Affinity graphs are widely used in deep architectures, including graph convolutional neural networks and attention networks. Thus far, the literature has focused on abstracting features from such graphs, while the learning of the affinities themselves has been overlooked. Here we propose a principled method to directly supervise the learning of weights in affinity graphs, to exploit meaningful connections between entities in the data source. Applied to a visual attention network, our affinity supervision improves relationship recovery between objects, even without the use of manually annotated relationship labels. We further show that affinity learning between objects boosts scene categorization performance and that the supervision of affinity can also be applied to graphs built from mini-batches, for neural network training. In an image classification task we demonstrate consistent improvement over the baseline, with diverse network architectures and datasets.

We consider how to directly extract a road map (also known as a topological representation) of an initially-unknown 2-dimensional environment via an on-line procedure which robustly computes a retraction of its boundaries. While such approaches are well known for their theoretical elegance, computing such representations in practice is complicated when the data is sparse and noisy. In this paper we present the online construction of a topological map and the implementation of a control law for guiding the robot to the nearest unexplored area. The proposed method operates by allowing the robot to localize itself on a partially constructed map, calculate a path to unexplored parts of the environment (frontiers), compute a robust terminating condition when the robot has fully explored the environment, and achieve loop closure detection. The proposed algorithm results in smooth safe paths for the robot's navigation needs. The presented approach is an any-time-algorithm which allows for the active creation of topological maps from laser-scan data, as it is being acquired. The resulting map is stable under variations to noise and the initial conditions. The key idea is the use of a flux-based skeletonization algorithm on the latest occupancy grid map. We also propose a navigation strategy based on a heuristic where the robot is directed towards nodes in the topological map that open to empty space. The method is evaluated on both synthetic data and in the context of active exploration using a Turtlebot 2. Our results demonstrate complete mapping of different environments with smooth topological abstraction without spurious edges.

Currently, the increasing rate of routing lookups in Internet routers, the large number of prefixes and also the transition from IPV4 to IPV6, have caused Internet designers to propose new lookup algorithms and try to reduce the memory cost and the prefix search and update procedures times. Recently, some new algorithms are proposed trying to store the prefixes in a balanced tree to reduce the worst case prefix search and update times. These algorithms improve the search and update times compared to previous range based trees. In this paper it is shown that there is no need to treat the prefixes as ranges. It is only required to compare them like scalar values using a predefined rule. The method "scalar prefix search" which is presented here, is built on this concept and combining it with the proposed store and search methods, interprets each prefix as a number without any encoding, the need to convert it to the prefix end points or to use the trie based algorithms whose performance completely depends on IP address length. This method can be applied to many different tree structures. It is implemented using the binary search tree and some other balanced trees such as RB-tree, AVL-tree and B-tree for both IPV4 and IPV6 prefixes. Comparison results show better lookup and update performance or superior storage requirements for scalar prefix search in both average and worst cases, against current solutions like PIBT (Lu and Sahni, 2005) and LPFST (Wnn et al., 2005).