Earthquake disaster avoidance learning system using deep learning

The popularity of deep learning has influenced the field of surveillance and human safety. We adopt the advantages of deep learning techniques to recognize potentially harmful objects inside living rooms, offices, and dining rooms during earthquakes. In this study, we propose an educational system to teach earthquake risks using indoor object recognition based on deep learning algorithms. The system is based on the You Look Only Once (YOLO) deployed on our cloud-based server named Earthquake Situation Learning System (ESLS) for the detection of harmful objects associated with risk tags. ESLS is trained on our own indoor images dataset. The user interacts with the ESLS server through video or image files, and the object detection algorithm using YOLO recognizes the indoor objects with associated risk tags. Results show that the service time of ESLS is low enough to serve it to users in 0.8 s on average, including processing and communication times. Furthermore, the accuracy of the harmful object detection is 96% in the general indoor lighting situation. The results show that the proposed ESLS is applicable to real service for teaching the earthquake disaster avoidance.     

Stimulus specificity in avoidance learning

The present experiments address the issue of stimulus specificity in fear and avoidance learning. First, it was established that light stimuli are effective warning signals (WS) in shuttle avoidance. Then light stimuli were shown to produce conditioned suppression in the conditioned emotional response situation comparable to that produced by noise conditioned stimuli. Finally, the effectiveness of noise onset and noise offset as feedback signals was tested. This was assessed under conditions of immediate and delayed termination of a light WS. Delayed termination of a light WS interfered with avoidance learning and the introduction of noise offset as a feedback signal enhanced it. The only demonstration of stimulus specificity was the failure of noise onset to function as a feedback signal.     

Species-specific defense reactions and avoidance learning

Bolles (1970) proposed a theory of avoidance learning, the species-specific defense reaction (SSDR) hypothesis, which emphasized innate constraints on the response repertoire of rats in aversive situations and minimized the role of reinforcement in avoidance learning. The present paper describes Bolles’ (1970, 1971, 1972, 1975, 1978) development of SSDR theory and reviews the empirical tests of its assumptions and predictions. It is concluded that the SSDRs described by Bolles, along with some others, are highly probable in aversive situations but that the response repertoire is not limited to them. Further, there is strong evidence for reinforcement effects in the establishment and maintenance of at least some avoidance responses.     

Myopic regret avoidance: Feedback avoidance and learning in repeated decision making

Decision makers can become trapped by myopic regret avoidance in which rejecting feedback to avoid short-term outcome regret (regret associated with counterfactual outcome comparisons) leads to reduced learning and greater long-term regret over continuing poor decisions. In a series of laboratory experiments involving repeated choices among uncertain monetary prospects, participants primed with outcome regret tended to decline feedback, learned the task slowly or not at all, and performed poorly. This pattern was reversed when decision makers were primed with self-blame regret (regret over an unjustified decision). Further, in a final experiment in which task learning was unnecessary, feedback was more often rejected in the self-blame regret condition than in the outcome regret condition. We discuss the findings in terms of a distinction between two regret components, one associated with outcome evaluation, the other with the justifiability of the decision process used in making the choice.     

Successive positive contrast in one-way avoidance learning

The main finding of these experiments was a positive contrast effect in one-way avoidance learning. Experiment 1 showed that increasing safety time during one-way avoidance training led to improved performance, surpassing that of a control group that had received the high reward (safe time) from the beginning of training. Experiment 2 showed that a similar positive contrast effect occurred when the time spent in the danger compartment before the onset of the warning signal was shortened. These results suggest that time spent in a safe context acts as a reinforcer of the avoidance response; however, its incentive value depends not only on its duration, but also on the length of the time spent in the danger compartment before the onset of the signal. Overall, results also suggest that the avoidance response is a mixture of flight (motivated by fear) and approach (to a safe place) behaviour. The specific weight of the flight or approach component may be a function of the time and the amount of activation of each emotional state (fear or relief) due to opponent homeostatic compensatory processes that occur in the danger and safe compartments during one-way avoidance learning.     

Aposematic colouration and avoidance learning in chicks

Two groups of male chicks (Gallus gallus) were allowed to peck at distasteful food (chick crumbs flavoured with quinine and mustard) that either matched or contrasted with the background on which the food was presented. The experiment consisted of a series of trials, each of which was terminated when a fixed amount of food (four crumbs) had been ingested. Within-trial pecking rate was initially higher in the group given constrasting food, confirming that degree of contrast between a prey item and its background increases the likelihood of detection by a naive predator. However, during the series of training trials pecking rate declined faster, and reached a lower level, in the group given contrasting food than in the group given matching food. Thus, rate and strength of avoidance learning are affected by the degree to which distasteful food contrasts with its background. This may explain why many distasteful prey species are brightly coloured, despite the fact that bright colouration increases the likelihood of initial detection by a predator.     

Dual effects of response blocking following avoidance learning

It was shown that either enhanced or decreased avoidance responding by rats could be produced by the same response blocking procedure occurring after avoidance training. The particular result produced depended on (1) the duration of the blocking trial and (2) when during the retention interval the blocking trial occurred. Results were interpreted in terms of a memory reactivation explanation and implications for the therapeutic uses of extinction were discussed.