UPLB ICS Peak One

This study presents and implements a Convolutional Neural Network(CNN) that will be used to classify steganographic materials. Image is one of the most used form of data in the current technological generation and in this study, the goal is to classify JPEG images whether these images contain suspicious data. Training the CNN involves two separate image datasets. The first dataset contains 6000 images and uses F5, JSteg and Outguess from Salgado’s work and the second set of images uses BOSS(Break Our Steganograpy System) dataset which contains HUGO, UNIWARD, and WOW. Results show that the images that use F5, JSteg, and Outguess can only hide relatively small amount of data which makes the CNN more confused and thus reduce the accuracy. The same model was applied to the BOSS dataset and shows that the model is not capable of beating the accuracy obtained by Couchot et al.’s model.

Open Shortest Path First module was implemented in Scalable Cisco IOS Simulator for Virtual networks to help the students configure networks using this routing protocol. OSPF is used to calculate the best path to be taken by the packets passing through networks.

This paper presents a software library for emulating the Posit number format written in the Python Programming lan- guage. The posit number format was proposed by John Gustafson in February 2017 and intends to be a drop-in replacement for the IEEE 754 Floating Point [1]. The software implementation implements basic arithmetic operations and transcendental func- tions. Along with the Posit, the library also supports the Quire accumulator as a large fixed-point storage that defers rounding. The Quire is leveraged in the implementation of transcendental function to achieve more accurate results. Similar to the Posit, the Quire is accessible to the programmer as a class. The research was able to build and publish the library at the Python Package Index with appropriate unit testing scripts. This paper describes the design and techniques used throughout the development of the library. Finally, this paper evaluates the resulting implementation by providing demonstrations of its many applications.

Most scientific applications require high perfor- mance computing (HPC) which utilizes parallel processing to run tasks quickly and efficiently. MPI clusters are often used to cater this type of tasks but the hardware required can be costly. Peak-Two Cloud (P2C) can host HPC applications in a cloud environment which is relatively cheaper and more convenient. One of its features is vCluster, a tool that can deploy MPI clusters on demand [1]. We have developed SkyLab, a web interface for vCluster. It aims to simplify the process of running MPI tasks from the perspective of the end-user. Here, we describe the system design and features of SkyLab.

With XMPPs real-time capabilities and the avail- ability, SIP’s multimedia features for communication and porta- bility of smartphones, they make good tools for communication. The researcher incorporate XMPP and SIP into Android device to create an intranet Instant Messaging application were users can chat and make free audio calls within networks.The Univer- sity of the Philippines Los Ba ̃nos has a wide network extending throughout the campus. The study is only set up in a small network using test servers but can potentially be applied to larger networks like UPLBs where it can be more useful.