Optimal ensemble construction via meta-evolutionary ensembles

In this paper, we propose a meta-evolutionary approach to improve on the performance of individual classifiers. In the proposed system, individual classifiers evolve, competing to correctly classify test points, and are given extra rewards for getting difficult points right. Ensembles consisting of multiple classifiers also compete for member classifiers, and are rewarded based on their predictive performance. In this way we aim to build small-sized optimal ensembles rather than form large-sized ensembles of individually-optimized classifiers. Experimental results on 15 data sets suggest that our algorithms can generate ensembles that are more effective than single classifiers and traditional ensemble methods.     

浅谈移动通信基站天线系统设计

对基站天线的基本原理、天线选型、天线参数设置、天线安装等方面进行了描述，对天线下倾角、分集距离等进行了重点讨论。     

MIMO系统中的分组全分集全码率空时编码

针对未编码的多输入多输出（MIMO，Multi—Input Multi—Output）系统，提出一种复杂度适中的分组全分集全码率（GFDFR，Group—wise Full Diversity Full Rate）空时编码方案。该方案通过在发送端进行天线分组，各组独立编码，减小全分集全码率（FDFR，Full Diversity Full Rate）编码块的大小从而降低系统编解码复杂度；在频率选择性信道中，进一步对子载波分组进行独立编码，获得频率分集（或多径分集），以适中的复杂度在不降低系统分集度的情况下保证了信息的全码率传输，是一种在MIMO信道中极具实用价值的空时编码方案。     

2D artificial life system using network-type assembly-like language: influence of change in environment with costs of instructions

To unlock the full potential of evolution in digital media, a Tierra-like system using network-type assembly-like language has been constructed. In the system, like Avida, digital creatures, self-replicating programs, live in a discrete 2D torus space and an interaction between creatures is restricted locally. Bearing a genetic network in mind, network structure is introduced. In the previous works, it is shown that the possibility that the network-type model has more potential of evolution than a linear-type model like Avida in a simple environment. In this paper, to study the potential of evolution more precisely, we model the effects of environment other than creatures. As one of the simplest models of such environment, the cost of execution, i.e. executing time, is introduced. The difficult environment to live in costs high to execute instructions, the easy environment does low. In computer experiments, we have investigated the influence of change in environment by analyzing the process of evolution and diversity of the system. Experimental results show that the network-type system keeps more stable diversity than the linear-type system does, even when the environment changes drastically. This indicates the possibility that the network-type system has more potential of evolution than the linear-type system has. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

A probabilistic simulator for population dynamics of quasispecies

We constructed a probabilistic simulator that allows all the events in population dynamics such as death, birth, mutation, and suppression/stimulation to be described by probabilistic rules. The simulator also facilitates a lattice used for expressing distribution and diversity (number of distinct strains) of quasispecies. The simulator is used to investigate the diversity threshold in HIV and T-cell interaction. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

High nutrient loading and climatic parameters influence the dominance and dissimilarity of toxigenic cyanobacteria in northern bays of Lake Victoria

Eutrophication of Lake Victoria led to changes in its phytoplankton communities. However, different levels of eutrophication exist in the open lake and the bays, and between embayments. This study utilized spatial and temporal sampling of Napoleon Gulf and Murchison Bay, exhibiting different trophic conditions. Over one year, we investigated phytoplankton biomass, richness, diversity and dissimilarity, and related the dynamics of the dominant species to the limnological and climatic conditions. The results confirmed that Napoleon Gulf and Murchison Bay showed large differences in eutrophication status, with lower nutrient concentrations in Napoleon Gulf than in Murchison Bay, where a strong gradient was observed from inshore to offshore areas. These nutrient dynamics resulted in a 4 to 10 fold higher chlorophyll-a in Murchison Bay than in Napoleon Gulf. From the embayments, 135 phytoplankton taxa were recorded with no significant differences in alpha diversity. However, high dissimilarity in community structure was observed in beta diversity, mostly due to a turnover among the dominant toxigenic species. Thus, from a similar species pool, there was a shift in the dominant toxigenic cyanobacteria from Microcystis flos-aquae and M. aeruginosa in Murchison Bay, Dolichospermum circinale and Planktolyngbya circumcreta in Napoleon Gulf to D. circinale in the offshore stations. These cyanobacteria are toxigenic taxa with known health hazards. Using partial least square models, we showed that both climatic variables (e.g. wind, solar radiation) and levels of inorganic dissolved nutrients (e.g. SRP, NO₃^–, and NH₄⁺) are the main drivers of differences and dominance in cyanobacteria communities in northern Lake Victoria.     

Random Forests with ensemble of feature spaces

Random Forests receive much attention from researchers because of their excellent performance. As Breiman suggested, the performance of Random Forests depends on the strength of the weak learners in the forests and the diversity among them. However, in the literature, many researchers only considered pre-processing of the data or post-processing of the Random Forests models. In this paper, we propose a new method to increase the diversity of each tree in the forests and thereby improve the overall accuracy. During the training process of each individual tree in the forest, different rotation spaces are concatenated into a higher space at the root node. Then the best split is exhaustively searched within this higher space. The location where the best split lies decides which rotation method to be used for all subsequent nodes. The performance of the proposed method here is evaluated on 42 benchmark data sets from various research fields and compared with the standard Random Forests. The results show that the proposed method improves the performance of the Random Forests in most cases.     

Creature grammar for creative modeling of 3D monsters

Monsters and strange creatures are frequently demanded in 3D games and movies. Modeling such kind of objects calls for creativity and imagination. Especially in a scenario where a large number of monsters with various shapes and styles are required, the designing and modeling process becomes even more challenging. We present a system to assist artists in the creative design of a large collection of various 3D monsters. Starting with a small set of shapes manually selected from different categories, our system iteratively generates sets of monster models serving as the artist’s reference and inspiration. The key component of our system is a so-called creature grammar, which is a shape grammar tailored for the generation of 3D monsters. Creature grammar governs the evolution from creatures with regular structures gradually into monsters with more and more abnormal structures through evolving the arrangement and number of shape parts, while preserving the semantics prescribed as prior knowledge. Experiments show that even starting with a small set of shapes from a few categories of common creatures (e.g., humanoids, bird-like creatures and quadrupeds), our system can produce a large set of unexpected monsters with both shape diversity and visual plausibility, thus providing great support for the user’s creative design.     

Analysis of effect of nicotine on microbial community structure in sediment using PCR-DGGE fingerprinting

Solid or liquid waste containing a high concentration of nicotine can pollute sediment in rivers and lakes, and may destroy the ecological balance if it is directly discharged into the environment without any treatment. In this study, the polymerase chain reaction(PCR) and denaturing gradient gel electrophoresis(DGGE) method was used to analyze the variation of the microbial community structure in the control and nicotinecontaminated sediment samples with nicotine concentration and time of exposure. The results demonstrated that the growth of some bacterial species in the nicotine-contaminated sediment samples was inhibited during the exposure. Some bacteria decreased in species diversity and in quantity with the increase of nicotine concentration or time of exposure, while other bacteria were enriched under the effect of nicotine, and their DGGE bands changed from undertones to deep colors. The microbial community structure, however, showed a wide variation in the nicotinecontaminated sediment samples, especially in the sediment samples treated with high-concentration nicotine. The Jaccard index was only 35.1%between the initial sediment sample and the sediment sample with a nicotine concentration of 0.030 mg/g after 28 d of exposure. Diversity indices showed that the contaminated groups had a similar trend over time. The diversity indices of contaminated groups all decreased in the first7 d after exposure, then increased until day 42. It has been found that nicotine decreased the diversity of the microbial community in the sediment.     

基于探索与利用平衡理论的灾变粒子群算法

基于算法只有适应优化问题的特性才能表现出优异性能的观点,在探索与利用平衡的理论框架下将灾变机制引入粒子群算法。在对灾变的强度和范围进行深入研究的基础上,提出4种控制灾变的方法,并通过多组正交实验研究最佳的灾变触发方式。通过实验分析得出如下结论：灾变对高维问题的作用有限;灾变强度控制在15%以下为宜;以种群多样性作为灾变的触发条件,能得到较好效果。以上述结论为基础提出自适应灾变粒子群算法,并通过与其他算法对比验证文中算法具有较好性能。