Reinforcement learning of pedagogical policies in adaptive and intelligent educational systems

In an adaptive and intelligent educational system (AIES), the process of learning pedagogical policies according the students needs fits as a Reinforcement Learning (RL) problem. Previous works have demonstrated that a great amount of experience is needed in order for the system to learn to teach properly, so applying RL to the AIES from scratch is unfeasible. Other works have previously demonstrated in a theoretical way that seeding the AIES with an initial value function learned with simulated students reduce the experience required to learn an accurate pedagogical policy. In this paper we present empirical results demonstrating that a value function learned with simulated students can provide the AIES with a very accurate initial pedagogical policy. The evaluation is based on the interaction of more than 70 Computer Science undergraduate students, and demonstrates that an efficient and useful guide through the contents of the educational system is obtained.     

A GPU framework for parallel segmentation of volumetric images using discrete deformable models

Despite the ability of current GPU processors to treat heavy parallel computation tasks, its use for solving medical image segmentation problems is still not fully exploited and remains challenging. A lot of difficulties may arise related to, for example, the different image modalities, noise and artifacts of source images, or the shape and appearance variability of the structures to segment. Motivated by practical problems of image segmentation in the medical field, we present in this paper a GPU framework based on explicit discrete deformable models, implemented over the NVidia CUDA architecture, aimed for the segmentation of volumetric images. The framework supports the segmentation in parallel of different volumetric structures as well as interaction during the segmentation process and real-time visualization of the intermediate results. Promising results in terms of accuracy and speed on a real segmentation experiment have demonstrated the usability of the system.     

Distributed parameters for low-frequency dielectric characterization of liquids with open-ended coaxial cell

The theory of distributed parameters is applied at low frequencies to the open-ended coaxial-line dielectric characterization technique. A procedure is proposed to determine the permittivity of liquids when the sample is placed between two dielectric beads to ensure tight closure. The technique leads to a complex transcendental equation that the authors solve. An ad hoc sample holder is built to implement the proposed method. Measurements of both polar and nonpolar liquids demonstrate the adequacy of the method and its advantages with respect to the lumped-circuit theory in the case of polar samples.     

Antioxidant mechanism of grape procyanidins in muscle tissues: Redox interactions with endogenous ascorbic acid and α-tocopherol

The present study investigates the antioxidant mechanism of grape procyanidins and, in particular, their aptitude to establish redox interactions with two important components of the endogenous antioxidant system of muscle tissues, α-tocopherol (α-TOH) and ascorbic acid (AA). To this end, the progress of lipid oxidation was monitored in fish muscle supplemented with grape procyanidins at the concentrations usually employed in antioxidant food applications, and then related to the redox stability of the endogenous α-TOH and AA. In addition to the lipid oxidation protective effect, the incorporation of procyanidins also provided an improvement of the redox stability of the endogenous components in a straight procyanidinic concentration-dependent manner. Results showed the capacity of procyanidins to repair oxidised α-TOH at medium-long term, and to delay the AA depletion. Therefore, such cooperative redox interaction of exogenous procyanidins adequately complements the natural α-TOH regenerative system supplied by AA that is efficient at the early post mortem stages.     

An efficient algorithm for optimal routing applied to convoy merging manoeuvres in urban environments

This paper describes an innovative routing strategy for intelligent transportation units willing to perform merging manoeuvres with a moving convoy. In particular, we consider a transportation unit located inside a city (pursuer unit), and which wishes to join a convoy that is constantly moving around the city. We first describe a solution that considers idealistic conditions, i.e., the traveling time along each street is constant. We then go on to improve our first approach to deal with the realistic random nature of the traveling times experienced by the pursuer and by the convoy leader. Our search strategy applies Dynamic Programming to achieve a meeting point that is optimal in two ways: on one hand, the optimal destination is the one closest to the current pursuer??s position; on the other hand, the optimal meeting point must minimize the time that elapses until the pursuer meets the convoy (considering that the pursuer must always arrive first). Calculating the optimal path to every possible destination is highly inefficient, error prone and time consuming. On the contrary, we propose an efficient search strategy that will achieve the meeting point and the path to it, both optimally and in a very short time. This enables the real-time application of our approach either for finding new solutions or for re-planning old ones owing to unexpected real-conditions.     

Application of systems identification to the implementation of motion camouflage in mobile robots

Motion camouflage is an animal stealth behaviour in which a shadower—or predator—moves in the vicinity of a shadowee—or prey—in such a way that the later perceives no apparent motion apart from the self motion. Despite some light has been shed on the control mechanism generating this pursuit strategy, it is not fully understood. Motion camouflage represents an interesting challenge in biological motion, and although simulated controllers can be found in the literature, no implementation on real robots has been done so far. This paper presents the first implementation of motion camouflage in real wheeled robots through a polynomial NARMAX model controller. The trajectories to adjust the model are generated using a heuristic approach. The NARMAX model outperforms the heuristic approach in terms of computational time and generates good camouflage trajectories in real robots and simulation. The transparency of polynomial models can also shed some light over this complex animal behaviour.     

Vehicle speed measurement: Cosine error correction

Although laser devices are commonly used to measure distances and vehicle speed in traffic safety-related studies, they suffer from a drawback known as the cosine effect. Furthermore, such traffic safety studies analyze both vehicle speed and a large amount of geographical data. As such, the development of efficient and cost-effective techniques for correcting the cosine error and managing the data involved in vehicle speed studies are needed. This paper presents an algorithm for correcting cosine error based on a Geographic Information System (GIS) that makes use of vehicle speed measurements and Global Positioning System (GPS) coordinates. An experiment using more than 350 vehicle speeds was conducted to test the algorithm, and the case study shows that controlling the quality of laser-measured speeds in order to suppress low quality data or correct measurement errors due to the cosine effect is very useful.     

English speakers attend more strongly than Spanish speakers to manner of motion when classifying novel objects and events.

Three experiments provide evidence that the conceptualization of moving objects and events is influenced by one's native language, consistent with linguistic relativity theory. Monolingual English speakers and bilingual Spanish/English speakers tested in an English-speaking context performed better than monolingual Spanish speakers and bilingual Spanish/English speakers tested in a Spanish-speaking context at sorting novel, animated objects and events into categories on the basis of manner of motion, an attribute that is prominently marked in English but not in Spanish. In contrast, English and Spanish speakers performed similarly at classifying on the basis of path, an attribute that is prominently marked in both languages. Similar results were obtained regardless of whether categories were labeled by novel words or numbered, suggesting that an English-speaking tendency to focus on manner of motion is a general phenomenon and not limited to word learning. Effects of age of acquisition of English were also observed on the performance of bilinguals, with early bilinguals performing similarly in the 2 language contexts and later bilinguals showing greater contextual variation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)