dennett (1989), for instance, highlights the puzzle of determining which principles can decide whether a rock, a thermostat, or a robot each possess agency. We here address this puzzle from the viewpoint of reinforcement learning by arguing that agency is fundamentally frame-dependent: Any measurement of a system's agency must be made relative to a reference frame. We support this claim by presenting a philosophical argument that each of the essential properties of agency proposed by Barandiaran et al. (2009) and Moreno (2018) are themselves frame-dependent. We conclude that any basic science of agency requires frame-dependence, and discuss the implications of this claim for reinforcement learning.
2025-02-06
ArXiv (prépublication)
arxiv.org
arxiv.org
Agency Is Frame-Dependent
David Abel
Andre Barreto
Michael Bowling
Will Dabney
Shi Dong
Steven Hansen
Anna Harutyunyan
Khimya Khetarpal
Clare Lyle
Razvan Pascanu
Georgios Piliouras
Doina Precup
Jonathan Richens
Mark Rowland
Tom Schaul
Satinder Singh
Agency is a system's capacity to steer outcomes toward a goal, and is a central topic of study across biology, philosophy, cognitive science… (voir plus), and artificial intelligence. Determining if a system exhibits agency is a notoriously difficult question: Dennett (1989), for instance, highlights the puzzle of determining which principles can decide whether a rock, a thermostat, or a robot each possess agency. We here address this puzzle from the viewpoint of reinforcement learning by arguing that agency is fundamentally frame-dependent: Any measurement of a system's agency must be made relative to a reference frame. We support this claim by presenting a philosophical argument that each of the essential properties of agency proposed by Barandiaran et al. (2009) and Moreno (2018) are themselves frame-dependent. We conclude that any basic science of agency requires frame-dependence, and discuss the implications of this claim for reinforcement learning.
2025-02-06
ArXiv (prépublication)
doi.org
arxiv.org