In this paper we apply active learning algorithms for dynamic pricing in a prominent e-commerce website. Dynamic pricing involves changing the price of items on a regular basis, and uses the feedback from the pricing decisions to update prices of the items. Most popular approaches to dynamic pricing use a passive learning approach, where the algorithm uses historical data to learn various parameters of the pricing problem, and uses the updated parameters to generate a new set of prices. We show that one can use active learning algorithms such as Thompson sampling to more efficiently learn the underlying parameters in a pricing problem. We apply our algorithms to a real e-commerce system and show that the algorithms indeed improve revenue compared to pricing algorithms that use passive learning.
Jan 11 2018 cs.CV
Recent data-driven approaches to scene interpretation predominantly pose inference as an end-to-end black-box mapping, commonly performed by a Convolutional Neural Network (CNN). However, decades of work on perceptual organization in both human and machine vision suggests that there are often intermediate representations that are intrinsic to an inference task, and which provide essential structure to improve generalization. In this work, we explore an approach for injecting prior domain structure into neural network training by supervising hidden layers of a CNN with intermediate concepts that normally are not observed in practice. We formulate a probabilistic framework which formalizes these notions and predicts improved generalization via this deep supervision method. One advantage of this approach is that we are able to train only from synthetic CAD renderings of cluttered scenes, where concept values can be extracted, but apply the results to real images. Our implementation achieves the state-of-the-art performance of 2D/3D keypoint localization and image classification on real image benchmarks, including KITTI, PASCAL VOC, PASCAL3D+, IKEA, and CIFAR100. We provide additional evidence that our approach outperforms alternative forms of supervision, such as multi-task networks.
Dec 15 2017 cs.SE
With high penetration of distributed renewable energy resources along with sophisticated automation and information technology, cyber-physical energy systems (CPES, i.e. Smart Grids here) requires a holistic approach to evaluate the integration at a system level, addressing all relevant domains. Hybrid cloud SCADA (Supervisory, Control And Data Acquisition), allowing laboratories to be linked in a consistent infrastructure can provide the support for such multi-platform experiments. This paper presents the procedure to implement a CIM (Common Information Model) compliant hybrid cloud SCADA, with database and client adaptive to change in system topology, as well as CIM library update. This innovative way ensures interoperability among the partner platforms and provides support to multi-platform holistic approach for CPES assessment.
Dec 13 2017 cs.SY
In the hierarchical control of an islanded microgrid, secondary control could be centralized or distributed. The former control strategy has several disadvantages, such as single point of failure at the level of the central controller as well as high investment of communication infrastructure. In this paper a three-layer architecture of distributed control is given, including the device layer, the control layer as well as the agent layer. The agent layer is a multi-agent system in which each agent is in charge of a distributed generation unit. Due to communication network constraints, agents are connected only to nearby neighbors. However, by using consensus algorithms the agents can discover the required global information and compute new references for the control layer. The proposed control system is tested on a microgrid scenario which includes paralleled inverter sources. For this, the system is implemented on a real-time cyber-physical test platform that combines real-time simulation models running in OPAL-RT with a network of ARM-based computers, representing the agents.
We propose a novel framework for the differentially private ERM, input perturbation. Existing differentially private ERM implicitly assumed that the data contributors submit their private data to a database expecting that the database invokes a differentially private mechanism for publication of the learned model. In input perturbation, each data contributor independently randomizes her/his data by itself and submits the perturbed data to the database. We show that the input perturbation framework theoretically guarantees that the model learned with the randomized data eventually satisfies differential privacy with the prescribed privacy parameters. At the same time, input perturbation guarantees that local differential privacy is guaranteed to the server. We also show that the excess risk bound of the model learned with input perturbation is $O(1/n)$ under a certain condition, where $n$ is the sample size. This is the same as the excess risk bound of the state-of-the-art.
Oct 09 2017 cs.SY
The integration of distributed renewable energy sources and the multi-domain behaviours inside the cyber-physical energy system (smart grids) draws up major challenges. Their validation and roll out requires careful assessment, in term of modelling, simulation and testing. The traditional approach focusing on a particular object, actual hardware or a detailed model, while drastically simplifying the remainder of the system under test, is no longer sufficient. Real-time simulation and Hardware-in-the-Loop (HIL) techniques emerge as indispensable tools for validating the behaviour of renewable sources as well as their impact/interaction to with the cyber-physical energy system. This paper aims to provide an overview of the present status-quo of real-time and HIL approaches used for smart grids and their readiness for cyber-physical experiments. We investigate the current limitations of HIL techniques and point out necessary future developments. Subsequently, the paper highlights challenges that need specific attention as well as ongoing actions and further research directions.
Oct 09 2017 cs.SY
Composed of a large variety of technologies and applications with unprecedented complexity, the smart grid, as a cyber-physical energy system, needs careful investigation of the interactions between the various domains involved, especially the coupling between the power and information systems. In this paper, two modern ways of modeling and simulating complex cyber-physical energy systems are considered: Co-simulation and Power-Hardware-in-the-Loop experiments. An analysis of these two approaches shows that a complementary and joint setup with realistic behaviors of hardware equipment under a variety of complex environments, co-simulated by several simulators from different domains, create a complete and high performance environment to achieve a holistic approach for smart grid validation and roll out. In the scope of coupling these two techniques, major technical challenges are identified and advanced solutions are outlined.
Oct 09 2017 cs.SY
Smart grid systems are characterized by high complexity due to interactions between a traditional passive network and active power electronic components, coupled using communication links. Additionally, automation and information technology plays an important role in order to operate and optimize such cyber-physical energy systems with a high(er) penetration of fluctuating renewable generation and controllable loads. As a result of these developments the validation on the system level becomes much more important during the whole engineering and deployment process, today. In earlier development stages and for larger system configurations laboratory-based testing is not always an option. Due to recent developments, simulation-based approaches are now an appropriate tool to support the development, implementation, and roll-out of smart grid solutions. This paper discusses the current state of simulation-based approaches and outlines the necessary future research and development directions in the domain of power and energy systems.
Jun 26 2017 cs.CL
Recurrent Neural Network models are the state-of-the-art for Named Entity Recognition (NER). We present two innovations to improve the performance of these models. The first innovation is the introduction of residual connections between the Stacked Recurrent Neural Network model to address the degradation problem of deep neural networks. The second innovation is a bias decoding mechanism that allows the trained system to adapt to non-differentiable and externally computed objectives, such as the entity-based F-measure. Our work improves the state-of-the-art results for both Spanish and English languages on the standard train/development/test split of the CoNLL 2003 Shared Task NER dataset.
Mar 28 2017 cs.LG
Multiple instance data are sets or multi-sets of unordered elements. Using metrics or distances for sets, we propose an approach to several multiple instance learning tasks, such as clustering (unsupervised learning), classification (supervised learning), and novelty detection (semi-supervised learning). In particular, we introduce the Optimal Sub-Pattern Assignment metric to multiple instance learning so as to provide versatile design choices. Numerical experiments on both simulated and real data are presented to illustrate the versatility of the proposed solution.
Mar 13 2017 cs.RO
This paper presents a new optimal filter namely past observation-based extended Kalman filter for the problem of localization of Internet-based mobile robot in which the control input and the feedback measurement suffer from communication delay. The filter operates through two phases: the time update and the data correction. The time update predicts the robot position by reformulating the kinematics model to be non-memoryless. The correction step corrects the prediction by extrapolating the delayed measurement to the present and then incorporating it to the being estimate as there is no delay. The optimality of the incorporation is ensured by the derivation of a multiplier that reflects the relevance of past observations to the present. Simulations in MATLAB and experiments in a real networked robot system confirm the validity of the proposed approach.
This study proposes behavior-based navigation architecture, named BBFM, to deal with the problem of navigating the mobile robot in unknown environments in the presence of obstacles and local minimum regions. In the architecture, the complex navigation task is split into principal sub-tasks or behaviors. Each behavior is implemented by a fuzzy controller and executed independently to deal with a specific problem of navigation. The fuzzy controller is modified to contain only the fuzzification and inference procedures so that its output is a membership function representing the behavior's objective. The membership functions of all controllers are then used as the objective functions for a multi-objective optimization process to coordinate all behaviors. The result of this process is an overall control signal, which is Pareto-optimal, used to control the robot. A number of simulations, comparisons, and experiments were conducted. The results show that the proposed architecture outperforms some popular behavior-based architectures in term of accuracy, smoothness, traveled distance, and time response.
While Multiple Instance (MI) data are point patterns -- sets or multi-sets of unordered points -- appropriate statistical point pattern models have not been used in MI learning. This article proposes a framework for model-based MI learning using point process theory. Likelihood functions for point pattern data derived from point process theory enable principled yet conceptually transparent extensions of learning tasks, such as classification, novelty detection and clustering, to point pattern data. Furthermore, tractable point pattern models as well as solutions for learning and decision making from point pattern data are developed.
Clustering is one of the most common unsupervised learning tasks in machine learning and data mining. Clustering algorithms have been used in a plethora of applications across several scientific fields. However, there has been limited research in the clustering of point patterns - sets or multi-sets of unordered elements - that are found in numerous applications and data sources. In this paper, we propose two approaches for clustering point patterns. The first is a non-parametric method based on novel distances for sets. The second is a model-based approach, formulated via random finite set theory, and solved by the Expectation-Maximization algorithm. Numerical experiments show that the proposed methods perform well on both simulated and real data.
Point patterns are sets or multi-sets of unordered elements that can be found in numerous data sources. However, in data analysis tasks such as classification and novelty detection, appropriate statistical models for point pattern data have not received much attention. This paper proposes the modelling of point pattern data via random finite sets (RFS). In particular, we propose appropriate likelihood functions, and a maximum likelihood estimator for learning a tractable family of RFS models. In novelty detection, we propose novel ranking functions based on RFS models, which substantially improve performance.
Dec 09 2016 cs.CV
Monocular 3D object parsing is highly desirable in various scenarios including occlusion reasoning and holistic scene interpretation. We present a deep convolutional neural network (CNN) architecture to localize semantic parts in 2D image and 3D space while inferring their visibility states, given a single RGB image. Our key insight is to exploit domain knowledge to regularize the network by deeply supervising its hidden layers, in order to sequentially infer intermediate concepts associated with the final task. To acquire training data in desired quantities with ground truth 3D shape and relevant concepts, we render 3D object CAD models to generate large-scale synthetic data and simulate challenging occlusion configurations between objects. We train the network only on synthetic data and demonstrate state-of-the-art performances on real image benchmarks including an extended version of KITTI, PASCAL VOC, PASCAL3D+ and IKEA for 2D and 3D keypoint localization and instance segmentation. The empirical results substantiate the utility of our deep supervision scheme by demonstrating effective transfer of knowledge from synthetic data to real images, resulting in less overfitting compared to standard end-to-end training.
This paper deals with the localization problem of mobile robot subject to communication delay and packet loss. The delay and loss may appear in a random fashion in both control inputs and observation measurements. A unified state-space representation is constructed to describe these mixed uncertainties. Based on it, the optimal linear estimator is developed. The main idea is the derivation of a relevance factor to incorporate delayed measurements to the being estimate. The estimator is then extended for nonlinear systems. The performance of this method is tested within the simulations in MATLAB and the experiments in a real robot system. The good localization results prove the efficiency of the method for the purpose of localization of networked mobile robot.
This paper addresses the localization problem. The extended Kalman filter (EKF) is employed to localize a unicycle-like mobile robot equipped with a laser range finder (LRF) sensor and an omni-directional camera. The LRF is used to scan the environment which is described with line segments. The segments are extracted by a modified least square quadratic method in which a dynamic threshold is injected. The camera is employed to determine the robot's orientation. The prediction step of the EKF is performed by extracting parameters from the kinematic model and input signal of the robot. The correction step is conducted with the implementation of a line matching algorithm and the comparison between line's parameters of the local and global maps. In the line matching algorithm, a conversion matrix is introduced to reduce the computation cost. Experiments have been carried out in a real mobile robot system and the results prove the applicability of the method for the purpose of localization.
Tracking the movement of human eyes is expected to yield natural and convenient applications based on human-computer interaction (HCI). To implement an effective eye-tracking system, eye movements must be recorded without placing any restriction on the user's behavior or user discomfort. This paper describes an eye movement recording system that offers free-head, simple configuration. It does not require the user to wear anything on her head, and she can move her head freely. Instead of using a computer, the system uses a visual digital signal processor (DSP) camera to detect the position of eye corner, the center of pupil and then calculate the eye movement. Evaluation tests show that the sampling rate of the system can be 300 Hz and the accuracy is about 1.8 degree/s.
Sep 02 2016 cs.CV
This study proposes an approach to segment human object from a depth image based on histogram of depth values. The region of interest is first extracted based on a predefined threshold for histogram regions. A region growing process is then employed to separate multiple human bodies with the same depth interval. Our contribution is the identification of an adaptive growth threshold based on the detected histogram region. To demonstrate the effectiveness of the proposed method, an application in driver distraction detection was introduced. After successfully extracting the driver's position inside the car, we came up with a simple solution to track the driver motion. With the analysis of the difference between initial and current frame, a change of cluster position or depth value in the interested region, which cross the preset threshold, is considered as a distracted activity. The experiment results demonstrated the success of the algorithm in detecting typical distracted driving activities such as using phone for calling or texting, adjusting internal devices and drinking in real time.
Recently, spectrum sharing has been considered as a promising solution to improve the spectrum utilization. It however may be vulnerable to security problems as the primary and secondary network access the same resource. Therefore, in this paper, we focus on the performance analysis of a cognitive radio network in the presence of an eavesdropper (EAV) who illegally listens to the primary user (PU) communication in which the transmit power of the secondary transmitter (SU-Tx) is subject to the joint constraint of peak transmit power of the SU-Tx and outage probability of the PU. Accordingly, an adaptive transmit power policy and an analytical expression of symbol error probability are derived for the SU. Most importantly, security evaluations of primary network in terms of the probability of existence of non-zero secrecy capacity and outage probability of secrecy capacity are obtained. Numerical results reveal a fact that the security of the primary network does not only depends on the channel mean powers between primary and secondary networks, but also strongly depends on the channel condition of the SU-Tx to EAV link and transmit power policy of the SU-Tx.
Apr 05 2013 cs.DB
Given a replicated database, a divergent design tunes the indexes in each replica differently in order to specialize it for a specific subset of the workload. This specialization brings significant performance gains compared to the common practice of having the same indexes in all replicas, but requires the development of new tuning tools for database administrators. In this paper we introduce RITA (Replication-aware Index Tuning Advisor), a novel divergent-tuning advisor that offers several essential features not found in existing tools: it generates robust divergent designs that allow the system to adapt gracefully to replica failures; it computes designs that spread the load evenly among specialized replicas, both during normal operation and when replicas fail; it monitors the workload online in order to detect changes that require a recomputation of the divergent design; and, it offers suggestions to elastically reconfigure the system (by adding/removing replicas or adding/dropping indexes) to respond to workload changes. The key technical innovation behind RITA is showing that the problem of selecting an optimal design can be formulated as a Binary Integer Program (BIP). The BIP has a relatively small number of variables, which makes it feasible to solve it efficiently using any off-the-shelf linear-optimization software. Experimental results demonstrate that RITA computes better divergent designs compared to existing tools, offers more features, and has fast execution times.
We present some of the experiments we have performed to best test our design for a library for MathScheme, the mechanized mathematics software system we are building. We wish for our library design to use and reflect, as much as possible, the mathematical structure present in the objects which populate the library.