PORTLAND, Ore. -- First IBM's Deep Blue computer beat the world champion Garry Kasparov at chess in 1997. Then in 2008 the University of California at San Diego (UCSD) authored an open source poker bot -- Fell Omen -- that could beat humans. Then Apple's Siri dazzled the world in 2010 by answering natural language questions on the iPhone. The very next year, 2011, IBM pitted its Watson supercomputer (since moved to the cloud) against former winners of the game show Jeopardy. Watson beat them and took the first-place prize of $1 million.
As a result of all this publicity, capital is flowing into "deep knowledge" startups like Vicarious FPC Inc. of San Francisco, which just last year claimed to have conjured the first artificial intelligence (AI) to pass the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPCHA).
Now a startup with $143 million in funding says, "You ain't seen nothin' yet," describing a sentient distributed artificial intelligence that sounds like a nice-guy version of Skynet from the cinema flick Terminator. According to the technology gurus at Sentient Technologies Holdings Ltd. of San Francisco, the software for sentient computers, which they are already installing at key customer sites, goes beyond natural language recognition, unstructured searching, machine learning, and deep knowledge.
One step beyond neural networks: Neural networks are good at learning to recognize particular patterns. The next step consists of evolutionary algorithms (EA), which use mutation to create pools of possible solutions on different servers, each of which ranks its pool of possible solutions. An evolutionary coordinator then compares the solutions and sends the best of the crop for validation against big data sets it has never seen.
"Reasoning and logic are one thing, but beyond that is true intelligence -- what we call sentience," Babak Hojat, cofounder and chief scientist at Sentient Technologies," tells EE Times.
"Sentience is being aware, having perceptions, being mindful, and has implications of autonomy," chief technology officer Nigel Duffy said in a video (below) that you really ought to watch to get the big picture.
Sentient Technologies is aiming for a globally distributed artificial intelligence the world has never seen. (Source: Sentient)
According to Hojat, excepting Horizons Ventures, Sentient Technologies has shunned the usual venture capital route to glory by instead enlisting support from its actual users in the fields of finance, consumer, food/beverage, and real estate for the last seven years, quietly showing demonstrations of how its distributed artificial intelligence, crafted in collaborative research with the Massachusetts Institute of Technology (MIT) and other partners, works for medical research, financial trading, and other big data applications. However, its latest Series C round of $103.5 million -- for marketing its distributed AIs -- included strategic investor Access Industries, with credits for backing Warner Brothers, Spotify, and Tata Communications.
Sentient has also addede two new advisors to bolster their drive to market Sentient globally-- Nello Cristianini, a professsor of AI at the University of Bristol, reported by Thomson Reuters as one of the "most influential scientists of the decade" for his books on machine learning.
To help with direct marketing, advisor Michael Scharff has two decades of retail experiene in e-commerce with well-known retailers such as Toys R Us, Staples and Best Buy.
Sentient just had one of its pending patents granted--the newly granted patent protects the unique way Sentient distributes its Evolutionary Algorithms (EA) among computers on a grid, so that Big Data does not have to be in one place at the same time.
Thank you so much for your insightful analysis. I searched for "we" and the only instances of "we" I can find in the story are quotes, where we refers to Sentient itself. Regarding the dangers of AI, as predicted by Stephen Hawkings, Sentient says they are a long, long way from their AI becoming dangerously autonomous. Hopefully the Computer Professionals for Social Responsibility will have plenty of time deal with these issues. Here I am only reporting on how the software works. Perhaps in a future article these issues can be discussed (if Sentient even becomes a success--the vast majority of startups fail). Thanks again for you excellent comment.
The Sentient story deserves to be discussed in more sophisticated ways.
For example, when the word "we" is invoked, to whom does the author refer -- a relative handful of mostly male, mostly North Americans? What ethical principle would justify allowing them to change the conditions of life for the rest of humanity?
If democracy is widely preferred over alternatives in the political sphere, why do so few consider its advantages (and disadvantages) in the technological-economic sphere?
Given that historians of technology repeatedly have shown that expert predictions are often wrong, why do so many respondents argue over their guesses about the unknowable facts instead of asking about the procedures for coping with being wrong?
Given that human and organizational learning tends to take a very long time, who is asking that the accelerator be used less and the brakes more -- thereby allowing time for patient trial and error, interpretation, and revised trajectories?
As Computer Professionals for Social Responsibility puts it, "Technology is Driving the Future, It is Up to Us to Do the Steering." I see almost no intelligent discussion -- here or elsewhere -- of what it would take in terms of political, economic, and cultural change to begin to approximate CPSR's goal.
Rick Merrit: What do they need the $143 million for? Is there a chip or data center they have to build?
ANSWER: They've alredy spent $40 million (in Series A and B) to develop their software suite, and say they are using the $103 Series C to market their software to the rest of the world (plus continue perfecting its algorithm). They are a software only company and install their code on the user's datacenters. One ofits investors, Tata Communications will be the preferred provider to Sentient customers without enough of their own datacenters, with its more than one million square feet of datacenters in 44 locations, serving 240 countries and territories. Eventually Sentient hopes to have a global network of users all of whom's data centers are connected by their software and who are willing to let others use idle time on their datacenters in return for the favor when they are running a job that could benefit from massive parallelism. So the answer is no and no--no proprietary chip and no datacenters of their own. They'll run their software on their investors datacenters, the vast majority of which are also their customers with the $130 million Series C to market their software to every other potential customer dealing with Big Data.
The stated goal of Sentient Technologies is to build a sentient computer. "Sentience is being aware, having perceptions, being mindful, and has implications of autonomy," chief technology officer Nigel Duffy said.
"Hojat says not to worry about the network "waking up" like Skynet".
Hmmmm....., my personal logic algorythms seem to have flagged a contradiction. Maybe he meant to say "waking up" yet? If he means it never will, then just what is the goal here?
How would he know if the network, or part of it "woke up", and became self-aware? Would it announce that to the world, so that we can install safeguards against unforeseen consequences, or keep that information to itself so we don't?
I've had some complaints about Sentient being called an MIT spinoff as well as a builder of Sentient computers. So let me set the record straight. Sentient had one big collaboration project with MIT so far and continues to plan future collaborations (mostly using MIT Big Data to make predictions using Sentient software) but no former MIT personel work at Sentient. MIT is not an investor in Sentient and Sentient does not license any IP from MIT (or anybody else for that matter--they have their own proprietary patents covering their IP). Regarding Sentient computers, Sentient Technologies Holding Ltd. is a software-only company. Of couse, their software has to run on a computer of some kind, but they do not sell any computers.