Intelligent Cooperative Control Architecture: A Framework for Performance Improvement Using Safe Learning

Planning for multi-agent systems such as task assignment for teams of limited-fuel unmanned aerial vehicles (UAVs) is challenging due to uncertainties in the assumed models and the very large size of the planning space. Researchers have developed fast cooperative planners based on simple models (e.g., linear and deterministic dynamics), yet inaccuracies in assumed models will impact the resulting performance. Learning techniques are capable of adapting the model and providing better policies asymptotically compared to cooperative planners, yet they often violate the safety conditions of the system due to their exploratory nature. Moreover they frequently require an impractically large number of interactions to perform well. This paper introduces the intelligent Cooperative Control Architecture (iCCA) as a framework for combining cooperative planners and reinforcement learning techniques. iCCA improves the policy of the cooperative planner, while reduces the risk and sample complexity of the learner. Empirical results in gridworld and task assignment for fuel-limited UAV domains with problem sizes up to 9 billion state-action pairs verify the advantage of iCCA over pure learning and planning strategies.     

Short communication: Comparison of the fecal bacterial communities in diarrheic and nondiarrheic dairy calves from multiple farms in southeastern Pennsylvania

Diarrhea is a major cause of illness and death in preweaned calves and causes significant economic losses to producers. A better understanding of the fecal microbiota in diarrheic and nondiarrheic calves could lead to improved treatment and prevention strategies. The purpose of this study was to compare the fecal microbiota of diarrheic and nondiarrheic calves to improve our understanding of what constitutes a healthy fecal microbiota in preweaned calves. At each of 7 farms, fecal samples were obtained from 1 to 3 diarrheic Holstein dairy calves (2 to 17 d old at sampling time) and age-matched (within 5 d) nondiarrheic controls for a total of 20 samples. Calves were fed either acidified bulk milk, pasteurized or unpasteurized waste milk, or milk replacer depending on farm. Fecal samples were extracted for genomic DNA, PCR-amplified for the V1–V2 region of the 16S rRNA bacterial gene, sequenced on the Illumina MiSeq (Illumina Inc., San Diego, CA) platform, and analyzed using QIIME2. Firmicutes and Bacteroidetes were the most abundant phyla in both groups; Fusobacteria was numerically more abundant in the diarrheic group, whereas Proteobacteria and Actinobacteria were numerically more abundant in the nondiarrheic group. At the genus level, Bacteroides was the most abundant genus in both groups and was numerically more abundant in the nondiarrheic group. Results from the mixed-effects regression model showed that Faecalibacterium and Butyricimonas were more abundant in the nondiarrheic calves, whereas Clostridium and Peptostreptococcus were more abundant in the diarrheic calves. Our results indicate that commensal bacteria acquired in the neonatal period may have been replaced with potential pathogens in diarrheic calves, which may have contributed to the incidence of diarrhea either directly or indirectly.     

Dynamic models of Fabry-Perot interferometers

Long-baseline, high-finesse Fabry-Perot interferometers can be used to make distance measurements that are precise enough to detect gravity waves. This level of sensitivity is achieved in part when the interferometer mirrors are isolated dynamically, with pendulum mounts and high-bandwidth cavity length control servos to reduce the effects of seismic noise. We present dynamical models of the cavity fields and signals of Fabry-Perot interferometers for use in the design and evaluation of length control systems for gravity-wave detectors. Models are described and compared with experimental data.     

Demographic variables, smoking variables, and outcome across five studies.

Objective: Intervention effectiveness can potentially be affected by membership in different demographic subgroups (race, ethnicity, gender, age, and education level) or smoking behavior variables (time to first cigarette, longest previous quit attempt, number of attempts in the past year, number of cigarettes, and stage of change). Previous research on these 2 sets of variables has produced mixed results. Design: This secondary data analysis combined data from 5 effectiveness trials (a random-digit-dial sample [N = 1,358], members of an HMO [N = 207], parents of students recruited for a school-based study [N = 347], patients from an insurance provider list [N = 535], and employees [N = 175]) in which smokers were all proactively recruited from a defined population and all received the same expert system intervention. The intervention produced a consistent 22% to 26% point prevalence cessation rate across the 5 studies. Main Outcome Measures: The main outcome measures were 24-hr point prevalence, 7-day point prevalence, 30-day prolonged abstinence, and 6-month prolonged abstinence. Results: There were no significant differences in outcome across gender, race, and ethnicity subgroups. There were significant differences and small effect sizes for age and education subgroups. There were significant differences and large effect sizes for all 5 smoking behavior variables. Discussion: Demographic variables are static variables, whereas the smoking variables are more dynamic, that is, open to change. Given the dynamic nature of the smoking variables and the large effect sizes, interventions tailored on the smoking variables should be more successful. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Calculation of diffraction effects on the average phase of an optical field

We report on algorithms for the computation of the average phase of a beam over a detector in the near field. The basic idea is to reconstruct the optical field numerically and then use a quadrature algorithm to evaluate the quantity of interest. The various algorithms that employ discrete Fourier transform techniques for the computation of the field are described, and numerical tests that assess the accuracy of these algorithms are presented. No particular algorithm delivers the desired accuracy over the entire range of Fresnel numbers of interest, but each can produce satisfactory results within a particular range. Finally, new methods to evaluate the average phase are introduced, and their efficiency is assessed.     

Generalization of prism adaptation.

Prism exposure produces 2 kinds of adaptive response. Recalibration is ordinary strategic remapping of spatially coded movement commands to rapidly reduce performance error. Realignment is the extraordinary process of transforming spatial maps to bring the origins of coordinate systems into correspondence. Realignment occurs when spatial discordance signals noncorrespondence between spatial maps. In Experiment 1, generalization of recalibration aftereffects from prism exposure to postexposure depended upon the similarity of target pointing limb postures. Realignment aftereffects generalized to the spatial maps involved in exposure. In Experiment 2, the 2 kinds of aftereffects were measured for 3 test positions, one of which was the exposure training position. Recalibration aftereffects generalized nonlinearly, while realignment aftereffects generalized linearly, replicating Bedford (1989, 1993a) using a more familiar prism adaptation paradigm. Recalibration and realignment require methods for distinguishing their relative contribution to prism adaptation. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Efficient calculation of finite Gabor transforms

The Gabor transform may be viewed as a collection of localized Fourier transforms and as such is useful for analysis of nonstationary signals and images. We present a new approach to analyzing the Gabor transform and use it to study the various critically sampled discretizations that form the infinite-discrete, periodic finite-discrete, and nonperiodic finite-discrete versions of the transform. In particular, we distinguish between the analysis and synthesis forms of the transform, and introduce an intermediate operation that decomposes both forms into collections of independent Toeplitz operators. In the continuous, the infinite-discrete, and the periodic finite-discrete cases, this decomposition allows us to show that, for appropriate windows, the analysis and synthesis transforms are inverses of each other. In the nonperiodic finite-discrete case this relation no longer holds, but we are still able to use the decomposition and results on Toeplitz matrices to show that both the transform and the inverse transform of P discrete samples are computable in O(P log P) operations (after a setup cost of O(Plog²P)). Furthermore, we use the decomposition to study in detail the differences between the periodic and nonperiodic versions of the transform and to compare their conditioning     

Clinical evaluation and microbiology of oropharyngeal infection due to fluconazole-resistant Candida in human immunodeficiency virus-infected patients

The aim of the present study was to investigate pregnancy rates ensuing from transfer of embryos with multinucleated blastomeres. In our in-vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) programme, 1735 embryo transfers were performed from January 1, 1995 to August 31, 1996. In 136 of these transfers at least one embryo with one or more multinucleated blastomeres was present per transfer (study group). For each of these 136 transfers, two matched controls with transfer of exclusively mononucleated embryos were selected (control group). Matching was carried out according to age, method of fertilization (IVF or ICSI), number of transferred embryos and quality score of transferred embryos. In the study group, there were eight transfers of exclusively multinucleated embryos from which one pregnancy ensued and 128 transfers in which multinucleated and mononucleated embryos were transferred together leading to 23 pregnancies. The overall clinical pregnancy rate per transfer was 16.9% in the study group versus 28.7% in the control group (P = 0.01). The ongoing pregnancy rate per transfer was 13.2% in the study group versus 23.2% in the control group (P = 0.03). The implantation rate per transferred embryo was 6.0% in the study group versus 11.3% in the control group (P = 0.003). This study shows that embryos with one or more multinucleated blastomeres have a poorer implantation potential than embryos with mononucleated blastomeres. Transfer of embryos with multinucleated blastomeres should hence only be considered when insufficient numbers of embryos with only mononucleated blastomeres are present.     

On the extraction of 3D models from airborne video sensors for geolocation

Geolocation of a feature in a video sequence collected from a moving platform is an activity that must be undertaken in video exploitation, especially surveillance and reconnaissance applications. Examples of sensor systems that are the focus of this work include manned and unmanned aerial vehicles. The approach described here uses positional information from three sources to compute refined coordinates in three dimensions for any feature in the video sequence. These three sources are: first, sensor-platform metadata describing the likely sensor footprint based on sensor-platform positional and attitudinal information; second, 3D information of a scene inherent in a video sequence collected from a moving platform; and third, reference imagery of the region of interest that is geolocated and georectified such as aerial photography. We describe the overall steps involved in this process and the progress made to date.