Mergers & Acquisitions


S&P Global Makes $550 Million Bet on AI with Kensho Buy - WatersTechnology.com

#artificialintelligence

The information giant is paying that amount in cash and stock to pick up Kensho Technologies, a fintech startup backed by names including Goldman Sachs, which is widely regarded as one of the most promising new breed of technology firms. Kensho, formed in 2013 by Daniel Nadler, a Harvard PhD graduate, and which counts technology specialists from Google, Apple and others among its ranks, specializes in AI for finance and national security.


Interest of global tech giants including Apple, Intel revives in Indian startup space - The Economic Times

#artificialintelligence

BENGALURU HYDERABAD: Apple's recent acquisition of Indian machine-learning startup Tuplejump offers further evidence of a revival in the interest of global technology giants in the country's startup space, especially in areas such as artificial intelligence, cloud infrastructure and automation. Enthusiasm for tech startups in India had waned in the last two years with fewer exits, a funding crunch and an inability to scale up. That seems to be changing with Intel, Apple and Nutanix shopping around for companies and the people who work there. Intel bought Soft Machines, a Silicon Valley chip designer with offices in Hyderabad, for 300 million in September. The company was cofounded by former Intel veteran Mahesh Lingareddy.


The Merger of Telecom and Artificial Intelligence

#artificialintelligence

The definition of artificial intelligence (AI) is a bit fuzzy, so when AT&T claims that it has been using AI for more than 20 years, it should be kept in mind that the AI of 1996 is a lot different from the AI of 2016. The bigger point is that the newer version is doing a lot for the carrier. AT&T is settling on an AI platform that can be used for different things instead of developing "one-off" solutions every time a task requiring the predictive capabilities and massive number-crunching abilities of AI presents itself. AI can be leveraged to anticipate rather than simply react to events, as less sophisticated AI platforms have done in the past. The driver is software-defined networks (SDNs), according to Computerworld.


M&A Activity In Artificial Intelligence Up 7x Since 2011

#artificialintelligence

As we've mentioned, corporate giants in tech are competing to acquire private artificial intelligence startups, with Google emerging as the most active acquirer over the last 5 years. So far in 2016, there have already been over 20 acquisitions of AI startups, the most recent being Twitter's acquisition of visual-processing startup Magic Pony Technology. We analyzed M&A and IPO activity in this space from 2011. Our artificial intelligence category includes companies developing core AI algorithms as well as those applying AI solutions to specific industries like healthcare and cybersecurity. We used our database to analyze at what stage the AI companies exited.


Efficiently Merging Symbolic Rules into Integrated Rules

AAAI Conferences

Neurules are a type of neuro-symbolic rules integrating neurocomputing and production rules. Each neurule is represented as an adaline unit. Neurules exhibit characteristics such as modularity, naturalness and ability to perform interactive and integrated inferences. One way of producing a neurule base is through conversion of an existing symbolic rule base yielding an equivalent but more compact rule base. The conversion process merges symbolic rules having the same conclusion into one or more neurules. Due to the inability of the adaline unit to handle inseparability, more than one neurule for each conclusion may be produced. In this paper, we define criteria concerning the ability or inability to convert a rule set into a single neurule. Definition of criteria determining whether a set of symbolic rules can (or cannot) be converted into a single, equivalent but more compact rule is of general representational interest. With application of such criteria, the conversion process of symbolic rules into neurules becomes more time- and space-efficient by omitting useless trainings. Experimental results are promising.