In the first of this two-part series on Cluster Analysis, our blogger discussed how Cluster Analysis can help tackle issues of cyber-attacks and malware better and more. In this article, we will analyze patent filing trend and identify segments where cluster analysis is applicable. With data… (Featured image source is for representational purpose alone and has been sourced from https://pixabay.com/en/artificial-intelligence-507813/)
In the first of this two-part series on Cluster Analysis, our blogger discussed how Cluster Analysis can help tackle issues of cyber-attacks and malware better and more. In this article, we will analyze patent filing trend and identify segments where cluster analysis is applicable …
With data mining becoming more and more important in many research areas, cluster analysis has played an important role in various application domains. In the UK, cluster analysis has been used by Police Department to predict crimes in a particular street or region of a city by using publically available crime-related data of past few years.
Cybersecurity and crime pattern detection: Clustering brings a dynamic edge in the detection of malicious content such as URLs and spam emails. It is also being used to detect frauds by auditors and forensic accountants. A recent study suggests prediction of riots by clustering twitter data. These advancements are also a direct result of increased IP activity in text and image analytics in recent years.
Deep Learning: Deep learning uses artificial neural networks having multiple hidden layers for feature extraction. In recent times, there has been extensive interest in using deep neural networks for clustering. This also justifies the increased overlap of cluster analysis and knowledge-based systems in patent filings post 2015. Attempts are also being made to combine fuzzy logic and deep learning for creating better-classifying systems.
Language Understanding: Great amount of work is being done currently in natural language processing. A recent report by IBM suggests that half of the CEOS plan to adapt cognitive computing by 2019. Cognitive computers will be able to understand the reason, learn, and interact with humans in natural language. Google is showing promise as a frontrunner in NLP technology with its Neural Machine Translation System and contextual approach to solve NLP tasks using clustering and classification techniques. Four major European institutions developed a new machine translation engine which they made available publically on Github. Research advancement has led to a surge in patent filings in recent years.
Leaders in Cluster Analysis
Google, Microsoft, Facebook, and Apple are leading R&D and acquisitions in machine learning and AI. In 2016, Amazon, Facebook, Google, Microsoft and IBM joined hands to create “Partnership on AI”. IBM, Google and Microsoft figure among the top companies patenting in this space. IBM and Google also appeared among the most innovative companies in machine learning in 2017.
- IBM: IBM’s Watson computer has remained in news since its appearance on ‘Jeopardy!’ in 2011. Currently, Watson is being leveraged in a number of areas such as healthcare, coding, education, weather forecast, and finance. IBM has also acquired startups like Explorys and Alchemy API to further increase its prowess in the field.
- Google: In 2014, Google acquired DeepMind for $400 million, making it one of the biggest acquisitions in the machine learning domain. In 2015, Google released TensorFlow, an open source software library for machine learning tasks including clustering.
- Microsoft: Microsoft provides machine learning capabilities through its Cortana Intelligence suite. It has also released a machine learning library for Apache Spark. Microsoft’s Machine Learning Library (MLlib) provides tools such as algorithms to offer classification, regression, clustering, and filtering. Microsoft’s Project Oxford (2015) aims to understand its users with face, emotion and speech application program interfaces (APIs). Microsoft is also investing in various machine learning startups through Microsoft Ventures.
Machine learning is growing exponentially, yet there is a long journey ahead to develop what can be called complete artificially intelligence. Advances in cluster analysis will result in avoiding repetitive tasks, with lesser scope for errors. It will also provide valuable data-based insights. The computation power required for dealing with a vast amount of data may be solved by the amalgamation of quantum computing with machine learning. In the near future, we can anticipate greater personalization, more collaboration, better security, sounder health and machine assistance in everyday life.
(Featured image source is for representational purpose alone and has been sourced from https://pixabay.com/en/artificial-intelligence-507813/)