Author

Lacroix

1,407 papers · 1753–2026

1967peak 87 in 20242026

Frequent co-authors

Papers

showing 1–25 of 1,407 RSS

LLaMA: Open and Efficient Foundation Language Models

Hugo Touvron, Thibaut Lavril, Gautier Izacard +11 more

cs.CL · 27 Feb 2023 · cited 21,804× · arXiv ↗ · more like this ·

We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community.

Mistral 7B

Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch +15 more

cs.CL · 10 Oct 2023 · cited 3,967× · arXiv ↗ · more like this ·

We introduce Mistral 7B v0.1, a 7-billion-parameter language model engineered for superior performance and efficiency. Mistral 7B outperforms Llama 2 13B across all evaluated benchmarks, and Llama 1 34B in reasoning, mathematics, and code generation. Our model leverages grouped-query attention (GQA) for faster inference, coupled with sliding window attention (SWA) to effectively handle sequences of arbitrary length with a reduced inference cost. We also provide a model fine-tuned to follow instructions, Mistral 7B -- Instruct, that surpasses the Llama 2 13B -- Chat model both on human and automated benchmarks. Our models are released under the Apache 2.0 license.

BLOOM: A 176B-Parameter Open-Access Multilingual Language Model

BigScience Workshop, :, Teven Le Scao +391 more

cs.CL · November 2022 · cited 3,043× · arXiv ↗ · more like this ·

Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers. BLOOM is a decoder-only Transformer language model that was trained on the ROOTS corpus, a dataset comprising hundreds of sources in 46 natural and 13 programming languages (59 in total). We find that BLOOM achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning. To facilitate future research and applications using LLMs, we…

Mixtral of Experts

Albert Q. Jiang, Alexandre Sablayrolles, Antoine Roux +23 more

cs.LG · 8 Jan 2024 · cited 2,186× · arXiv ↗ · more like this ·

We introduce Mixtral 8x7B, a Sparse Mixture of Experts (SMoE) language model. Mixtral has the same architecture as Mistral 7B, with the difference that each layer is composed of 8 feedforward blocks (i.e. experts). For every token, at each layer, a router network selects two experts to process the current state and combine their outputs. Even though each token only sees two experts, the selected experts can be different at each timestep. As a result, each token has access to 47B parameters, but only uses 13B active parameters during inference. Mixtral was trained with a context size of 32k tokens and it outperforms or matches Llama 2 70B and GPT-3.5 across all evaluated benchmarks. In particular, Mixtral vastly outperforms Llama 2 70B on mathematics, code generation, and multilingual benchmarks. We also provide a model fine-tuned to follow instructions, Mixtral 8x7B - Instruct, that…

Quantum error correction below the surface code threshold

Rajeev Acharya, Laleh Aghababaie-Beni, Igor Aleiner +246 more

quant-ph · 24 Aug 2024 · cited 1,441× · arXiv ↗ · more like this ·

Quantum error correction provides a path to reach practical quantum computing by combining multiple physical qubits into a logical qubit, where the logical error rate is suppressed exponentially as more qubits are added. However, this exponential suppression only occurs if the physical error rate is below a critical threshold. In this work, we present two surface code memories operating below this threshold: a distance-7 code and a distance-5 code integrated with a real-time decoder. The logical error rate of our larger quantum memory is suppressed by a factor of $Λ$ = 2.14 $\pm$ 0.02 when increasing the code distance by two, culminating in a 101-qubit distance-7 code with 0.143% $\pm$ 0.003% error per cycle of error correction. This logical memory is also beyond break-even, exceeding its best physical qubit's lifetime by a factor of 2.4 $\pm$ 0.3. We maintain below-threshold…

Marriage Market, Divorce Legislation, and Household Labor Supply

Pierre‐André Chiappori, Bernard Fortin, Guy Lacroix

Gender Studies · Journal of Political Economy · February 2002 · cited 891× · DOI ↗ · more like this ·

This paper provides a theoretical framework for analyzing the impact of the marriage market and divorce legislation on household labor supply. In our approach, the sex ratio in the marriage market and the rules governing divorce are examples of "distribution factors." These factors are defined as variables that affect the household members' bargaining position but not preferences or the joint budget set. We extend the collective labor supply model developed by Chiappori to allow for distribution factors. We show that our model imposes new restrictions on the labor supply functions and eases the identification of individual preferences and the intrahousehold decision process. The model is estimated using PSID data for the year 1988. Our results do not reject the restrictions imposed by the model. Also, the sex ratio and divorce laws deemed favorable to women are found to affect labor…

Maintaining Mobility in Late Life. II. Smoking, Alcohol Consumption, Physical Activity, and Body Mass Index

Andrea Z. LaCroix, Jack M. Guralnik, Lisa Berkman +2 more

Health · American Journal of Epidemiology · 15 Apr 1993 · cited 524× · DOI ↗ · more like this ·

While positive health behaviors have been shown to extend life, their association with extending active life has not been well investigated. In this report, several health behaviors were investigated in relation to maintaining mobility during 4 years of follow-up among 6,981 men and women aged 65 years and older with intact mobility at baseline between 1981 and 1983 who lived in one of three communities: East Boston, Massachusetts; Iowa and Washington counties, Iowa; and New Haven, Connecticut. Intact mobility, defined as the ability to climb up and down stairs and walk a half mile, was determined annually by interview, and study subjects were classified into one of three categories at the end of 4 years of follow-up: 1) maintained mobility (55.1%); 2) lost mobility (36.2%); or 3) died without evidence of having lost mobility prior to death (8.7%). After adjustment for age and all of…

Canonical Tensor Decomposition for Knowledge Base Completion

Timothée Lacroix, Nicolas Usunier, Guillaume Obozinski

stat.ML · 19 Jun 2018 · cited 455× · arXiv ↗ · more like this ·

The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) (Hitchcock, 1927) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear $p$-norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.

Patterns of Technology Use Among Older Adults With and Without Disabilities

Nancy Gell, Dori E. Rosenberg, George Demiris +2 more

Demography · The Gerontologist · 30 Dec 2013 · cited 450× · DOI ↗ · more like this ·

PURPOSE OF THE STUDY: The purpose of this study was to describe prevalence of technology use among adults ages 65 and older, particularly for those with disability and activity-limiting symptoms and impairments. DESIGN AND METHODS: Data from the 2011 National Health and Aging Trends Study, a nationally representative sample of community-dwelling Medicare beneficiaries (N = 7,609), were analyzed. Analysis consisted of technology use (use of e-mail/text messages and the internet) by sociodemographic and health characteristics and prevalence ratios for technology usage by disability status. RESULTS: Forty percent of older adults used e-mail or text messaging and 42.7% used the internet. Higher prevalence of technology use was associated with younger age, male sex, white race, higher education level, and being married (all p values <.001). After adjustment for sociodemographic and health…

PyTorch-BigGraph: A Large-scale Graph Embedding System

Adam Lerer, Ledell Wu, Jiajun Shen +4 more

cs.LG · March 2019 · cited 422× · arXiv ↗ · more like this ·

Graph embedding methods produce unsupervised node features from graphs that can then be used for a variety of machine learning tasks. Modern graphs, particularly in industrial applications, contain billions of nodes and trillions of edges, which exceeds the capability of existing embedding systems. We present PyTorch-BigGraph (PBG), an embedding system that incorporates several modifications to traditional multi-relation embedding systems that allow it to scale to graphs with billions of nodes and trillions of edges. PBG uses graph partitioning to train arbitrarily large embeddings on either a single machine or in a distributed environment. We demonstrate comparable performance with existing embedding systems on common benchmarks, while allowing for scaling to arbitrarily large graphs and parallelization on multiple machines. We train and evaluate embeddings on several large social…

Tensor Decompositions for temporal knowledge base completion

Timothée Lacroix, Guillaume Obozinski, Nicolas Usunier

stat.ML · 10 Apr 2020 · cited 387× · arXiv ↗ · more like this ·

Most algorithms for representation learning and link prediction in relational data have been designed for static data. However, the data they are applied to usually evolves with time, such as friend graphs in social networks or user interactions with items in recommender systems. This is also the case for knowledge bases, which contain facts such as (US, has president, B. Obama, [2009-2017]) that are valid only at certain points in time. For the problem of link prediction under temporal constraints, i.e., answering queries such as (US, has president, ?, 2012), we propose a solution inspired by the canonical decomposition of tensors of order 4. We introduce new regularization schemes and present an extension of ComplEx (Trouillon et al., 2016) that achieves state-of-the-art performance. Additionally, we propose a new dataset for knowledge base completion constructed from Wikidata, larger…

Draft, Sketch, and Prove: Guiding Formal Theorem Provers with Informal Proofs

Albert Q. Jiang, Sean Welleck, Jin Peng Zhou +6 more

cs.AI · October 2022 · cited 361× · arXiv ↗ · more like this ·

The formalization of existing mathematical proofs is a notoriously difficult process. Despite decades of research on automation and proof assistants, writing formal proofs remains arduous and only accessible to a few experts. While previous studies to automate formalization focused on powerful search algorithms, no attempts were made to take advantage of available informal proofs. In this work, we introduce Draft, Sketch, and Prove (DSP), a method that maps informal proofs to formal proof sketches, and uses the sketches to guide an automated prover by directing its search to easier sub-problems. We investigate two relevant setups where informal proofs are either written by humans or generated by a language model. Our experiments and ablation studies show that large language models are able to produce well-structured formal sketches that follow the same reasoning steps as the informal…

Quantum Computing for High-Energy Physics: State of the Art and Challenges. Summary of the QC4HEP Working Group

Alberto Di Meglio, Karl Jansen, Ivano Tavernelli +43 more

quant-ph · 6 Jul 2023 · cited 314× · arXiv ↗ · more like this ·

Quantum computers offer an intriguing path for a paradigmatic change of computing in the natural sciences and beyond, with the potential for achieving a so-called quantum advantage, namely a significant (in some cases exponential) speed-up of numerical simulations. The rapid development of hardware devices with various realizations of qubits enables the execution of small scale but representative applications on quantum computers. In particular, the high-energy physics community plays a pivotal role in accessing the power of quantum computing, since the field is a driving source for challenging computational problems. This concerns, on the theoretical side, the exploration of models which are very hard or even impossible to address with classical techniques and, on the experimental side, the enormous data challenge of newly emerging experiments, such as the upgrade of the Large Hadron…

A Test of the Unitary and Collective Models of Household Labour Supply

Bernard Fortin, Guy Lacroix

Gender Studies · The Economic Journal · July 1997 · cited 312× · DOI ↗ · more like this ·

This paper tests the unitary and the collective household labour supply models within a structural framework. A nonlinear unrestricted household labour supply system is postulated and the set of parametric restrictions imposed by each model is derived. Our results generally reject the income pooling restrictions embodied in the unitary model. Also, the symmetry of cross‐wage effects is rejected. However, the collective model is not rejected for groups with no pre‐school children.

HyperTree Proof Search for Neural Theorem Proving

Guillaume Lample, Marie-Anne Lachaux, Thibaut Lavril +5 more

cs.CL · 23 May 2022 · cited 270× · arXiv ↗ · more like this ·

We propose an online training procedure for a transformer-based automated theorem prover. Our approach leverages a new search algorithm, HyperTree Proof Search (HTPS), inspired by the recent success of AlphaZero. Our model learns from previous proof searches through online training, allowing it to generalize to domains far from the training distribution. We report detailed ablations of our pipeline's main components by studying performance on three environments of increasing complexity. In particular, we show that with HTPS alone, a model trained on annotated proofs manages to prove 65.4% of a held-out set of Metamath theorems, significantly outperforming the previous state of the art of 56.5% by GPT-f. Online training on these unproved theorems increases accuracy to 82.6%. With a similar computational budget, we improve the state of the art on the Lean-based miniF2F-curriculum dataset…

Preferences; Putting More Knowledge into Queries

Michel Lacroix, P. Lavency

Computer Networks and Communications · Very Large Data Bases · September 1987 · cited 242× · OpenAlex ↗ · more like this ·

Classical query languages provide a way to express mandatory qualifications on the data to be retrieved. They do not feature facilities for expressing preferences or desirable qualifications. The need for preferences is illustrated in a software engineering framework. A preference mechanism is then presented as an extension of a language of the Domain Relational Calculus family, and the expressive power of the resulting language is discussed. The proposed mechanisms are shown to effectively allow the use of queries for supporting software configuration management functions.

Evaluation of Social Isolation, Loneliness, and Cardiovascular Disease Among Older Women in the US

Natalie M. Golaszewski, Andrea Z. LaCroix, Job Godino +9 more

Health · JAMA Network Open · 2 Feb 2022 · cited 217× · DOI ↗ · more like this ·

Importance: Social isolation and loneliness are increasing public health concerns and have been associated with increased risk of cardiovascular disease (CVD) among older adults. Objective: To examine the associations of social isolation and loneliness with incident CVD in a large cohort of postmenopausal women and whether social support moderated these associations. Design, Setting, and Participants: This prospective cohort study, conducted from March 2011 through March 2019, included community-living US women aged 65 to 99 years from the Women's Health Initiative Extension Study II who had no history of myocardial infarction, stroke, or coronary heart disease. Exposures: Social isolation and loneliness were ascertained using validated questionnaires. Main Outcomes and Measures: The main outcome was major CVD, which was physician adjudicated using medical records and included coronary…

A distributed tasks allocation scheme in multi-UAV context

Thomas Lemaire, Rachid Alami, Simon Lacroix

Computer Networks and Communications · 2004 · cited 196× · DOI ↗ · more like this ·

This paper deals with the task allocation problem in multi-robot systems. We propose a completely distributed architecture, where robots dynamically allocate their tasks while they are building their plans. We first focus on the problem of simple "goto" tasks allocation: our approach involves an incremental task allocation algorithm based on the Contract-Net protocol. We introduce a parameter called equity coefficient in order to equilibrate the workload between the different robots and to control the triggering of the auction process. Then, we address the problem raised by temporal constraints between tasks by dynamically specifying temporary hierarchies among the tasks. Tests run in simulation quantify the benefits of our improvements.

Autonomous Rover Navigation on Unknown Terrains: Functions and Integration

Simon Lacroix, Anthony Mallet, David Bonnafous +4 more

Computer Vision and Pattern Recognition · The International Journal of Robotics Research · October 2002 · cited 184× · DOI ↗ · more like this ·

Autonomous long-range navigation in partially known planetary-like terrains is still an open challenge for robotics. Navigating hundreds of meters without any human intervention requires the robot to be able to build various representations of its environment, to plan and execute trajectories according to the kind of terrain traversed, to control its motions and to localize itself as it moves. All these activities have to be scheduled, triggered, controlled and interrupted according to the rover context. In this paper, we briefly review some functionalities that have been developed in our laboratory, and implemented on board the Marsokhod model robot, Lama. We then present how the various concurrent instances of the perception, localization and motion generation functionalities are integrated. Experimental results illustrate the functionalities throughout the paper.

Images of Animated Others: The Orientalization of Disney's Cartoon Heroines From The Little Mermaid to The Hunchback of Notre Dame

Celeste Lacroix

Gender Studies · Popular Communication · November 2004 · cited 180× · DOI ↗ · more like this ·

Because of the success of the first six animated feature films produced in the "new era" of Disney animation (The Little Mermaid, 1989; Beauty and the Beast, 1991; Aladdin, 1993; The Lion King, 1994; Pocahontas, 1995; The Hunchback of Notre Dame, 1996) and their mass merchandising, Disney animated characters became ubiquitous for children of the 1990s. Although Bell, Haas, and Sells (1995) suggested that Disney films present a "sanitization of violence, sexuality, and political struggle concomitant with an erasure or repression of difference" (p. 7), an increasing emphasis on sexuality and the exotic is evident in the construction of the female heroines in these films, particularly in the female characters of color. This article analyzes what may be referred to in Said's (1978) terminology as the orientalization of women of color in five of these six Disney animated films and posits how…

Canadian Refugee Policy and the Social Construction of the Refugee Claimant Subjectivity: Understanding Refugeeness

Marie Lacroix

Sociology and Political Science · Journal of Refugee Studies · June 2004 · cited 179× · DOI ↗ · more like this ·

This paper introduces a theoretical framework to fill a gap in our understanding of refugeeness. Based on material gathered from a series of interviews undertaken as part of an exploratory study completed in Montreal, Canada, refugee claimant subjectivity is empirically explored by examining the social construction of refugees' subjectivity in three major spheres of life: work, family and state. Within this theoretical perspective two linked research questions are addressed, namely: (1) What is the impact of refugee policy on refugee claimants? and (2) How is their subjectivity altered by the refugee determination process? The objective is to understand the impact of a particular discourse on an individual's subjectivity in relation to that discourse. Refugee claimant subjectivity is analysed through those conscious thoughts and emotions expressed by subjects, their way of understanding…

Psychological Barriers to Energy Conservation Behavior: The Role of Worldviews and Climate Change Risk Perception

Karine Lacroix, Robert Gifford

Sociology and Political Science · Environment and Behavior · 23 Jun 2017 · cited 166× · DOI ↗ · more like this ·

We proposed and tested a conceptual model of how cultural cognition worldviews, climate change risk perception, and psychological barriers are related to reported energy conservation behavior frequency. Egalitarian and communitarian worldviews were correlated with heightened climate change risk perception, and egalitarian worldviews were correlated with weaker perceived barriers to reported energy conservation behavior. Heightened climate change risk perception was, in turn, associated with fewer perceived barriers to engagement in energy conservation behavior and more reported energy conservation behaviors. The relation between cultural worldviews and perceived barriers was partly mediated by climate change risk perception. Individuals with distinct worldviews perceived psychological barriers differently, and some barrier components were more strongly related to energy conservation…

Robust regression through the Huber’s criterion and adaptive lasso penalty

Sophie Lambert‐Lacroix, Laurent Zwald

Statistics and Probability · Electronic Journal of Statistics · 2011 · cited 149× · DOI ↗ · more like this ·

The Huber’s Criterion is a useful method for robust regression. The adaptive least absolute shrinkage and selection operator (lasso) is a popular technique for simultaneous estimation and variable selection. The adaptive weights in the adaptive lasso allow to have the oracle properties. In this paper we propose to combine the Huber’s criterion and adaptive penalty as lasso. This regression technique is resistant to heavy-tailed errors or outliers in the response. Furthermore, we show that the estimator associated with this procedure enjoys the oracle properties. This approach is compared with LAD-lasso based on least absolute deviation with adaptive lasso. Extensive simulation studies demonstrate satisfactory finite-sample performance of such procedure. A real example is analyzed for illustration purposes.

Awakening Islam: the politics of religious dissent in contemporary Saudi Arabia

Stéphane Lacroix

Sociology and Political Science · Choice Reviews Online · February 2012 · cited 130× · DOI ↗ · more like this ·

Explores the political culture in Saudi Arabia, describing how the Middle Eastern country became a place of refuge for Islamist militants beginning in the 1950s, which led to a social movement that focused on political activism and local religious ideas, and discussing the effects of this social movement on Saudi society.

Realizing Quantum Convolutional Neural Networks on a Superconducting Quantum Processor to Recognize Quantum Phases

Johannes Herrmann, Sergi Masot Llima, Ants Remm +15 more

quant-ph · 13 Sep 2021 · cited 127× · arXiv ↗ · more like this ·

Quantum computing crucially relies on the ability to efficiently characterize the quantum states output by quantum hardware. Conventional methods which probe these states through direct measurements and classically computed correlations become computationally expensive when increasing the system size. Quantum neural networks tailored to recognize specific features of quantum states by combining unitary operations, measurements and feedforward promise to require fewer measurements and to tolerate errors. Here, we realize a quantum convolutional neural network (QCNN) on a 7-qubit superconducting quantum processor to identify symmetry-protected topological (SPT) phases of a spin model characterized by a non-zero string order parameter. We benchmark the performance of the QCNN based on approximate ground states of a family of cluster-Ising Hamiltonians which we prepare using a…