UNLEARNING GEO-CULTURAL STEREOTYPES IN MULTILINGUAL LLMS
Alireza Dehghanpour Farashah
Aditi Khandelwal
As multilingual generative models become more widely used, most safety and fairness evaluation techniques still focus on English-language re… (voir plus)sources, while overlooking important cross-cultural factors. This limitation raises concerns about fairness and safety, particularly regarding geoculturally situated stereotypes that hinder the models’ global inclusivity. In this work, we present preliminary findings on the impact of stereotype unlearning across languages, specifically in English, French, and Hindi. Using an adapted version of the SeeGULL dataset, we analyze how unlearning stereotypes in one language influences other languages within multilingual large language models. Our study evaluates two model families, Llama-3.1-8B and Aya-Expanse-8B, to assess whether unlearning in one linguistic context transfers across languages, potentially mitigating or exacerbating biases in multilingual settings.
WebMMU: A Benchmark for Multimodal Multilingual Website Understanding and Code Generation
Rabiul Awal
Mahsa Massoud
Zichao Li
Aarash Feizi
Suyuchen Wang
David Vazquez
Juan A. Rodriguez
Perouz Taslakian
Spandana Gella
Sai Rajeswar
Understanding diverse web data and automating web development presents an exciting challenge for agentic AI. While existing benchmarks addre… (voir plus)ss isolated web-based tasks—such as website-based Visual Question Answering (VQA) and UI-to-code generation—they lack a unified evaluation suite for assessing web agents that interact with and reason about web environments. We introduce WebMMU, a large-scale benchmark for evaluating AI-driven web agents across multilingual website VQA, HTML/CSS/JavaScript code editing, and sketch-to-code generation. WebMMU provides a comprehensive evaluation suite with real-world website data, multi-step reasoning tasks, and functional UI understanding. Benchmarking state-of-the-art multimodal models on WebMMU reveals significant limitations in web-based reasoning, layout understanding, and structured code generation, particularly in preserving UI hierarchy, handling multilingual content, and producing robust, functional code. While most existing models are optimized for English-only settings, WebMMU highlights the challenges of cross-lingual adaptation in real-world web development. These findings expose critical gaps in current models’ ability to understand website structures, execute user instructions, and generate high-quality web code, underscoring the need for more advanced multimodal reasoning in AI-driven web understanding and development.
Automated diagnosis of usual interstitial pneumonia on chest CT via the mean curvature of isophotes
Peter Savadjiev
Morteza Rezanejad
Sahir Bhatnagar
David Camirand
Claude Kauffmann
Ronald J Dandurand
Patrick Bourgouin
Carl Chartrand-Lefebvre
Alexandre Semionov
AI Automatons: AI Systems Intended to Imitate Humans
Solon Barocas
Su Lin Blodgett
Lisa Egede
Alicia DeVrio
Myra Cheng
There is a growing proliferation of AI systems designed to mimic people's behavior, work, abilities, likenesses, or humanness -- systems we … (voir plus)dub AI automatons. Individuals, groups, or generic humans are being simulated to produce creative work in their styles, to respond to surveys in their places, to probe how they would use a new system before deployment, to provide users with assistance and companionship, and to anticipate their possible future behavior and interactions with others, just to name a few applications. The research, design, deployment, and availability of such AI systems have, however, also prompted growing concerns about a wide range of possible legal, ethical, and other social impacts. To both 1) facilitate productive discussions about whether, when, and how to design and deploy such systems, and 2) chart the current landscape of existing and prospective AI automatons, we need to tease apart determinant design axes and considerations that can aid our understanding of whether and how various design choices along these axes could mitigate -- or instead exacerbate -- potential adverse impacts that the development and use of AI automatons could give rise to. In this paper, through a synthesis of related literature and extensive examples of existing AI systems intended to mimic humans, we develop a conceptual framework to help foreground key axes of design variations and provide analytical scaffolding to foster greater recognition of the design choices available to developers, as well as the possible ethical implications these choices might have.
Beyond Cosine Decay: On the effectiveness of Infinite Learning Rate Schedule for Continual Pre-training
Vaibhav Singh
Paul Janson
Paria Mehrbod
Adam Ibrahim
Benjamin Thérien
The ever-growing availability of unlabeled data presents both opportunities and challenges for training artificial intelligence systems. Whi… (voir plus)le self-supervised learning (SSL) has emerged as a powerful paradigm for extracting meaningful representations from vast amounts of unlabeled data, existing methods still struggle to adapt to the non-stationary, non-IID nature of real-world data streams without forgetting previously learned knowledge. Recent works have adopted a repeated cosine annealing schedule for large-scale continual pre-training; however, these schedules (1) inherently cause forgetting during the re-warming phase and (2) have not been systematically compared to existing continual SSL methods. In this work, we systematically compare the widely used cosine schedule with the recently proposed infinite learning rate schedule and empirically find the latter to be a more effective alternative. Our extensive empirical evaluation across diverse image and language datasets demonstrates that the infinite learning rate schedule consistently enhances continual pre-training performance compared to a repeated cosine decay without being restricted to a fixed iteration budget. For instance, in a small-scale MAE pre-training setup, it outperforms several strong baselines from the literature. We then scale up our experiments to larger MAE pre-training and autoregressive language model pre-training. Our results show that the infinite learning rate schedule remains effective at scale, surpassing repeated cosine decay for both MAE pre-training and zero-shot LM benchmarks.
EarthView: A Large Scale Remote Sensing Dataset for Self-Supervision
Diego Velazquez
Pau Rodriguez
Sergio Alonso
Josep M. Gonfaus
Jordi Gonzalez
Gerardo Richarte
Javier Marin
Alexandre Lacoste
This paper presents EarthView, a comprehensive dataset specifically designed for self-supervision on remote sensing data, intended to enhanc… (voir plus)e deep learning applications on Earth monitoring tasks. The dataset spans 15 tera pixels of global remote-sensing data, combining imagery from a diverse range of sources, including NEON, Sentinel, and a novel release of 1m spatial resolution data from Satellogic. Our dataset provides a wide spectrum of image data with varying resolutions, harnessed from different sensors and organized coherently into an accessible HuggingFace dataset in parquet format. This data spans five years, from 2017 to 2022. Accompanying the dataset, we introduce EarthMAE, a tailored Masked Autoencoder, developed to tackle the distinct challenges of remote sensing data. Trained in a self-supervised fashion, EarthMAE effectively processes different data modalities such as hyperspectral, multispectral, topographical data, segmentation maps, and temporal structure. This model helps us show that pre-training on Satellogic data improves performance on downstream tasks. While there is still a gap to fill in MAE for heterogeneous data, we regard this innovative combination of an expansive, diverse dataset and a versatile model adapted for self-supervised learning as a stride forward in deep learning for Earth monitoring.
Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Experts
Marta Skreta
Tara Akhound-Sadegh
Viktor Ohanesian
Roberto Bondesan
Alan Aspuru-Guzik
Arnaud Doucet
Rob Brekelmans
Alexander Tong
While score-based generative models are the model of choice across diverse domains, there are limited tools available for controlling infere… (voir plus)nce-time behavior in a principled manner, e.g. for composing multiple pretrained models. Existing classifier-free guidance methods use a simple heuristic to mix conditional and unconditional scores to approximately sample from conditional distributions. However, such methods do not approximate the intermediate distributions, necessitating additional 'corrector' steps. In this work, we provide an efficient and principled method for sampling from a sequence of annealed, geometric-averaged, or product distributions derived from pretrained score-based models. We derive a weighted simulation scheme which we call Feynman-Kac Correctors (FKCs) based on the celebrated Feynman-Kac formula by carefully accounting for terms in the appropriate partial differential equations (PDEs). To simulate these PDEs, we propose Sequential Monte Carlo (SMC) resampling algorithms that leverage inference-time scaling to improve sampling quality. We empirically demonstrate the utility of our methods by proposing amortized sampling via inference-time temperature annealing, improving multi-objective molecule generation using pretrained models, and improving classifier-free guidance for text-to-image generation. Our code is available at https://github.com/martaskrt/fkc-diffusion.
Hardware Synthesizable Exceptions using Continuations
Paul Teng
LLM-Safety Evaluations Lack Robustness
Tim Beyer
Sophie Xhonneux
Simon Geisler
Leo Schwinn
Stephan Günnemann
In this paper, we argue that current safety alignment research efforts for large language models are hindered by many intertwined sources of… (voir plus) noise, such as small datasets, methodological inconsistencies, and unreliable evaluation setups. This can, at times, make it impossible to evaluate and compare attacks and defenses fairly, thereby slowing progress. We systematically analyze the LLM safety evaluation pipeline, covering dataset curation, optimization strategies for automated red-teaming, response generation, and response evaluation using LLM judges. At each stage, we identify key issues and highlight their practical impact. We also propose a set of guidelines for reducing noise and bias in evaluations of future attack and defense papers. Lastly, we offer an opposing perspective, highlighting practical reasons for existing limitations. We believe that addressing the outlined problems in future research will improve the field's ability to generate easily comparable results and make measurable progress.
Variation Matters: from Mitigating to Embracing Zero-Shot NAS Ranking Function Variation
Pavel Rumiantsev
Neural Architecture Search (NAS) is a powerful automatic alternative to manual design of a neural network. In the zero-shot version, a fast … (voir plus)ranking function is used to compare architectures without training them. The outputs of the ranking functions often vary significantly due to different sources of randomness, including the evaluated architecture's weights' initialization or the batch of data used for calculations. A common approach to addressing the variation is to average a ranking function output over several evaluations. We propose taking into account the variation in a different manner, by viewing the ranking function output as a random variable representing a proxy performance metric. During the search process, we strive to construct a stochastic ordering of the performance metrics to determine the best architecture. Our experiments show that the proposed stochastic ordering can effectively boost performance of a search on standard benchmark search spaces.
CrystalGym: A New Benchmark for Materials Discovery Using Reinforcement Learning
Prashant Govindarajan
Mathieu Reymond
Antoine Clavaud
Mariano Phielipp
Santiago Miret
*In silico* design and optimization of new materials primarily relies on high-accuracy atomic simulators that perform density functional the… (voir plus)ory (DFT) calculations. While recent works showcase the strong potential of machine learning to accelerate the material design process, they mostly consist of generative approaches that do not use direct DFT signals as feedback to improve training and generation mainly due to DFT's high computational cost. To aid the adoption of direct DFT signals in the materials design loop through online reinforcement learning (RL), we propose **CrystalGym**, an open-source RL environment for crystalline material discovery. Using CrystalGym, we benchmark value- and policy-based reinforcement learning algorithms for designing various crystals conditioned on target properties. Concretely, we optimize for challenging properties like the band gap, bulk modulus, and density, which are directly calculated from DFT in the environment. While none of the algorithms we benchmark solve all CrystalGym tasks, our extensive experiments and ablations show different sample efficiencies and ease of convergence to optimality for different algorithms and environment settings. Our goal is for CrystalGym to serve as a test bed for reinforcement learning researchers and material scientists to address these real-world design problems with practical applications. Furthermore, we introduce a novel class of challenges for reinforcement learning methods dealing with time-consuming reward signals, paving the way for future interdisciplinary research for machine learning motivated by real-world applications.
Development and Feasibility Study of HOPE Model for Prediction of Depression Among Older Adults Using Wi-Fi-based Motion Sensor Data: Machine Learning Study.
Shayan Nejadshamsi
Vania Karami
Negar Ghourchian
Howard Bergman
Roland Grad
Machelle Wilchesky
Vladimir Khanassov
Isabelle Vedel
BACKGROUND Depression, characterized by persistent sadness and loss of interest in daily activities, greatly reduces quality of life. Early … (voir plus)detection is vital for effective treatment and intervention. While many studies use wearable devices to classify depression based on physical activity, these often rely on intrusive methods. Additionally, most depression classification studies involve large participant groups and use single-stage classifiers without explainability. OBJECTIVE This study aims to assess the feasibility of classifying depression using nonintrusive Wi-Fi-based motion sensor data using a novel machine learning model on a limited number of participants. We also conduct an explainability analysis to interpret the model's predictions and identify key features associated with depression classification. METHODS In this study, we recruited adults aged 65 years and older through web-based and in-person methods, supported by a McGill University health care facility directory. Participants provided consent, and we collected 6 months of activity and sleep data via nonintrusive Wi-Fi-based sensors, along with Edmonton Frailty Scale and Geriatric Depression Scale data. For depression classification, we proposed a HOPE (Home-Based Older Adults' Depression Prediction) machine learning model with feature selection, dimensionality reduction, and classification stages, evaluating various model combinations using accuracy, sensitivity, precision, and F1-score. Shapely addictive explanations and local interpretable model-agnostic explanations were used to explain the model's predictions. RESULTS A total of 6 participants were enrolled in this study; however, 2 participants withdrew later due to internet connectivity issues. Among the 4 remaining participants, 3 participants were classified as not having depression, while 1 participant was identified as having depression. The most accurate classification model, which combined sequential forward selection for feature selection, principal component analysis for dimensionality reduction, and a decision tree for classification, achieved an accuracy of 87.5%, sensitivity of 90%, and precision of 88.3%, effectively distinguishing individuals with and those without depression. The explainability analysis revealed that the most influential features in depression classification, in order of importance, were "average sleep duration," "total number of sleep interruptions," "percentage of nights with sleep interruptions," "average duration of sleep interruptions," and "Edmonton Frailty Scale." CONCLUSIONS The findings from this preliminary study demonstrate the feasibility of using Wi-Fi-based motion sensors for depression classification and highlight the effectiveness of our proposed HOPE machine learning model, even with a small sample size. These results suggest the potential for further research with a larger cohort for more comprehensive validation. Additionally, the nonintrusive data collection method and model architecture proposed in this study offer promising applications in remote health monitoring, particularly for older adults who may face challenges in using wearable devices. Furthermore, the importance of sleep patterns identified in our explainability analysis aligns with findings from previous research, emphasizing the need for more in-depth studies on the role of sleep in mental health, as suggested in the explainable machine learning study.