Damage Detection Method for Buildings with Machine-Learning Techniques Utilizing Images of Automobile Running Surveys Aftermath of the 2016 Kumamoto Earthquake

In order to understand the damage situation immediately after the occurrence of a disaster and to support disaster response, we developed a method to classify the degree of building damage in three stages with machine-learning using road-running survey images obtained immediately after the Kumamoto earthquake. Machine-learning involves a learning phase and a discrimination phase. As training data, we used images from a camera installed in the travel direction of an automobile, in which the degree of damage was visually categorized. In the learning phase, class separation is carried out by support vector machine (SVM) on a basis of a feature calculated from training patch images for each extracted damage category. In the discrimination phase, input images are provided with raster scan so that the class separation is carried out in units of the patch image. In this manner, learning, discrimination, and parameter tuning are repeated. By doing so, we developed a damage-discrimination model for each patch image and validated the discrimination accuracy using a cross-validation method. Furthermore, we developed a method using an optical flow for preventing double counting of damaged areas in cases where an identical building is captured in multiple photos.

Inhibitory and excitatory mechanisms in the human cingulate-cortex support reinforcement learning

The dorsal anterior cingulate cortex (dACC) is crucial for motivation, reward- and error-guided decision-making, yet its excitatory and inhibitory mechanisms remain poorly explored in humans. In particular, the balance between excitation and inhibition (E/I), demonstrated to play a role in animal studies, is difficult to measure in behaving humans. Here, we used magnetic-resonance-spectroscopy (1H-MRS) to examine these mechanisms during reinforcement learning with three different conditions: high cognitive load (uncertainty); probabilistic discrimination learning; and a control null-condition. Subjects learned to prefer the gain option in the discrimination phase and had no preference in the other conditions. We found increased GABA levels during the uncertainty condition, suggesting recruitment of inhibitory systems during high cognitive load when trying to learn. Further, higher GABA levels during the null (baseline) condition correlated with improved discrimination learning. Finally, excitatory and inhibitory levels were correlated during high cognitive load. The result suggests that availability of dACC inhibitory resources enables successful learning. Our approach establishes a novel way to examine the contribution of the balance between excitation and inhibition to learning and motivation in behaving humans.

Dear Virginia

Congratulations on your admission and fellowship for graduate work at ………. It should be the beginning of a very promising career and I'll be glad to give you whatever advice I can on the subject.In answer to your question, there is probably less overt discrimination against women now. Graduate admissions and fellowships look fairly good on the record. And although departments are still lagging in professorial positions and salaries for women, some of them are making an effort to catch up. It shouldn't be too long before most departments have their “quota” of assistant professors. The problem now is with the subtler forms of discrimination — “Phase II” of whatever it is we are involved in. (Phase II, like Phase I, uses sex as a basis for choice, but its arguments are more sophisticated. Phase I says “A woman's place is in the home.” Phase II says “We'd like to hire a woman, but …”)