Publications

My goal is to make humanoid robots walk, in many places.

I'm interested in 3D bipedal walking, that is to say, walking over terrains where foot contacts between the robot and the ground can be sketchy. In this general case, simplifications previously used for walking on flat floors disappear, and we are left with a number of exciting questions and solutions to develop. My understanding of these questions is summed up in the following presentation:

Below is my list of publications, organized by topics. Pre-prints correspond to ongoing works. Thankfully the robotics community is making the move to open archives, publishing pre-prints of recently-submitted works on repositories like HAL or arXiv. I hope we can take it to the next step by curating our own epijournals as well.

Pre-prints

  • Capturability-based Pattern Generation for Walking with Variable Height
    Stéphane Caron, Adrien Escande, Leonardo Lanari and Bastien Mallein. Submitted January 2018. Revised October 2018. (pdf)

    Capturability analysis of the linear inverted pendulum (LIP) model enabled walking with constrained height based on the capture point. We generalize this analysis to the variable-height inverted pendulum (VHIP) and show how it enables 3D walking over uneven terrains based on capture inputs. Thanks to a tailored optimization scheme, we can compute these inputs fast enough for real-time model predictive control. We implement this approach as open-source software and demonstrate it in dynamic simulations.

  • Stair Climbing Stabilization of the HRP-4 Humanoid Robot using Whole-body Admittance Control
    Stéphane Caron, Abderrahmane Kheddar and Olivier Tempier. Submitted September 2018. (pdf)

    This paper considers dynamic stair climbing with the HRP-4 humanoid robot as part of an Airbus manufacturing use-case demonstrator. Apart from the fact that HRP-4 had never been challenged to this scenario before, we share experimental knowledge gathered in achieving this task. We also show how walking stabilization based on linear inverted pendulum tracking can be extended with quadratic programming-based wrench distribution and a whole-body admittance controller that relies on both end-effector and CoM strategies. While existing stabilizers tend to use either one or the other of these two approaches, experiments suggest that their combination can improve tracking performance. We demonstrate this solution in a hardware experiment where HRP-4 climbs a staircase with 18.5-cm steps, and release our walking controller as open source software.

Walking stabilization

  • Stair Climbing Stabilization of the HRP-4 Humanoid Robot using Whole-body Admittance Control
    Stéphane Caron, Abderrahmane Kheddar and Olivier Tempier. Submitted September 2018. (pdf)

    This paper considers dynamic stair climbing with the HRP-4 humanoid robot as part of an Airbus manufacturing use-case demonstrator. Apart from the fact that HRP-4 had never been challenged to this scenario before, we share experimental knowledge gathered in achieving this task. We also show how walking stabilization based on linear inverted pendulum tracking can be extended with quadratic programming-based wrench distribution and a whole-body admittance controller that relies on both end-effector and CoM strategies. While existing stabilizers tend to use either one or the other of these two approaches, experiments suggest that their combination can improve tracking performance. We demonstrate this solution in a hardware experiment where HRP-4 climbs a staircase with 18.5-cm steps, and release our walking controller as open source software.

  • Walking on Gravel with Soft Soles using Linear Inverted Pendulum Tracking and Reaction Force Distribution
    Adrien Pajon, Stéphane Caron, Giovanni De Magistris, Sylvain Miossec and Abderrahmane Kheddar. Humanoids 2017, Birmingham, United Kingdom, November 2017. (pdf)

    Soft soles absorb impacts and cast ground un-evenness during locomotion on rough terrains. However, they introduce passive degrees of freedom (deformations under the feet) that complexify the tasks of state estimation and overall robot stabilization. We address this problem by developing a control loop that stabilizes humanoid robots when walking with soft soles on flat and uneven terrain. Our closed-loop controller minimizes the errors on the center of mass (COM) and the zero moment point (ZMP) with an admittance control of the feet based on a simple deformation estimator. We demonstrate its effectiveness in real experiments on the HRP-4 humanoid.

Walking pattern generation

  • Balance control using both ZMP and COM height variations: A convex boundedness approach
    Stéphane Caron and Bastien Mallein. ICRA 2018, Brisbane, Australia, May 2018. (pdf)

    Developments for 3D control of the center of mass (CoM) of biped robots are currently located in two local minima: on the one hand, methods that allow CoM height variations but only work in the 2D sagittal plane; on the other hand, nonconvex direct transcriptions of centroidal dynamics that are delicate to handle. This paper presents an alternative that controls the CoM in 3D via an indirect transcription that is both low-dimensional and solvable fast enough for real-time control. The key to this development is the notion of boundedness condition, which quantifies the capturability of 3D CoM trajectories.

  • When to make a step? Tackling the timing problem in multi-contact locomotion by TOPP-MPC
    Stéphane Caron and Quang-Cuong Pham. Humanoids 2017, Birmingham, United Kingdom, November 2017. (pdf)

    We present a model predictive controller (MPC) for multi-contact locomotion where predictive optimizations are realized by time-optimal path parameterization (TOPP). A key feature of this solution is that, contrary to existing planners where step timings are provided as inputs, here the timing between contact switches is computed as output of a fast nonlinear optimization. This is appealing to multi-contact locomotion, where proper timings depend on terrain topology and suitable heuristics are unknown. We show how to formulate legged locomotion as a TOPP problem and demonstrate the behavior of the resulting TOPP-MPC controller in simulations with a model of the HRP-4 humanoid robot.

  • Dynamic Walking over Rough Terrains by Nonlinear Predictive Control of the Floating-base Inverted Pendulum
    Stéphane Caron and Abderrahmane Kheddar. IROS 2017, Vancouver, Canada, September 2017. (pdf)

    We present a real-time rough-terrain dynamic walking pattern generator. Our method automatically finds step durations, which is a critical issue over rough terrains where they depend on terrain topology. To achieve this level of generality, we introduce the Floating-base Inverted Pendulum (FIP) model where the center of mass can translate freely and the zero-tilting moment point is allowed to leave the contact surface. We show that this model is equivalent to the linear-inverted pendulum mode with variable center of mass height, aside from the fact that its equations of motion remain linear. Our design then follows three steps: (i) we characterize the FIP contact-stability condition; (ii) we compute feedforward controls by solving a nonlinear optimization over receding-horizon FIP trajectories. Despite running at 30 Hz in a model-predictive fashion ...

  • Multi-contact Walking Pattern Generation based on Model Preview Control of 3D COM Accelerations
    Stéphane Caron and Abderrahmane Kheddar. Humanoids 2016, Cancún, Mexico, November 2016. (pdf)

    We present a multi-contact walking pattern generator based on preview-control of the 3D acceleration of the center of mass (COM). A key point in the design of our algorithm is the calculation of contact-stability constraints. Thanks to a mathematical observation on the algebraic nature of the frictional wrench cone, we show that the 3D volume of feasible COM accelerations is a always an upward-pointing cone. We reduce its computation to a convex hull of (dual) 2D points, for which optimal \({\cal O}(n \log n)\) algorithms are readily available. This reformulation brings a significant speedup compared to previous methods, which allows us to compute time-varying contact-stability criteria fast enough for the control loop. Next, we propose a conservative trajectory-wide contact-stability criterion, which can be derived from COM-acceleration volumes ...

Contact stability conditions

  • ZMP Support Areas for Multi-contact Mobility Under Frictional Constraints
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. IEEE Transactions on Robotics. Submitted October 2015. Published December 2016. (pdf)

    We propose a method for checking and enforcing multi-contact stability based on the Zero-tilting Moment Point (ZMP). The key to our development is the generalization of ZMP support areas to take into account (a) frictional constraints and (b) multiple non-coplanar contacts. We introduce and investigate two kinds of ZMP support areas. First, we characterize and provide a fast geometric construction for the support area generated by valid contact forces, with no other constraint on the robot motion. We call this set the full support area. Next, we consider the control of humanoid robots using the Linear Pendulum Mode (LPM). We observe that the constraints stemming from the LPM induce a shrinking of the support area, even for walking on horizontal floors. We ...

  • Leveraging Cone Double Description for Multi-contact Stability of Humanoids with Applications to Statics and Dynamics
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. RSS 2015, Roma, Italy, July 2015. (pdf)

    We build on previous works advocating the use of the Gravito-Inertial Wrench Cone (GIWC) as a general contact stability criterion (a "ZMP for non-coplanar contacts"). We show how to compute this wrench cone from the friction cones of contact forces by using an intermediate representation, the surface contact wrench cone, which is the minimal representation of contact stability for each surface contact. The observation that the GIWC needs to be computed only once per stance leads to particularly efficient algorithms, as we illustrate in two important problems for humanoids : "testing robust static equilibrium" and "time-optimal path parameterization". We show, through theoretical analysis and in physical simulations, that our method is more general and/or outperforms existing ones.

  • Stability of Surface Contacts for Humanoid Robots Closed-Form Formulae of the Contact Wrench Cone for Rectangular Support Areas
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. ICRA 2015, Seattle, USA, May 2015. (pdf)

    Humanoids locomote by making and breaking contacts with their environment. Thus, a crucial question for them is to anticipate whether a contact will hold or break under effort. For rigid surface contacts, existing methods usually consider several point-contact forces, which has some drawbacks due to the underlying redundancy. We derive a criterion, the Contact Wrench Cone (CWC), which is equivalent to any number of applied forces on the contact surface, and for which we provide a closed-form formula.

Multi-contact motion control

  • Post-Impact Adaptive Compliance for Humanoid Falls Using Predictive Control of a Reduced Model
    Vincent Samy, Stéphane Caron, Karim Bouyarmane and Abderrahmane Kheddar. Humanoids 2017, Birmingham, United Kingdom, November 2017. (pdf)

    We consider control of a humanoid robot in active compliance just after the impact consecutive to a fall. The goal of this post-impact braking is to absorb undesired linear momentum accumulated during the fall, using a limited supply of time and actuation power. The gist of our method is an optimal distribution of undesired momentum between the robot's hand and foot contact points, followed by the parallel resolution of Linear Model Predictive Control (LMPC) at each contact. This distribution is made possible thanks to emph{torque-limited friction polytopes}, an extension of friction cones that takes actuation limits into account. Individual LMPC results are finally combined back into a feasible CoM trajectory sent to the robot's whole-body controller. We validate the solution ...

  • Multi-Contact Motion Planning and Control
    Karim Bouyarmane, Stéphane Caron, Adrien Escande and Abderrahmane Kheddar. Humanoid Robotics: a Reference. Edited by Ambarish Goswami and Prahlad Vadakkepat. Springer. (pdf)

    The essence of humanoid robots is their ability to reproduce human skills in locomotion and manipulation. Early efforts in humanoid research were dedicated to bipedal walking, first on flat terrains and recently on uneven ones, while the manipulation capabilities inherit from the literature in bimanual and dexterous-hand manipulation. In practice, the two problems interact largely. Locomotion in cluttered spaces benefits from extra contacts between any part of the robot and the environment, such as when grippers grasp a handrail during stair climbing, while legs can conversely enhance manipulation capabilities, such as when arching the whole-body to augment contact pressure at an end-effector. The two problems share the same background: they are governed by non-smooth dynamics (friction and impacts ...

  • Supervoxel Plane Segmentation and Multi-Contact Motion Generation for Humanoid Stair Climbing
    Tianwei Zhang, Stéphane Caron and Yoshihiko Nakamura. International Journal of Humanoid Robotics. Submitted August 2016. Published March 2017. (pdf)

    Stair climbing is still a challenging task for humanoid robots, especially in unknown environments. In this paper, we address this problem from perception to execution. Our first contribution is a real-time plane segment estimation method using unorganized lidar data without prior models of the staircase. We then integrate this solution with humanoid motion planning. Our second contribution is a stair-climbing motion generator where estimated plane segments are used to compute footholds and stability polygons. We evaluate our method on various staircases. We also demonstrate the feasibility of the generated trajectories in a real-life experiment with the humanoid robot HRP-4.

  • Teleoperation System Design of Valve Turning Motions in Degraded Communication Conditions
    Stéphane Caron and Yoshihiko Nakamura. RSJ 2015, Tokyo, Japan, September 2015. (pdf)

    During the DARPA Robotics Challenge (DRC), robots were expected to solve a number of tasks under teleoperation by a human operator. Limits in execution time and teleoperation bandwidth required teams to implement some level of autonomy on their robots, yet meaningful input could still be provided by the operator on a regular basis via the team’s Operator Control System (OCS). The purpose of the present paper is to report on the development of Team Hydra’s OCS for the DRC in the context of the valve-turning task. We describe the design of the system as well as the technical choices made, meanwhile pointing out the underlying research questions and directions for future work.

Force sensing

  • Multi-Contact Interaction Force Sensing from Whole-Body Motion Capture
    Tu-Hoa Pham, Stéphane Caron and Abderrahmane Kheddar. IEEE Transactions on Industrial Informatics. Submitted November 2016. Published October 2017. (pdf)

    We present a novel technique that unobtrusively estimates forces exerted by human participants in multi-contact interaction with rigid environments. Our method uses motion capture only, thus circumventing the need to setup cumbersome force transducers at all potential contacts between the human body and the environment. This problem is particularly challenging, as the knowledge of a given motion only characterizes the resultant force, which can generally be caused by an infinity of force distributions over individual contacts. We collect and release a large-scale dataset on how humans instinctively regulate interaction forces on diverse multi-contact tasks and motions. The force estimation framework we propose leverages physics-based optimization and neural networks to reconstruct force distributions that are physically realistic and compatible with real interaction ...

  • Whole-Body Contact Force Sensing From Motion Capture
    Tu-Hoa Pham, Adrien Bufort, Stéphane Caron and Abderrahmane Kheddar. SII 2016, Sapporo, Japan, December 2016. Best Paper Award. (pdf)

    In this paper, we challenge the estimation of contact forces backed with ground-truth sensing in human whole-body interaction with the environment, from motion capture only. Our novel method makes it possible to get rid of cumbersome force sensors in monitoring multi-contact motion together with force data. This problem is very challenging. Indeed, while a given force distribution uniquely determines the resulting kinematics, the converse is generally not true in multi-contact. In such scenarios, physics-based optimization alone may only capture force distributions that are physically compatible with a given motion rather than the actual forces being applied. We address this indeterminacy by collecting a large-scale dataset on whole-body motion and contact forces humans apply in multi-contact scenarios. We then train recurrent neural ...

Motion planning

  • Completeness of Randomized Kinodynamic Planners with State-based Steering
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. Robotics and Autonomous Systems. Submitted November 2015. Published December 2016. (pdf)

    Probabilistic completeness is an important property in motion planning. Although it has been established with clear assumptions for geometric planners, the panorama of completeness results for kinodynamic planners is still incomplete, as most existing proofs rely on strong assumptions that are difficult, if not impossible, to verify on practical systems. In this paper, we focus on an important class of kinodynamic planners, namely those that interpolate trajectories in the state space. We provide a proof of probabilistic completeness for these planners under assumptions that can be readily verified from the system’s equations of motion and the user-defined interpolation function. Our proof relies crucially on a property of interpolated trajectories, termed second-order continuity (SOC), which we show is tightly related to the ...

  • Admissible Velocity Propagation: Beyond Quasi-Static Path Planning for High-Dimensional Robots
    Quang-Cuong Pham, Stéphane Caron, Puttichai Lertkultanon and Yoshihiko Nakamura. International Journal of Robotics Research. Submitted November 2014. Published November 2016. (pdf)

    Path-velocity decomposition is an intuitive yet powerful approach to address the complexity of kinodynamic motion planning. The difficult trajectory planning problem is solved in two separate adn simpler steps: first, find a path in the configuration space that satisfies the geometric constraints (path planning), and second, find a time-parameterization of that path satisfying the kinodynamic constraints. A fundamental requirement is that the path found in the first step should be time-parameterizable. Most existing works fulfill this requirement by enforcing quasi-static constraints in the path planning step, resulting in an important loss in completeness. We propose a method that enables path-velocity decomposition to discover truly dynamic motions, i.e. motions that are not quasi-statically executable. At the heart of the ...

  • Planning with the Center-of-Mass rather than Stances for Humanoids Walking on Uneven Terrains
    Stéphane Caron and Yoshihiko Nakamura. IFToMM 2015, Taipei, Taiwan, October 2015. (pdf)

    In the current literature for non-gaited humanoid motion planning, stances (i.e., contact locations) are usually planned in a first step, after which joint-angle trajectories are interpolated or planned themselves. In this paper, we propose an alternative where planning is driven by center-of-mass motions rather than stances. Our approach uses a randomized motion planner as its first layer to explore the space of horizontal CoM coordinates. At a lower level, we propose a custom method to extend stances based on a desired CoM position. We evaluate the ability of the resulting planner in a rubble-field 3D environment with a model of the HYDRA humanoid robot.

  • Kinodynamic Motion Retiming for Humanoid Robots
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. RSJ 2014, Fukuoka, Japan, September 2014. (pdf)

    In this paper, we advocate the use of Time-Optimal Path Parameterization (TOPP) to enable planning of dynamic motions for humanoid robots. We extend the existing formulation of ZMP constraints to arbitrary polygonal areas and provide an original approach to incorporate frictional contact constraints in TOPP. We evaluate our algorithm experimentally with the HRP-4 robot performing a stepping motion. Given a slow and quasi-static input motion, our method automatically produces a 2x-faster dynamic motion successfully executed on the real robot (4x faster in simulation).

  • Completeness of Randomized Kinodynamic Planners with State-based Steering
    Stéphane Caron, Quang-Cuong Pham and Yoshihiko Nakamura. ICRA 2014, Hong-Kong, China, June 2014. (pdf)

    The panorama of probabilistic completeness results for kinodynamic planners is still confusing. Most existing completeness proofs require strong assumptions that are difficult, if not impossible, to verify in practice. To make completeness results more useful, it is thus sensible to estabish a classification of the various types of constraints and planning methods, and then attack each class with specific proofs and hypotheses that can be verified in practice. We propose such a classification, and provide a proof of probabilistic completeness for an important class of planners, namely those whose steering method is based on the interpolation of system trajectories in the state space. We also provide design guidelines for the interpolation function and discuss two criteria arising from our analysis: local boundedness and acceleration compliance.

  • Kinodynamic Motion Planners based on Velocity Interval Propagation
    Stéphane Caron, Yoshihiko Nakamura and Quang-Cuong Pham. RSJ 2013, Tokyo, Japan, September 2013. (pdf)

    Humanoid robotics has spawned several fields of active research. When it comes to dynamic motion control, three lines of work stand out: reduced models (combined with inverse kinematics), local controllers and global planning. In the present paper, we present and further develop a motion planning approach recently proposed in [10], which allows planning with dynamics constraints while staying in the configuration space; thus avoiding the complexity explosion mentioned above. We briefly discuss how this approach can be extended to handle ZMP constraints, which may give rise to a new family of efficient motion planners for humanoid robots.

  • Kinodynamic planning in the configuration space via Admissible Velocity Propagation
    Quang-Cuong Pham, Stéphane Caron and Yoshihiko Nakamura. RSS 2013, Berlin, Germany, June 2013. (pdf)

    We propose a method that enables kinodynamic planning in the configuration space (of dimension n) instead of the state space (of dimension 2n), thereby potentially cutting down the complexity of usual kinodynamic planning algorithms by an exponential factor. At the heart of this method is a new technique – called Velocity Interval Propagation – which, given a path in the configuration space and an interval of reachable velocities at the beginning of that path, computes exactly and efficiently the interval of all the velocities the system can reach after traversing the path while respecting the system kinodynamic constraints. Combining this technique with usual sampling-based methods gives rise to a family of new motion planners that can appropriately handle kinodynamic constraints while avoiding the complexity explosion and, to some ...

PhD thesis

Non-robotics research

Multi-armed bandits
  • Mixing bandits A recipe for improved cold-start recommendations in a social network
    Stéphane Caron and Smriti Bhagat. SNAKDD 2013. (pdf)

    Recommending items to new or “cold-start” users is a challenging problem for recommender systems. Collaborative filtering approaches fail when the preference history of users is not available. A promising direction that has been explored recently [12] is to utilize the information in the social networks of users to improve the quality of cold-start recommendations. That is, given that users are part of a social network, a new user shows up in the network with no preference history and limited social links, the recommender system tries to learn the user’s tastes as fast as possible. In this work, we model the learning of preferences of cold-start users using multi-armed bandits [5] embedded in a social network. We propose two novel strategies leveraging neighborhood estimates to improve the learning rate of bandits for ...

  • Leveraging Side Observations in Stochastic Bandits
    Stéphane Caron, Branislav Kveton, Marc Lelarge and Smriti Bhagat. UAI 2012. (pdf)

    This paper considers stochastic bandits with side observations, a model that accounts for both the exploration/exploitation dilemma and relationships between arms. In this setting, after pulling an arm i, the decision maker also observes the rewards for some other actions related to i. We will see that this model is suited to content recommendation in social networks, where users' reactions may be endorsed or not by their friends. We provide efficient algorithms based on upper confidence bounds (UCBs) to leverage this additional information and derive new bounds improving on standard regret guarantees. We also evaluate these policies in the context of movie recommendation in social networks: experiments on real datasets show substantial learning rate speedups ranging from 2.2x to 14x on dense networks.

Smart grid energy management
  • Incentive-based Energy Consumption Scheduling Algorithms for the Smart Grid
    Stéphane Caron and George Kesidis. SmartGridComm 2010. (pdf)

    In this paper, we study Demand Response (DR) problematics for different levels of information sharing in a smart grid. We propose a dynamic pricing scheme incentivizing consumers to achieve an aggregate load profile suitable for utilities, and study how close they can get to an ideal flat profile depending on how much information they share. When customers can share all their load profiles, we provide a distributed algorithm, set up as a cooperative game between consumers, which significantly reduces the total cost and peak-to-average ratio (PAR) of the system. In the absence of full information sharing (for reasons of privacy), when users have only access to the instantaneous total load on the grid, we provide distributed stochastic strategies that successfully exploit this information to improve the overall load profile. Simulation results confirm ...

P2P storage systems
  • P2P Storage Systems: Study of Different Placement Policies
    Stéphane Caron, Frédéric Giroire, Dorian Mazauric, Julian Monteiro and Stéphane Pérennes. Peer-to-Peer Networking and Applications, Springer, March 2013. (pdf)

    In a P2P storage system using erasure codes, a data block is encoded in many redundancy fragments. These fragments are then sent to distinct peers of the network. In this work, we study the impact of different placement policies of these fragments on the performance of storage systems. Several practical factors (easier control, software reuse, latency) tend to favor data placement strategies that preserve some degree of locality. We compare three policies: two of them are local, in which the data are stored in logical neighbors, and the other one, global, in which the data are spread randomly in the whole system. We focus on the study of the probability to lose a data block and the bandwidth consumption to maintain ...

  • Data Life Time for Different Placement Policies in P2P Storage Systems
    Stéphane Caron, Frédéric Giroire, Dorian Mazauric, Julian Monteiro and Stéphane Pérennes. Globe 2010. (pdf)

    Peer-to-peer systems are foreseen as an efficient solution to achieve reliable data storage at low cost. To deal with common P2P problems such as peer failures or churn, such systems encode the user data into redundant fragments and distribute them among peers. The way they distribute it, known as placement policy, has a significant impact on their behavior and reliability. In this paper, we study the impact of different placement policies on the data life time. More precisely, we describe methods to compute and approximate the mean time before the system loses data (Mean Time to Data Loss). We compare this metric for three placement policies: two of them local, in which the data is stored in logical peer neighborhoods, and one of them global in ...

  • P2P Storage Systems Data Life Time for Different Placement Policies
    Stéphane Caron, Frédéric Giroire, Dorian Mazauric, Julian Monteiro and Stéphane Pérennes. AlgoTel 2010. (pdf)

    Peer-to-peer systems are foreseen as an efficient solution to achieve reliable data storage at low cost. To deal with common P2P problems such as peer failures or churn, such systems encode the user data into redundant fragments and distribute them among peers. The way they distribute it, known as placement policy, has a significant impact on their behavior and reliability. In this report, after a brief state-of-the-art of the technology used in P2P storage systems, we compare three different placement policies: two of them local, in which the data is stored in logical peer neighborhoods, and on of them global in which fragments are parted at random among the different peers. For each policy, we give either Markov Chain Models to efficiently compute the Mean Time ...

Patents

I'm not particularly proud of these. Despite stemming from a laudable principle (incenting companies to disclose technological knowledge lest it is lost when the company shuts down), the patent system is presently dead broken, as illustrated e.g. by patent trolls.

  • System and method for recommending items in a social network
    Stéphane Caron, Branislav Kveton, Marc Lelarge and Smriti Bhagat. US Patent Application US-20150199715-A1.

    The present principles consider stochastic bandits with side observations, a model that accounts for both the exploration/exploitation dilemma and relationships between arms. In this setting, after pulling an arm i, the decision maker also observes the rewards for some other actions related to i. The present principles provide a method and a system for efficiently leveraging additional information based on the responses provided by other users connected to the user via a computerized social network and derive new bounds improving on standard regret guarantees. We will see that this model is suited to content recommendation in social networks, where users' reactions may be endorsed or not by their friends.

  • Method for cold start of a multi-armed bandit in a recommender system
    Smriti Bhagat and Stéphane Caron. US Patent Application US-20150012345-A1. (pdf)

    A method performed by a recommender system to recommend items to a new user includes calculating reward estimates from multiple multi-armed bandit models of a user and her social network friends. The new user's social network friends have multi-armed bandit models that are well established. The mixed multi-armed bandit estimates are processed to select the arm that maximizes the estimated reward to the new user. The multi-armed bandit arm of the greatest reward estimate is played and the new user responds by providing feedback so that the new user's multi-armed bandit model is updated as time progresses.

Others
  • 3D hardware canaries
    Sébastien Briais, Stéphane Caron, Jean-Michel Cioranesco, Jean-Luc Danger, Sylvain Guilley, Jacques-Henri Jourdan, Arthur Milchior, David Naccache, and Thibault Porteboeuf. Cryptographic Hardware and Embedded Systems, Springer, 2012. (pdf)

    3D integration is a promising advanced manufacturing process offering a variety of new hardware security protection opportunities. This paper presents a way of securing 3D ICs using Hamiltonian paths as hardware integrity verification sensors. As 3D integration consists in the stacking of many metal layers, one can consider surrounding a security-sensitive circuit part by a wire cage. After exploring and comparing different cage construction strategies (and reporting preliminary implementation results on silicon), we introduce a "hardware canary". The canary is a spatially distributed chain of functions Fi positioned at the vertices of a 3D cage surrounding a protected circuit. A correct answer (Fn ∘ … ∘ F1)(m) to a challenge m attests ...

  • A model of network neutrality with usage-based prices
    Eitan Altman, Pierre Bernhard, Stéphane Caron, George Kesidis, Julio Rojas-Mora and Sulan Wong. Telecommunication Systems, Springer, June 2011. (pdf)

    Hahn and Wallsten (Econ. Voice 3(6):1–7, 2006) wrote that network neutrality “usually means that broadband service providers charge consumers only once for Internet access, do not favor one content provider over another, and do not charge content providers for sending information over broadband lines to end users.” In this paper we study the implications of non-neutral behaviors under a simple model of linear demand-response to usage-based prices. We take into account advertising revenues for the content provider and consider both cooperative and non-cooperative scenarios. In particular, we model the: impact of side-payments between service and content providers, consider an access provider that offers multiple service classes, and model leader-follower (Stackelberg game) dynamics. We finally study the additional option ...

  • Application Neutrality and a Paradox of Side Payments
    Stéphane Caron, George Kesidis and Eitan Altman. ReARCH 2010. (pdf)

    The ongoing debate over net neutrality covers a broad set of issues related to the regulation of public networks. In two ways, we extend an idealized usage-priced game-theoretic framework based on a common linear demand-response model. First, we study the impact of "side payments" among a plurality of Internet service (access) providers and content providers. In the non-monopolistic case, our analysis reveals an interesting "paradox" of side payments in that overall revenues are reduced for those that receive them. Second, assuming different application types (e.g., HTTP web traffic, peer-to-peer file sharing, media streaming, interactive VoIP), we extend this model to accommodate differential pricing among them in order to study the issue of application neutrality. Revenues for neutral and non-neutral pricing are compared for the case of two application types ...

  • A Study of Non-neutral Networks with Usage-Based Prices
    Eitan Altman, Pierre Bernhard, Stéphane Caron, George Kesidis, Julian Rojas-Mora and Sulan Wong. ETM 2010. (pdf)

    Hahn and Wallsten wrote that network neutrality "usually means that broadband service providers charge consumers only once for Internet access, do not favor one content provider over another, and do not charge content providers for sending information over broadband lines to end users." In this paper we study the implications of non-neutral behaviors under a simple model of linear demand-response to usage-based prices. We take into account advertising revenues and consider both cooperative and non-cooperative scenarios. In particular, we model the impact of side-payments between service and content providers. We also consider the effect of service discrimination by access providers, as well as an extension of our model to non-monopolistic content providers.

Pages of this website are under the CC-BY 4.0 license.