Nous utilisons des témoins pour analyser le trafic et l’utilisation de notre site web, afin de personnaliser votre expérience. Vous pouvez désactiver ces technologies à tout moment, mais cela peut restreindre certaines fonctionnalités du site. Consultez notre Politique de protection de la vie privée pour en savoir plus.
Paramètre des cookies
Vous pouvez activer et désactiver les types de cookies que vous souhaitez accepter. Cependant certains choix que vous ferez pourraient affecter les services proposés sur nos sites (ex : suggestions, annonces personnalisées, etc.).
Cookies essentiels
Ces cookies sont nécessaires au fonctionnement du site et ne peuvent être désactivés. (Toujours actif)
Cookies analyse
Acceptez-vous l'utilisation de cookies pour mesurer l'audience de nos sites ?
Multimedia Player
Acceptez-vous l'utilisation de cookies pour afficher et vous permettre de regarder les contenus vidéo hébergés par nos partenaires (YouTube, etc.) ?
ScatterMoE is an implementation of Sparse Mixture-of-Experts (SMoE) on GPUs. ScatterMoE builds upon techniques in existing implementations, … (voir plus)and overcoming some of the current limitations to improve batched inference, training speed, and memory footprint. This implementation achieves this by avoiding padding and making excessive copies of the input. We also fuse expert linear transforms and reordering operations with ParallelLinear, a module that can be used to extend the concept of SMoEs. We benchmark our implementation against Megablocks, and show that it enables a higher throughput and lower memory footprint. We also show how ParallelLinear enables extension of the Mixture-of-Experts concept by demonstrating with an implementation of Mixture-of-Attention.
Measuring personal disclosures made in human-chatbot interactions can provide a better understanding of users' AI literacy and facilitate pr… (voir plus)ivacy research for large language models (LLMs). We run an extensive, fine-grained analysis on the personal disclosures made by real users to commercial GPT models, investigating the leakage of personally identifiable and sensitive information. To understand the contexts in which users disclose to chatbots, we develop a taxonomy of tasks and sensitive topics, based on qualitative and quantitative analysis of naturally occurring conversations. We discuss these potential privacy harms and observe that: (1) personally identifiable information (PII) appears in unexpected contexts such as in translation or code editing (48% and 16% of the time, respectively) and (2) PII detection alone is insufficient to capture the sensitive topics that are common in human-chatbot interactions, such as detailed sexual preferences or specific drug use habits. We believe that these high disclosure rates are of significant importance for researchers and data curators, and we call for the design of appropriate nudging mechanisms to help users moderate their interactions.
Common self-improvement approaches for large language models (LLMs), such as STaR (Zelikman et al., 2022), iteratively fine-tune LLMs on sel… (voir plus)f-generated solutions to improve their problem-solving ability. However, these approaches discard the large amounts of incorrect solutions generated during this process, potentially neglecting valuable information in such solutions. To address this shortcoming, we propose V-STaR that utilizes both the correct and incorrect solutions generated during the self-improvement process to train a verifier using DPO that judges correctness of model-generated solutions. This verifier is used at inference time to select one solution among many candidate solutions. Running V-STaR for multiple iterations results in progressively better reasoners and verifiers, delivering a 4% to 17% test accuracy improvement over existing self-improvement and verification approaches on common code generation and math reasoning benchmarks with LLaMA2 models.
SETTING
In Canada's federated healthcare system, 13 provincial and territorial jurisdictions have independent responsibility to collect data… (voir plus) to inform health policies. During the COVID-19 pandemic (2020-2023), national and regional sero-surveys mostly drew upon existing infrastructure to quickly test specimens and collect data but required cross-jurisdiction coordination and communication.
INTERVENTION
There were 4 national and 7 regional general population SARS-CoV-2 sero-surveys. Survey methodologies varied by participant selection approaches, assay choices, and reporting structures. We analyzed Canadian pandemic sero-surveillance initiatives to identify key learnings to inform future pandemic planning.
OUTCOMES
Over a million samples were tested for SARS-CoV-2 antibodies from 2020 to 2023 but siloed in 11 distinct datasets. Most national sero-surveys had insufficient sample size to estimate regional prevalence; differences in methodology hampered cross-regional comparisons of regional sero-surveys. Only four sero-surveys included questionnaires. Sero-surveys were not directly comparable due to different assays, sampling methodologies, and time-frames. Linkage to health records occurred in three provinces only. Dried blood spots permitted sample collection in remote populations and during stay-at-home orders.
IMPLICATIONS
To provide timely, high-quality information for public health decision-making, routine sero-surveillance systems must be adaptable, flexible, and scalable. National capability planning should include consortiums for assay design and validation, defined mechanisms to improve test capacity, base documents for data linkage and material transfer across jurisdictions, and mechanisms for real-time communication of data. Lessons learned will inform incorporation of a robust sero-survey program into routine surveillance with strategic sampling and capacity to adapt and scale rapidly as a part of a comprehensive national pandemic response plan.
