Relationship Between BP (Baculovirus penaei) Energy Reserves in Larval and Postlarval Pacific White Shrimp Penaeus vannamei

The relationship between energy reserves of the penaeid shrimp Penaeus vannamei and Baculovirus penaei, or BP, were investigated in a series of experiments using mysis stage or early postlarval shrimp. Pre-exposure and post-exposure levels of protein and triacylgycerol (TAG) were determined. The effect of pre-exposure protein and TAG levels on susceptibility to BP infections was also investigated by starving a group of shrimp immediately prior to BP exposure. There was no consistent relationship between either pre-exposure or post-exposure protein levels and the percent of shrimp developing patent BP infections. There was, however, a significant positive correlation between TAG levels immediately prior to viral exposure and prevalence of infection 72 h later. Experimental reduction of TAG reserves prior to BP exposure delayed the development of a patent infection. In some, but not all, experiments there was a significant reduction in TAG levels of infected compared with uninfected shrimp 72 h post-exposure. The effect of patent BP infections on host TAG levels was subordinate to fluctuations in TAG content associated with the ontogeny of the hepatopancreas. Results of this study support histological observations that shrimp lipid levels can be altered by baculovirus infections. Furthermore, high levels of energy reserves in the form of TAG are associated with increased susceptibility to BP infection in larval and postlarval shrimp.

Balancing Exploration and Exploitation: A New Algorithm for Active Machine Learning

Active machine learning algorithms are used when large numbers of unlabeled examples are available and getting labels for them is costly (e.g. requiring consulting a human expert). Many conventional active learning algorithms focus on refining the decision boundary, at the expense of exploring new regions that the current hypothesis misclassifies. We propose a new active learning algorithm that balances such exploration with refining of the decision boundary by dynamically adjusting the probability to explore at each step. Our experimental results demonstrate improved performance on data sets that require extensive exploration while remaining competitive on data sets that do not. Our algorithm also shows significant tolerance of noise.