Hybrid Warfare and its Impacts on Pakistan

In modern times, the conventional means of warfare are increasingly becoming less usable. However, the states are involved in waging hybrid warfare to the maximum to fulfill their foreign policy goals. In nuclearized South Asia, direct war between India and Pakistan seems unlikely given that both the states know that escalation could lead to nuclear catastrophe in the region. This compels both the states to find other means of warfare to undermine each other’s interests. India wants to weaken Pakistan so that it may abandon claim on Indian occupied Jammu and Kashmir. For that, India is using all tools of hybrid warfare against Pakistan. In this context, this paper aims at to unearth India’s hybrid warfare in the region and its implications for Pakistan. The main focus of the paper is to explain tools and methods of India hybrid warfare. At the same time the research also tries to unravel few other case studies. It also notes how Pakistan can counter hybrid threats posed by its arch rival.

Image Based Recommendation System for Textile

The project was to develop an online portal for shoppers that provides recom- mendations based on similarities in the design patterns, taking a step beyond the usual searching on price, material, color etc. The process started with gathering data from multiple online resources by scraping the websites of dierent online stores within the city. Image Processing was the next crucial part as texture features had to be extracted from the images of printed clothes to recognize their patterns. Machine Learning came in the form of unsupervised learning where clustering techniques were explored on a sample data set. The best technique was chosen based on highest quality of clusters formed. K-means, Self Organizing Maps, and Adaptive Resonance Theory were the clustering algorithms applied on the image features extracted using Haralick fea- tures extraction technique and also on features from Auto encoders. The results from Haralick features undergone through Self Organising Maps clustering were used to nally generate suggestions for similar items on the website