Friday, 15 January 2016

Cybersecurity Risk to Machine Learning Algorithms

Cybersecurity is a very important and is becoming one of the biggest worries for companies. According to various surveys, companies are investing a lot of money in cyber security and training their employees in it. It is estimated that till 2020 investment in the cybersecurity will be around $170 billion. In today’s world as the data is rising, so the risks on it are also increasing. The pattern classification systems that machine-learning algorithm rely on themselves exhibit vulnerabilities that can be exploited by hackers. As we know machine learning algorithms train themselves with the training data set so it may be manipulated according to the needs as hacker wants. For example, Search-engine-optimization algorithm was trained and manipulate website content to boost results in the search ranking or senders of junk emails try to fool spam-filtering algorithm. Even the results of the public election can also be affected by 20% or more as it is found that the order in which candidates appear in search results can have significant impact on perception. To read more about Cybersecurity risks, follow the article by Dr. Kira Radinsky (CTO and Co-founder of SalesPredict) at:

No comments:

Post a Comment