Enhancement of Online Stochastic Gradient Descent using Backward Queried Images
Stochastic gradient descent (SGD) is one of the preferred online optimization algorithms. However, one of its major drawbacks is its predisposition to forgetting previous data when optimizing through a data stream, also known as catastrophic interference. In this project, we attempt to mitigate this drawback by proposing a new low-cost approach which incorporates backward queried images with SGD during online training. Under this new approach, we propose that for every new training sample through the data stream, the neural network is optimized using the corresponding backward queried image from the initial dataset. After compiling the accuracy of the proposed method and SGD under a data-stream of 50,000 training cases with 10,000 test cases and comparing our algorithm to SGD, we see substantial improvements in the performance of the neural network with two different MNIST datasets (Fashion and Kuzushiji), classifying the MNIST datasets at a high accuracy for the mean, minimum, lower quartile, median, and upper quartile, while maintaining lower standard deviation in performance, demonstrating that our proposed algorithm can be a potential alternative to online SGD.
The Waves Fish Controller
Our oceans, coasts, and estuaries are home to diverse living things. These organisms take many forms, from the tiniest single-celled plankton to the largest animal on Earth, the blue whale. Understanding the life cycles, habits, habitats, and inter-relationships of marine life contributes to our understanding of the planet as a whole. Human influence and reliance on these species, as well as changing environmental conditions, will determine the future health of these marine inhabitants Humans influence the whole environment even if they don’t notice , the growth of men and our increasing reproduction over the years results to an over consumption of nutritious products , which makes us exploit the wildlife more and more and in the same time take parts of its habitats for us to life in and throwing our non-needed materials in what’s left of the world. And that’s a big problem because the Eco-System was just fine before we started over exploiting it in a greedy and unreasonable manner, and since the ecosystem’s parts are related altogether in an ongoing circle , the absence or the destruction of one part of It may lead to the unbalance and even destruction of the whole organized system. And that’s why as humans, it is our first duty to take care of nature generally and both fauna and flora specifically, not because of a moral code of some kind; but to protect Humanity from ourselves, and to preserve the human kind from destruction and extinction. And that’s the main goal of our project, that’s to help us organize our fishing exploitative activities with how much can the environment handle from it.
Territorial behaviour of the Eurasian Wren (Troglodytes troglodytes) during autumn migration and wintering in the urban environment of Hradec Králové, Czech Republic
In the present paper I studied the winter territories of the Eurasian Wren (Troglodytes troglodytes) in the urban environment of the city of Hradec Králové. The males of this species were detected through the aggressive reactions to the playback of its conspecific call. The Eurasian Wren territories were detected at seven out of the ten observation points in the city. The frequency of the territories decreased during the observation period from October to January and it was also affected by the structure of the biotope. It seemed that the wrens preferred trees and avoided lawns and isolated growths. That illustrates how important it is to maintain wildlife corridors in the cities. Individuals of seven other bird species were seen reacting to the playback of Eurasian Wren songs or warning calls. Eurasian Wren may be applied as an indicator species of the ecological value of the urban environment. Through this species, we can access such quality also for the human inhabitants.