Extractive summarization is one of the vital tasks in text analysis and natural language processing. Although Hindi is one of the world’s highly speaking languages and produces thousands of online documents daily, most existing text summa- rization works focus on the English language. A Neural network- based summarizer is popular for abstractive summarization but has not been explored for extractive one except in a few recent studies. The present work uses a neural extractive summarizing model to develop a Hindi language extractive summarizer. The main contribution of the paper is two-fold. First, we generated a new Hindi-based text summarization data set from a popular Hindi news channel AajTak. The code to generate the data set is available at https://tinyurl.com/sonaa-hindi-text. Then we use this data set to train a Neural Extractive Summarization model. The model also learns the word embeddings while learning itself. The ROUGE-2-F1 and ROUGE-1-F1 results on test data show promising output with a score of 20.02 and 39.81, respectively.
In this modern era of technologies of scale, vast amounts of data are generated both by users and machines every day. This data comes as streams that may contain outliers. Detecting those outliers can be helpful in many ways, such as machine failures due to overload. Similarly, trends in social media posts are also outliers, and detecting them at different levels has great benefits. The current paper proposes an algorithm to approximate median and median absolute deviation from a stream of numerical values. The algorithm takes a fixed number of memory spaces and linear to the size of the memory. The median and median absolute deviation are then used to detect outliers and multi-level trends without being prone to noise in the data. Experimental results with CPU usage benchmark data and Twitter post data show the effectiveness of the proposed algorithms.
The talk introduced techniques developed by us for social network analysis using soft computing technique. The fuzzy set is used for handling uncertainities arises from the concepts itself and rough set is used to handle the uncertainities due to the granularity of the data.
The talk describs why the network science is important to solve different aspects of complex systems. It also provide the brief about the different research problems in the domain. Further it provide some of the solutions we developed at our lab and research group.
The talk summarizes different social media analytics techniques that can be useful for security services. For example, text analytics, sentiment analytics, topic molding, etc. In addition to that the talk shares recent developments that are being pursued in the Cognitive and Social Computing laboratory of the Department of Computer Science and Engineering at the Indian Institute of Technology Jodhpur.