We present a novel vector quantization (VQ) module for the two state-of-the-art long-range simultaneous localization and mapping (SLAM) algorithms. The VQ task in SLAM is generally performed using unsupervised methods. We provide an alternative approach trough embedding a semisupervised hyperbolic graph convolutional neural network (HGCN) in the VQ step of the SLAM processes. The SLAM platforms we have utilized for this purpose are fast appearance-based mapping (FABMAP) and oriented fast and rotated short (ORB), both of which rely on extracting the features of the captured images in their loop closure detection (LCD) module. For the first time, we have considered the space formed by these SURF features, robust image descriptors, as a graph, enabling us to apply an HGCN in the VQ section which results in an improved LCD performance. The HGCN vector quantizes the SURF feature space, leading to a bag-of-word (BoW) representation construction of the images. This representation is subsequently used to determine LCD accuracy and recall. Our approaches in this study are referred to as HGCN-FABMAP and HGCN-ORB. The main advantage of using HGCN in the LCD section is that it scales linearly when the features are accumulated. The benchmarking experiments show the superiority of our methods in terms of both trajectory generation accuracy in small-scale paths and LCD accuracy and recall for large-scale problems.

A large volume of unstructured data, especially text data, is generated and exchanged daily. Consequently, the importance of extracting patterns and discovering knowledge from textual data is significantly increasing. As the task of automatically recognizing the relations between two or more entities, semantic relation extraction has a prominent role in the exploitation of raw text. This article surveys different approaches and types of relation extraction in English and the most prominent proposed methods in Persian. We also introduce, analyze, and compare the most important datasets available for relation extraction in Persian and English. Furthermore, traditional and emerging evaluation metrics for supervised, semi-supervised, and unsupervised methods are described, along with pointers to commonly used performance evaluation datasets. Finally, we briefly describe challenges in extracting relationships in Persian and English and dataset creation challenges.

In this paper, a new mobile robot mapping algorithm inspired from the functionality of hippocampus cells is presented. Place cells in hippocampus can store a map of the environment. This model fuses odometry and vision data based on dimensionality reduction technique, hierarchically. These two types of data are first fused and then considered as inputs to the place cell model. Place cells do the clustering of places. The proposed Place cell model has two types of inputs: Grid cells input and input from the lateral entorhinal cortex (LEC). The LEC is modelled based on the dimension reduction technique. Therefore, the data that causes locations different to be inserted into the place cell from this layer. Another contribution is proposing a new unsupervised dimension reduction method based on k-means. The method can find perpendicular independent dimensions. Also, the distance of cluster centres found in these dimensions is maximised. The method was compared with LDA and PCA in standard functions. Although LDA is a supervised method, the result showed that the proposed unsupervised method outperformed. To evaluate the place cells model, sequences of images collected by a mobile robot was used and similar results to real place cells achieved.

We present improvements in maximum a-posteriori inference for Markov Logic, a widely used SRL formalism. Several approaches, including Cutting Plane Aggregation (CPA), perform inference through translation to Integer Linear Programs. Aggregation exploits context-specific symmetries independently of evidence and reduces the size of the program. We illustrate much more symmetries occurring in long ground clauses that are ignored by CPA and can be exploited by higher-order aggregations. We propose Full-Constraint-Aggregation, a superior algorithm to CPA which exploits the ignored symmetries via a lifted translation method and some constraint relaxations. RDBMS and heuristic techniques are involved to improve the overall performance. We introduce Xeggora as an evolutionary extension of RockIt, the query engine that uses CPA. Xeggora evaluation on real-world benchmarks shows progress in efficiency compared to RockIt especially for models with long formulas.

We present improvements in maximum a-posteriori inference for Markov Logic, a widely used SRL formalism. Inferring the most probable world for Markov Logic is NP-hard in general. Several approaches, including Cutting Plane Aggregation (CPA), perform inference through translation to Integer Linear Programs. Aggregation exploits context-specific symmetries independently of evidence and reduces the size of the program. We illustrate much more symmetries occurring in long ground clauses that are ignored by CPA and can be exploited by higher-order aggregations. We propose Full-Constraint-Aggregation, a superior algorithm to CPA which exploits the ignored symmetries via a lifted translation method and some constraint relaxations. RDBMS and heuristic techniques are involved to improve the overall performance. We introduce Xeggora as an evolutionary extension of RockIt, the query engine that uses CPA. Xeggora evaluation on real-world benchmarks shows progress in efficiency compared to RockIt especially for models with long formulas.

In this paper, we propose a kernel for nonlinear dimensionality reduction over the manifold of Symmetric Positive Definite (SPD) matrices in a Motor Imagery (MI)-based Brain Computer Interface (BCI) application. The proposed kernel, which is based on Riemannian geometry, tries to preserve the topology of data points in the feature space . Topology preservation is the main challenge in nonlinear dimensionality reduction (NLDR). Our main idea is to decrease the non-Euclidean characteristics of the manifold by modifying the volume elements . We apply a conformal transform over data-dependent isometric mapping to reduce the negative eigen fraction to learn a data dependent kernel over the Riemannian manifolds. Multiple experiments were carried out using the proposed kernel for a dimensionality reduction of SPD matrices that describe the EEG signals of dataset IIa from BCI competition IV. The experiments show that this kernel adapts to the input data and leads to promising results in comparison with the most popular manifold learning methods and the Common Spatial Pattern (CSP) technique as a reference algorithm in BCI competitions. The proposed kernel is strong, particularly in the cases where data points have a complex and nonlinear separable distribution.

Neurophysiological and computational studies have proposed that properties of natural images have a prominent role in shaping selectivity of neurons in the visual cortex. An important property of natural images that has been studied extensively is the inherent redundancy in these images. In this paper, the concept of category-specific redundancies is introduced to describe the complex pattern of dependencies between responses of linear filters to natural images. It is proposed that structural similarities between images of different object categories result in dependencies between responses of linear filters in different spatial scales. It is also proposed that the brain gradually removes these dependencies in different areas of the ventral visual hierarchy to provide a more efficient representation of its sensory input. The authors proposed a model to remove these redundancies and trained it with a set of natural images using general learning rules that are developed to remove dependencies between responses of neighbouring neurons. Results of experiments demonstrate the close resemblance of neuronal selectivity between different layers of the model and their corresponding visual areas.

Efficient coding hypothesis states that the goal of sensory system of the brain is to remove redundancies in the sensory input. Several models tried to remove redundancy in the visual input and successfully modeled the functional properties of neurons in the primary visual cortex. However there has been no progress to extend these models to simulate the properties of neurons in extrastriate visual areas. In this paper, we propose that visual cortex tries to remove higher order dependencies in a hierarchical architecture. In each layer a nonlinear mechanism removes redundancies in a local neighborhood. We used the biologically plausible divisive normalization mechanism in a two layer model network to remove nonlinear dependencies in the input. Units in this model can simulate the responses of neurons in area V2 to angle stimuli.

Statistical properties of natural signals is an important factor in forming neuronal selectivities of brain sensory system. One such property is the redundancy in visual and auditory inputs to the brain. In this paper, we introduce the concept of class specific redundancies in natural images and propose that the selectivity of neurons in extrastriate visual areas is developed to reveal these redundancies. In each extrastriate area, a redundancy reduction mechanism removes these revealed redundancies to provide a more efficient representation of the input image. To test this hypothesis, we implemented a model of area V2 and trained this model with a set of natural images. Experiments on artificial stimulus sets and natural images show the close similarity of model neurons to real V2 neurons and their preference for coding object contours over textures.

Network administrators are able to correlate log file entries manually. Large volume and low quality of log files justify the need for further log processing. The manual log processing is lack of flexibility. It is time consuming, and one doesn't get the general view of the log files in the network. Without this general view it is hard to correlate information between the network components. Events seemingly unessential by themselves can in reality be a piece of a larger threat. In this regard, different log correlation methods are proposed to improve alert quality and to give a comprehensive view of system security. In this paper, we show how different attacks categorized in three categories with different behavior: Denial of service (DoS) attacks, user-to-root (U2R) & remote-to-local (R2L) attacks and probing, are reflected in different logs and argue that some attacks are not evident when a single log is analyzed.

For a mobile robot to move in a known environment and operate successfully, first it needs to robustly determine its initial position and orientation relative to the map, and then update its position while moving in the environment. Thus determining robot’s position is one of the most important tasks in mobile robotics. This task consists of “global localization‟ and “robot’s pose tracking‟. In this paper two recent sample-based evolutionary methods for globally localizing the position of a mobile robot are proposed. The first method is a modified version of genetic algorithm called Differential Evolution (DE) which is based on natural selection. The second one is Particle Swarm Optimization (PSO) which is based on bird flocking. DE evaluates initial population using the probabilistic motion and observation models and the evolution of the individuals is performed by evolutionary operators. PSO adjusts the velocity and location of particles towards target (robot’s pose) through a problem space on the basis of information about each particle’s previous best location and the best previous location of its neighbors. Our results illustrate the excellence of these two methods over standard Monte Carlo localization algorithm with regard to convergence rate, speed and computational cost.

Mobile robot localization is taken into account as one of the most important topics in robotics. In this paper, the localization problem is extended to the cases in which estimating the position of multi robots is considered. To do so, the joint probabilistic data association filter (JPDAF) approach is applied for tracking the position of multiple robots. To characterize the motion of each robot, two models are used. First, a simple near constant velocity model is considered and then a variable velocity model is applied for tracking. This improves the performance when the robots change their velocity and conduct maneuvering movements. This issue gives an advantage to explore the movement of the manoeuvring objects which is common in many robotics problems such as soccer or rescue robots. Simulation results show the efficiency of the JPDAF algorithm in tracking multiple mobile robots with maneuvering movements.

the piezoelectric actuator (PEA) accuracy is limited due to hysteresis nonlinearity as well as mechanical loading effect. Investigation of the fundamental properties of the piezoceramics depicts that the external mechanical loads cause inclination in hysteresis loop can deteriorate tracking performance furthermore. Other limiting problem is the rate dependency of hysteresis nonlinearity in PEA. Herein a new controller scheme to compensate mentioned limiting factors in PEA is presented. In this study an adaptive inverse control method based on modified Prandtle-Ishlinskii operator is proposed. It compensates both the rate dependant hysteresis nonlinearity and the mechanical loading effect in PEA. The controller structure has a simple design and can be quickly identified. The control system is capable to achieve suitable tracking control only by measuring the applied external load and rate of control input. Therefore it is convenient to use and can be quickly applied to the practical PEA applications. To verify the efficiency of the proposed method, we have conducted simulation by using the PEA system model as the unknown piezo actuator (PEA) dynamic.

In this paper we present a linear time invariant controller for bilateral teleoperation of a pair of N-DOF linear robotic systems under constant time delay. Our framework uses position and velocity signals to achieve position tracking in free motion. It also uses force feedback for compensating the human/environment disturbances to achieve position coordination in contact tasks. In passifying position and velocity feedback terms, for simplicity reasons there is no need to consider the dynamics of the robotic systems as long as they satisfy Euler-Lagrangian dynamic system equation. However when the force feedback is applied, we have to consider the dynamics of the robot in controller design. We do passivity analysis in the frequency domain by using Parseval’s identity. Simulation has been performed to illustrate the proposed algorithm and to clarify some points in the design procedure.

Information overload is no longer news; the explosive growth of the Internet has made this issue increasingly serious for Web users. Users are very often overwhelmed by the huge amount of information and are faced with a big challenge to find the most relevant information in the right time. Recommender systems aim at pruning this information space and directing users toward the items that best meet their needs and interests. Web Recommendation has been an active application area in Web Mining and Machine Learning research. In this paper we propose a novel machine learning perspective toward the problem, based on reinforcement learning. Unlike other recommender systems, our system does not use the static patterns discovered from web usage data, instead it learns to make recommendations as the actions it performs in each situation. We model the problem as Q-Learning while employing concepts and techniques commonly applied in the web usage mining domain. We propose that the reinforcement learning paradigm provides an appropriate model for the recommendation problem, as well as a framework in which the system constantly interacts with the user and learns from her behavior. Our experimental evaluations support our claims and demonstrate how this approach can improve the quality of web recommendations.

Describes an interface for multi modal human robot interaction, which enables people to introduce a newcomer robot to different attributes of objects and places in the room through speech commands and hand gestures. The robot makes an environment map of the room based on knowledge learned through communication with human and uses this map for navigation. The developed system consists of several sections including: natural language processing, posture recognition, object localization and map generation. This system uses a combination of multiple sources of information and model matching to detect and track a human hand so that the user can point toward an object of interest and guide the robot to go near to it or locate that object's position in the room. The position of objects in the room is located by a monocular camera vision and depth from focus method.

The system introduced in the paper estimates the 3D position of a human's head and face by using the depth from focus method. First the system locates the human's head and face by using motion detection, Hough transform and a statistical color model. Then by changing pan, tilt and zoom of camera we put the face in the center of the camera's field of view. Then we use information from focusing the ring of an autofocus camera to measure the distance between the human face and the camera. By having information about the absolute position of the camera and its pan and tilt angles we can compute the absolute position of the person in the room. In the case of a person with small motions, this system measures distance with less than a 10 cm error. The measurable distance range is from 90 cm up to 340 cm.