Sankar Vema - Blog Space

Speaking to Recode’s Kara Swisher and MSNBC’s Ari Melber, Pichai said AI is “one of the most important things that humanity is working on. It’s more profound than, I don’t know, electricity or fire,” adding that people learned to harness fire for the benefits of humanity, but also needed to overcome its downsides, too. Pichai also said that AI could be used to help solve climate change issues, or to cure cancer. We are seeing some exciting things in the industry, Samsung’s massive 8K TVs apparently use AI to upscale lower resolution images for the big screen. Sony has created a new version of the Aibo robot dog, which this time promises more artificial intelligence. Travelmate’s robot suitcase will use AI to drive around and follow its owner wherever they go.  Kohler has invented Numi, a toilet that has Amazon’s Alexa voice assistant built in etc., But despite all this, it does leave me wondering: is artificial intelligence really what we should be calling this revolution?...

We all know AI is coming to the forefront of technology in the form of machine learning (ML) and deep neural networks, where AI can understand the surroundings and apply that information to the current context, whether it is about data pattern analysis or driving a vehicle. Coupled with increasing capabilities of AI hardware, in 2018 we can expect: AI research will reach new heights in the fields of neural networks and generative adversarial networks. Cooperative learning among networks will lead to hybrid intelligence. Continuous evolution of AI will empower enterprise-scale business processes with intelligence and agility. AI will redefine the IT landscape with smart machines and borg systems. Read the complete article on DXC technology trends for 2018

Since it is open sourced, Google TensorFlow is becoming a more and more popular choice for AI enthusiast and researchers. With its recent announcement of DeepMind using TensorFlow for all its future research activities Google is reiterating its confidence on TensorFlow. Other google AI research projects using TensorFlow includes DeepDream, Google's visual AI that could transform photos into psychedelic art Magenta - Google's quest to find "Can machines make music and art? If so, how? If not, why not?". To be officially launched on June 1st  References: Google's Magenta project

Abstract: Application of Machine learning algorithms to some of the real-world problems pertaining to areas, like fraud/intrusion detection, medical diagnosis/monitoring, bio-informatics, text categorization and et al. where data set are not approximately equally distributed suffer from the perspective of reduced performance. The imbalances in class distribution often causes machine learning algorithms to perform poorly on the minority class. The cost minority class mis-classification is often unknown at learning time and can be far too high. A number of technique in data sampling, predominantly over-sampling and under-sampling, are proposed to address issues related to imbalanced data without discussing exactly how or why such methods work or what underlying issues they address. This paper tries to highlight some of the key challenges related to classification of imbalanced data while applying standard classification technique. This discusses some of the prevalent methods related ...

Abstract: Motivation is a very complex psychological behavior arising out of ones current physiological and psychological state of affairs. Motivation in humans is always associated or studied with incentive theories. As per human psychology, our intrinsic motivation factors are centered around intrinsic rewards which are considered critical for the development of cognitive intelligence. In that case, can an artificially learning machine be motivated to develop cognitive intelligence? What are the factors that would lead a machine learning system to motivate itself intrinsically? This paper discusses some of these question based on the latest research work carried out in the fields of development psychology, active learning, neuroscience, adaptive curiosity et. al., and see how this can be applied to our context of developing intrinsically motivated systems. Awaiting session recording. Will post it soon.

Abstract: Increase in dimensions of the data decrease the performance of the machine learning systems as the increase in the dimensions increase the problem space under analysis make data sparse. As the efficiency of the machine learning algorithms directly relates to the volume of the test data, increased space demands more data for better learning opportunities. To address this challenge, most of the time we tend to reduce some of the data dimensions searching for dimensions which are not directly related to the problem under analysis. For efficient reduction of dimensions we need to address the question "what is the idea dimensionality we can address without compromising on the sensitive of the dimensions?" This paper outline the problem of dimensionality not just from the angle of issues with high dimensional data leading to the reduction of dimensions but analyses how to efficiency balance the dimensions through better data projection techniques for more accurate resu...

Abstract: Distributed Denial of Service (DDoS) attacks are one of the major challenges to Internet community. Attackers send legitimate packets with often changing information from various compromised systems at random and at a very high frequency, rendering the target non-responsive for normal traffic. DDoS attacks are difficult to detect with traditional detection methods and standard Intrusion Detection Systems (IDS). Standard IDS tries to analyze the network traffic or system logs trying to identify emerging patterns on the network traffic. But due to randomness of the package origins it is difficult segregate true, false positive and normal traffic. This paper proposes a model based on Artificial Neural Networks to identify anomalies and detect DDoS patterns. In the proposed system sets of known characteristic features, which can separate attacks from normal traffic, are fed to the system to train the Artificial Neural Networks (ANN). This self learn system improves with each n...

As discussed in my previous blog, Imbalanced data poses serious challenges in Machine Learning .  One of approach to combat this imbalance is data is to alter the training set in such a way as to create a more balanced class distribution so that the resulting sampled data set can be used with traditional data-mining algorithms. This can be achieved through...  Under-sample where the size of the majority class is reduced using different techniques like reducing redundancy, removing boundary candidates etc., Over-sample where the size of the minority class is increased by adding more candidates which can augment the data set. Hybrid approach where a combination of both oversampling of minority class and under sampling of majority class is attempted. Each of these techniques discussed below Random Over Sampling In random over-sampling, the minority class instances are duplicated in the data set until a more balanced distribution is reached. As a illustration, consider...

For many real world machine learning problem we see an imbalance in the data where one class under represented in relative to others. This leads to mis-classification of elements between classes. The cost of mis-classification is often unknown at learning time and can be far too high. We often see this type of imbalanced classification scenarios in fraud/intrusion detection, medical diagnosis/monitoring, bio-informatics, text categorization and et al. To better understand the problem, consider the “Mammography Data Set,” a collection of images acquired from a series of mammography examinations performed on a set of distinct patients. For such a data set, the natural classes that arise are “Positive” or “Negative” for an image representative of a “cancerous” or “healthy” patient, respectively. From experience, one would expect the number of noncancerous patients to exceed greatly the number of cancerous patients; indeed, this data set contains 10,923 “Negative” (majority class) and...