Showing 1–15 of 15 results for author: Gaucher, S

Improved Algorithms for Contextual Dynamic Pricing

Authors: Matilde Tullii, Solenne Gaucher, Nadav Merlis, Vianney Perchet

Abstract: In contextual dynamic pricing, a seller sequentially prices goods based on contextual information. Buyers will purchase products only if the prices are below their valuations. The goal of the seller is to design a pricing strategy that collects as much revenue as possible. We focus on two different valuation models. The first assumes that valuations linearly depend on the context and are further d… ▽ More In contextual dynamic pricing, a seller sequentially prices goods based on contextual information. Buyers will purchase products only if the prices are below their valuations. The goal of the seller is to design a pricing strategy that collects as much revenue as possible. We focus on two different valuation models. The first assumes that valuations linearly depend on the context and are further distorted by noise. Under minor regularity assumptions, our algorithm achieves an optimal regret bound of $\tilde{\mathcal{O}}(T^{2/3})$, improving the existing results. The second model removes the linearity assumption, requiring only that the expected buyer valuation is $β$-Hölder in the context. For this model, our algorithm obtains a regret $\tilde{\mathcal{O}}(T^{d+2β/d+3β})$, where $d$ is the dimension of the context space. △ Less

Submitted 17 June, 2024; originally announced June 2024.

Feature-Based Online Bilateral Trade

Authors: Solenne Gaucher, Martino Bernasconi, Matteo Castiglioni, Andrea Celli, Vianney Perchet

Abstract: Bilateral trade models the problem of facilitating trades between a seller and a buyer having private valuations for the item being sold. In the online version of the problem, the learner faces a new seller and buyer at each time step, and has to post a price for each of the two parties without any knowledge of their valuations. We consider a scenario where, at each time step, before posting price… ▽ More Bilateral trade models the problem of facilitating trades between a seller and a buyer having private valuations for the item being sold. In the online version of the problem, the learner faces a new seller and buyer at each time step, and has to post a price for each of the two parties without any knowledge of their valuations. We consider a scenario where, at each time step, before posting prices the learner observes a context vector containing information about the features of the item for sale. The valuations of both the seller and the buyer follow an unknown linear function of the context. In this setting, the learner could leverage previous transactions in an attempt to estimate private valuations. We characterize the regret regimes of different settings, taking as a baseline the best context-dependent prices in hindsight. First, in the setting in which the learner has two-bit feedback and strong budget balance constraints, we propose an algorithm with $O(\log T)$ regret. Then, we study the same set-up with noisy valuations, providing a tight $\widetilde O(T^{\frac23})$ regret upper bound. Finally, we show that loosening budget balance constraints allows the learner to operate under more restrictive feedback. Specifically, we show how to address the one-bit, global budget balance setting through a reduction from the two-bit, strong budget balance setup. This established a fundamental trade-off between the quality of the feedback and the strictness of the budget constraints. △ Less

Submitted 28 May, 2024; originally announced May 2024.

Supervised Contamination Detection, with Flow Cytometry Application

Authors: Solenne Gaucher, Gilles Blanchard, Frédéric Chazal

Abstract: The contamination detection problem aims to determine whether a set of observations has been contaminated, i.e. whether it contains points drawn from a distribution different from the reference distribution. Here, we consider a supervised problem, where labeled samples drawn from both the reference distribution and the contamination distribution are available at training time. This problem is moti… ▽ More The contamination detection problem aims to determine whether a set of observations has been contaminated, i.e. whether it contains points drawn from a distribution different from the reference distribution. Here, we consider a supervised problem, where labeled samples drawn from both the reference distribution and the contamination distribution are available at training time. This problem is motivated by the detection of rare cells in flow cytometry. Compared to novelty detection problems or two-sample testing, where only samples from the reference distribution are available, the challenge lies in efficiently leveraging the observations from the contamination detection to design more powerful tests. In this article, we introduce a test for the supervised contamination detection problem. We provide non-asymptotic guarantees on its Type I error, and characterize its detection rate. The test relies on estimating reference and contamination densities using histograms, and its power depends strongly on the choice of the corresponding partition. We present an algorithm for judiciously choosing the partition that results in a powerful test. Simulations illustrate the good empirical performances of our partition selection algorithm and the efficiency of our test. Finally, we showcase our method and apply it to a real flow cytometry dataset. △ Less

Submitted 9 April, 2024; originally announced April 2024.

Fair learning with Wasserstein barycenters for non-decomposable performance measures

Authors: Solenne Gaucher, Nicolas Schreuder, Evgenii Chzhen

Abstract: This work provides several fundamental characterizations of the optimal classification function under the demographic parity constraint. In the awareness framework, akin to the classical unconstrained classification case, we show that maximizing accuracy under this fairness constraint is equivalent to solving a corresponding regression problem followed by thresholding at level $1/2$. We extend thi… ▽ More This work provides several fundamental characterizations of the optimal classification function under the demographic parity constraint. In the awareness framework, akin to the classical unconstrained classification case, we show that maximizing accuracy under this fairness constraint is equivalent to solving a corresponding regression problem followed by thresholding at level $1/2$. We extend this result to linear-fractional classification measures (e.g., ${\rm F}$-score, AM measure, balanced accuracy, etc.), highlighting the fundamental role played by the regression problem in this framework. Our results leverage recently developed connection between the demographic parity constraint and the multi-marginal optimal transport formulation. Informally, our result shows that the transition between the unconstrained problems and the fair one is achieved by replacing the conditional expectation of the label by the solution of the fair regression problem. Finally, leveraging our analysis, we demonstrate an equivalence between the awareness and the unawareness setups in the case of two sensitive groups. △ Less

Submitted 1 September, 2022; originally announced September 2022.

The price of unfairness in linear bandits with biased feedback

Authors: Solenne Gaucher, Alexandra Carpentier, Christophe Giraud

Abstract: In this paper, we study the problem of fair sequential decision making with biased linear bandit feedback. At each round, a player selects an action described by a covariate and by a sensitive attribute. The perceived reward is a linear combination of the covariates of the chosen action, but the player only observes a biased evaluation of this reward, depending on the sensitive attribute. To chara… ▽ More In this paper, we study the problem of fair sequential decision making with biased linear bandit feedback. At each round, a player selects an action described by a covariate and by a sensitive attribute. The perceived reward is a linear combination of the covariates of the chosen action, but the player only observes a biased evaluation of this reward, depending on the sensitive attribute. To characterize the difficulty of this problem, we design a phased elimination algorithm that corrects the unfair evaluations, and establish upper bounds on its regret. We show that the worst-case regret is smaller than $\mathcal{O}(κ_*^{1/3}\log(T)^{1/3}T^{2/3})$, where $κ_*$ is an explicit geometrical constant characterizing the difficulty of bias estimation. We prove lower bounds on the worst-case regret for some sets of actions showing that this rate is tight up to a possible sub-logarithmic factor. We also derive gap-dependent upper bounds on the regret, and matching lower bounds for some problem instance.Interestingly, these results reveal a transition between a regime where the problem is as difficult as its unbiased counterpart, and a regime where it can be much harder. △ Less

Submitted 3 June, 2022; v1 submitted 18 March, 2022; originally announced March 2022.

Hierarchical transfer learning with applications for electricity load forecasting

Authors: Anestis Antoniadis, Solenne Gaucher, Yannig Goude

Abstract: The recent abundance of data on electricity consumption at different scales opens new challenges and highlights the need for new techniques to leverage information present at finer scales in order to improve forecasts at wider scales. In this work, we take advantage of the similarity between this hierarchical prediction problem and multi-scale transfer learning. We develop two methods for hierarch… ▽ More The recent abundance of data on electricity consumption at different scales opens new challenges and highlights the need for new techniques to leverage information present at finer scales in order to improve forecasts at wider scales. In this work, we take advantage of the similarity between this hierarchical prediction problem and multi-scale transfer learning. We develop two methods for hierarchical transfer learning, based respectively on the stacking of generalized additive models and random forests, and on the use of aggregation of experts. We apply these methods to two problems of electricity load forecasting at national scale, using smart meter data in the first case, and regional data in the second case. For these two usecases, we compare the performances of our methods to that of benchmark algorithms, and we investigate their behaviour using variable importance analysis. Our results demonstrate the interest of both methods, which lead to a significant improvement of the predictions. △ Less

Submitted 22 November, 2022; v1 submitted 16 November, 2021; originally announced November 2021.

Optimality of variational inference for stochastic block model with missing links

Authors: Solenne Gaucher, Olga Klopp

Abstract: Variational methods are extremely popular in the analysis of network data. Statistical guarantees obtained for these methods typically provide asymptotic normality for the problem of estimation of global model parameters under the stochastic block model. In the present work, we consider the case of networks with missing links that is important in application and show that the variational approxima… ▽ More Variational methods are extremely popular in the analysis of network data. Statistical guarantees obtained for these methods typically provide asymptotic normality for the problem of estimation of global model parameters under the stochastic block model. In the present work, we consider the case of networks with missing links that is important in application and show that the variational approximation to the maximum likelihood estimator converges at the minimax rate. This provides the first minimax optimal and tractable estimator for the problem of parameter estimation for the stochastic block model with missing links. We complement our results with numerical studies of simulated and real networks, which confirm the advantages of this estimator over current methods. △ Less

Submitted 5 November, 2021; originally announced November 2021.

Finite Continuum-Armed Bandits

Authors: Solenne Gaucher

Abstract: We consider a situation where an agent has $T$ ressources to be allocated to a larger number $N$ of actions. Each action can be completed at most once and results in a stochastic reward with unknown mean. The goal of the agent is to maximize her cumulative reward. Non trivial strategies are possible when side information on the actions is available, for example in the form of covariates. Focusing… ▽ More We consider a situation where an agent has $T$ ressources to be allocated to a larger number $N$ of actions. Each action can be completed at most once and results in a stochastic reward with unknown mean. The goal of the agent is to maximize her cumulative reward. Non trivial strategies are possible when side information on the actions is available, for example in the form of covariates. Focusing on a nonparametric setting, where the mean reward is an unknown function of a one-dimensional covariate, we propose an optimal strategy for this problem. Under natural assumptions on the reward function, we prove that the optimal regret scales as $O(T^{1/3})$ up to poly-logarithmic factors when the budget $T$ is proportional to the number of actions $N$. When $T$ becomes small compared to $N$, a smooth transition occurs. When the ratio $T/N$ decreases from a constant to $N^{-1/3}$, the regret increases progressively up to the $O(T^{1/2})$ rate encountered in continuum-armed bandits. △ Less

Submitted 3 November, 2020; v1 submitted 23 October, 2020; originally announced October 2020.

Electronic and magnetic properties of $α$-FeGe$_2$ films embedded in vertical spin valve devices

Authors: Dietmar Czubak, Samuel Gaucher, Lars Oppermann, Jens Herfort, Klaus Zollner, Jaroslav Fabian, Holger T. Grahn, Manfred Ramsteiner

Abstract: We studied metastable $α$-FeGe$_2$, a novel layered tetragonal material, embedded as a spacer layer in spin valve structures with ferromagnetic Fe$_3$Si and Co$_2$FeSi electrodes. For both types of electrodes, spin valve operation is demonstrated with a metallic transport behavior of the $α$-FeGe$_2$ spacer layer. The spin valve signals are found to increase both with temperature and spacer thickn… ▽ More We studied metastable $α$-FeGe$_2$, a novel layered tetragonal material, embedded as a spacer layer in spin valve structures with ferromagnetic Fe$_3$Si and Co$_2$FeSi electrodes. For both types of electrodes, spin valve operation is demonstrated with a metallic transport behavior of the $α$-FeGe$_2$ spacer layer. The spin valve signals are found to increase both with temperature and spacer thickness, which is discussed in terms of a decreasing magnetic coupling strength between the ferromagnetic bottom and top electrodes. The temperature-dependent resistances of the spin valve structures exhibit characteristic features, which are explained by ferromagnetic phase transitions between 55 and 110~K. The metallic transport characteristics as well as the low-temperature ferromagnetism are found to be consistent with the results of first-principles calculations. △ Less

Submitted 27 October, 2020; v1 submitted 9 October, 2020; originally announced October 2020.

Comments: 11 pages, 13 figures

Outliers Detection in Networks with Missing Links

Authors: Solenne Gaucher, Olga Klopp, Geneviève Robin

Abstract: Outliers arise in networks due to different reasons such as fraudulent behavior of malicious users or default in measurement instruments and can significantly impair network analyses. In addition, real-life networks are likely to be incompletely observed, with missing links due to individual non-response or machine failures. Identifying outliers in the presence of missing links is therefore a cruc… ▽ More Outliers arise in networks due to different reasons such as fraudulent behavior of malicious users or default in measurement instruments and can significantly impair network analyses. In addition, real-life networks are likely to be incompletely observed, with missing links due to individual non-response or machine failures. Identifying outliers in the presence of missing links is therefore a crucial problem in network analysis. In this work, we introduce a new algorithm to detect outliers in a network that simultaneously predicts the missing links. The proposed method is statistically sound: we prove that, under fairly general assumptions, our algorithm exactly detects the outliers, and achieves the best known error for the prediction of missing links with polynomial computation cost. It is also computationally efficient: we prove sub-linear convergence of our algorithm. We provide a simulation study which demonstrates the good behavior of the algorithm in terms of outliers detection and prediction of the missing links. We also illustrate the method with an application in epidemiology, and with the analysis of a political Twitter network. The method is freely available as an R package on the Comprehensive R Archive Network. △ Less

Submitted 1 December, 2020; v1 submitted 29 November, 2019; originally announced November 2019.

Ordered structure of FeGe$_2$ formed during solid-phase epitaxy

Authors: B Jenichen, M Hanke, S Gaucher, A Trampert, J Herfort, H Kirmse, B Haas, E Willinger, X Huang, S C Erwin

Abstract: Fe$_{3}$Si/Ge(Fe,Si)/Fe$_{3}$Si thin film stacks were grown by a combination of molecular beam epitaxy and solid phase epitaxy (Ge on Fe$_{3}$Si). The stacks were analyzed using electron microscopy, electron diffraction, and synchrotron X-ray diffraction. The Ge(Fe,Si) films crystallize in the well oriented, layered tetragonal structure FeGe$_{2}$ with space group P4mm. This kind of structure does… ▽ More Fe$_{3}$Si/Ge(Fe,Si)/Fe$_{3}$Si thin film stacks were grown by a combination of molecular beam epitaxy and solid phase epitaxy (Ge on Fe$_{3}$Si). The stacks were analyzed using electron microscopy, electron diffraction, and synchrotron X-ray diffraction. The Ge(Fe,Si) films crystallize in the well oriented, layered tetragonal structure FeGe$_{2}$ with space group P4mm. This kind of structure does not exist as a bulk material and is stabilized by solid phase epitaxy of Ge on Fe$_{3}$Si. We interpret this as an ordering phenomenon induced by minimization of the elastic energy of the epitaxial film. △ Less

Submitted 11 July, 2019; originally announced July 2019.

Journal ref: Physical Review Materials 2, 051402(R) (2018)

Maximum Likelihood Estimation of Sparse Networks with Missing Observations

Authors: Solenne Gaucher, Olga Klopp

Abstract: Estimating the matrix of connections probabilities is one of the key questions when studying sparse networks. In this work, we consider networks generated under the sparse graphon model and the in-homogeneous random graph model with missing observations. Using the Stochastic Block Model as a parametric proxy, we bound the risk of the maximum likelihood estimator of network connections probabilitie… ▽ More Estimating the matrix of connections probabilities is one of the key questions when studying sparse networks. In this work, we consider networks generated under the sparse graphon model and the in-homogeneous random graph model with missing observations. Using the Stochastic Block Model as a parametric proxy, we bound the risk of the maximum likelihood estimator of network connections probabilities , and show that it is minimax optimal. When risk is measured in Frobenius norm, no estimator running in polynomial time has been shown to attain the minimax optimal rate of convergence for this problem. Thus, maximum likelihood estimation is of particular interest as computationally efficient approximations to it have been proposed in the literature and are often used in practice. △ Less

Submitted 27 April, 2021; v1 submitted 27 February, 2019; originally announced February 2019.

Comments: We derive a variational approximation to the maximum likelihood estimator of the connection probabilities. We bound the risk of this tractable estimator

Ferromagnet/Semiconductor/Ferromagnet Hybrid Trilayers grown using solid phase epitaxy

Authors: Samuel Gaucher, Bernd Jenichen, Jens Herfort

Abstract: The direct growth of semiconductors over metals by molecular beam epitaxy is a difficult task due to the large differences in crystallization energy between these types of materials. This aspect is problematic in the context of spintronics, where coherent spin-injection must proceed via ballistic transport through sharp interfacial Schottky barriers. We report the realization of single-crystalline… ▽ More The direct growth of semiconductors over metals by molecular beam epitaxy is a difficult task due to the large differences in crystallization energy between these types of materials. This aspect is problematic in the context of spintronics, where coherent spin-injection must proceed via ballistic transport through sharp interfacial Schottky barriers. We report the realization of single-crystalline ferromagnet/semiconductor/ferromagnet hybrid trilayers using solid-phase epitaxy, with combinations of Fe3Si, Co2FeSi, and Ge. The slow annealing of amorphous Ge over Fe3Si results in a crystalline filmlm identified as FeGe2. When the annealing is performed over Co2FeSi, reflected high-energy electron diffraction and X-ray diffraction indicate the creation of a different crystalline Ge(Co,Fe,Si) compound, which also preserves growth orientation. It was possible to observe independent magnetization switching of the ferromagnetic layers in a Fe3Si/FeGe2/Co2FeSi sample, thanks to the different coercive fields of the two metals and to the quality of the interfaces. This result is a step towards the implementation of vertical spin-selective transistor-like devices. △ Less

Submitted 7 June, 2018; originally announced June 2018.

Comments: 4 pages

Specific heat and entropy of fractional quantum Hall states in the second Landau level

Authors: B. A. Schmidt, K. Bennaceur, S. Gaucher, G. Gervais, L. N. Pfeiffer, K. W. West

Abstract: Specific heat has had an important role in the study of superfluidity and superconductivity, and could provide important information about the fractional quantum Hall effect as well. However, traditional measurements of the specific heat of a two-dimensional electron gas are difficult due to the large background contribution of the phonon bath, even at very low temperatures. Here, we report measur… ▽ More Specific heat has had an important role in the study of superfluidity and superconductivity, and could provide important information about the fractional quantum Hall effect as well. However, traditional measurements of the specific heat of a two-dimensional electron gas are difficult due to the large background contribution of the phonon bath, even at very low temperatures. Here, we report measurements of the specific heat per electron in the second Landau level by measuring the thermalization time between the electrons and phonons. We observe activated behaviour of the specific heat of the 5/2 and 7/3 fractional quantum Hall states, and extract the entropy by integrating over temperature. Our results are in excellent agreement with previous measurements of the entropy via longitudinal thermopower. Extending the technique to lower temperatures could lead to detection of the non-Abelian entropy predicted for bulk quasiparticles at 5/2 filling △ Less

Submitted 7 June, 2017; v1 submitted 8 May, 2016; originally announced May 2016.

Journal ref: Phys. Rev. B 95, 201306 (2017)

Mechanical Flip-Chip for Ultra-High Electron Mobility Devices

Authors: K. Bennaceur, B. A. Schmidt, S. Gaucher, D. Laroche, M. P. Lilly, J. L. Reno, K. W. West, L. N. Pfeiffer, G. Gervais

Abstract: Electrostatic gates are of paramount importance for the physics of devices based on high-mobility two-dimensional electron gas (2DEG) since they allow depletion of electrons in selected areas. This field-effect gating enables the fabrication of a wide range of devices such as, for example, quantum point contacts (QPC), electron interferometers and quantum dots. To fabricate these gates, processing… ▽ More Electrostatic gates are of paramount importance for the physics of devices based on high-mobility two-dimensional electron gas (2DEG) since they allow depletion of electrons in selected areas. This field-effect gating enables the fabrication of a wide range of devices such as, for example, quantum point contacts (QPC), electron interferometers and quantum dots. To fabricate these gates, processing is usually performed on the 2DEG material, which is in many cases detrimental to its electron mobility. Here we propose an alternative process which does not require any processing of the 2DEG material other than for the ohmic contacts. This approach relies on processing a separate wafer that is then mechanically mounted on the 2DEG material in a flip-chip fashion. This technique proved successful to fabricate quantum point contacts on both GaAs/AlGaAs materials with both moderate and ultra-high electron mobility. △ Less

Submitted 17 March, 2015; originally announced March 2015.

Comments: 5 pages, 3 figures

Journal ref: Scientific Reports 5, Article number: 13494 (2015)