In terms of plans to deploy AI commercially in the near future (before the end of 2018), healthcare is clearly the vertical sector in which the largest percentage of companies worldwide will take action, followed by the consumer content and apps industry (including gaming), manufacturing, retail and automotive.
Nearly 140 private companies working to advance artificial intelligence technologies have been acquired since 2011, with over 40 acquisitions taking place in 2016 alone (as of 10/7/2016). Corporate giants like Google, IBM, Yahoo, Intel, Apple and Salesforce, are competing in the race to acquire private AI companies, with Samsung emerging as a new entrant this month with its acquisition of startup Viv Labs, which is developing a Siri-like AI assistant.
Source: The Asimov Institute
With new neural network architectures popping up every now and then, it’s hard to keep track of them all. Knowing all the abbreviations being thrown around (DCIGN, BiLSTM, DCGAN, anyone?) can be a bit overwhelming at first.
So I decided to compose a cheat sheet containing many of those architectures. Most of these are neural networks, some are completely different beasts. Though all of these architectures are presented as novel and unique, when I drew the node structures… their underlying relations started to make more sense…
Composing a complete list is practically impossible, as new architectures are invented all the time. Even if published it can still be quite challenging to find them even if you’re looking for them, or sometimes you just overlook some. So while this list may provide you with some insights into the world of AI, please, by no means take this list for being comprehensive; especially if you read this post long after it was written.
For each of the architectures depicted in the picture, I wrote a very, very brief description. You may find some of these to be useful if you’re quite familiar with some architectures, but you aren’t familiar with a particular one.
IDC just released the results of its 3rd annual survey of IoT decision-makers (press release here and webcast with IDC Vernon Turner and Carrie MacGillivray here). The IoT market is maturing, says IDC, going beyond its initial focus on connecting more and more things. Data management is fast becoming the overarching theme, with Analytics and the IoT Platform emerging as the main requirements of the 31.4% of organizations surveyed that have already launched IoT solutions , and the additional 43% looking to deploy in the next 12 months.
The survey was conducted in July and August 2016, with 4,500 decision-makers from more than 25 countries participating, from enterprises with more than 100 employees in a wide range of industries. Here are the highlights:
Survey results show that IBM and Microsoft have taken a leading role in almost all IoT segments, especially the ones ascending in importance—analytics, software, systems integration and providing an IoT platform. This is due, IDC says, to their success in blending a cloud strategy with analytics and software capabilities. To IDC’s question about most important digital transformation projects, survey respondents cited cloud transformation/transition (66%), IoT (32%), and big data/cognitive solutions (27%). IDC noted that these transformational initiatives are interlinked: The cloud gives IoT a platform on which to scale and the IoT lays the foundation for investments in big data and cognitive solutions, to make sense of all data generated by the IoT and residing in the cloud.
Originally published on Forbes.com
Under pressure to remove alleged human bias from its “Trending Topics” section, in August Facebook fired the editors who were selecting and writing headlines for the stories, explaining that this “will make the product more automated.” The results of trusting algorithms more than humans have continued to make headlines ever since with the Trending “product” promoting a fake news story about Fox News’ Meghan Kelly, a conspiracy article claiming the 9/11 twin towers collapsed because of “controlled demolition,” and Apple’s Tim Cook announcing that Siri will physically come out of the phone and do all the household chores (a story from an Indian satirical website, Faking News, that was Trending’s top story on the day of the iPhone 7 launch event), to mention just a few of the more embarrassing machine failures.
Silicon Valley has never displayed much love for fallible humans, but has shown a lot of confidence in the continuous improvement and now, self-improvement, of machines. Do humans still have an important role to play in our automated lives which are increasingly controlled by sophisticated algorithms and seemingly smarter machines?
In Only Humans Need Apply: Winners and Losers in the Age of Smart Machines, knowledge work and analytics expert Tom Davenport and Julia Kirby, a contributing editor for the Harvard Business Review, offer optimistic, upbeat and practical answers to this much-debated question. “The upside potential of the advancing technology is the promise of augmentation—in which humans and computers combine their strengths to achieve more favorable outcomes than either could do alone,” they write.
There is not much difference, contend Davenport and Kirby, between technologies of automation and technologies of augmentation. The difference lies in the goals and attitudes behind the application of these technologies. Automation is unidirectional and focuses “primarily or exclusively on cost reduction” via the elimination of human labor. In contrast, “augmentation approaches tend to be more likely to achieve value and innovation” and they are bidirectional, making “humans more capable of what they are good at” and “machines even better at what they do.”
It is a shortsighted (and short-term) strategy for companies to favor automation over augmentation: “If the goal is to provide truly exceptional or differentiated products and services at scale, only an augmentation arrangement can accomplish that,” write Davenport and Kirby. They advocate a “workplace that combine sophisticated machines and humans in partnerships of mutual augmentation” and mutual benefit.
Competitive considerations apply not only to companies in the race against the machine, but also to their employees. The book addresses primarily the plight of knowledge workers who thought they would escape the fate of factory workers but are now increasingly automated out of a job. “The advice on avoiding that fate,” say Davenport and Kirby, “has been noticeably thin. For the most part, the experts boil it down to a single, daunting task: Keep getting smarter. We are going to argue that there are other strategies, all of them featuring augmentation of human work by machines.”
The authors describe in detail—with vivid and engaging examples—five “options for augmentation:” Stepping Up or moving a level above the machines and making high level decisions about augmentation; Stepping Aside or choosing to pursue a job that computers are not good at, such as selling or motivating; Stepping In or monitoring and improving the computer’s automated decisions; Stepping Narrowly or finding a specialty area in a specific profession that wouldn’t be economical to automate; and Stepping Forward or becoming involved in creating the very technology that supports intelligent decisions.
These strategies will work for knowledge workers (or all workers) who are “willing to work to add value to machines, and who are willing to have machines add value to them.” They will also work for organizations that understand that “no matter how smart these machines get, there is still some potential value from human augmentation.”
What a refreshing perspective in these times of machine-worship, where Silicon Valley’s automation addiction has spread far and wide. Mark Fields, the chief executive of Ford Motor Company, recently promised completely self-driving cars by about 2025, displaying a very Silicon Valley (and silly) attitude by saying “a driver is not going to be required.”
Fields and the many other executives of established companies racing against the disruptive Silicon Valley machine should read Only Humans Need Apply where Davenport and Kirby warn that companies investing in self-driving cars “could find that they have put a lot of energy into developing vehicles that drive themselves but are stuck with regulations that require an alert driver with hands on the steering wheel and feet on the pedals. If that happens, perhaps a company whose strategy all along has given careful thought to how to redeploy the human attention that is freed up by the technology—will win big.”
Just because you can automate, doesn’t mean you should. This is the important lesson of this contrarian, timely, and well-argued book. Augmentation, say Davenport and Kirby, is something “societies should encourage in ways big and small.” Hear! Hear!
Originally published on Forbes.com
Source: @Annu297
Vincent Müller and Nick Bostrom of FHI conducted a poll of four groups of AI experts in 2012-13. Combined, the median date by which they gave a 10% chance of human-level AI was 2022, and the median date by which they gave a 50% chance of human-level AI was 2040.
According to Bostrom, the participants were asked when they expect “human-level machine intelligence” to be developed, defined as “one that can carry out most human professions at least as well as a typical human”. The results were as follows. The groups surveyed are described below.
| Response rate | 10% | 50% | 90% | |
| PT-AI | 43% | 2023 | 2048 | 2080 |
| AGI | 65% | 2022 | 2040 | 2065 |
| EETN | 10% | 2020 | 2050 | 2093 |
| TOP100 | 29% | 2022 | 2040 | 2075 |
| Combined | 31% | 2022 | 2040 | 2075 |
Figure 1: Median dates for different confidence levels for human-level AI, given by different groups of surveyed experts (from Bostrom, 2014).
Surveyed groups:
PT-AI: Participants at the 2011 Philosophy and Theory of AI conference. By the list of speakers, this appears to have contained a fairly even mixture of philosophers, computer scientists and others (e.g. cognitive scientists). According to the paper, they tend to be interested in theory, to not do technical AI work, and to be skeptical of AI progress being easy.
AGI: Participants at the 2012 AGI-12 and AGI Impacts conferences. These people mostly do technical work.
EETN: Members of the Greek Association for Artificial Intelligence, which only accepts published AI researchers.
TOP100: The 100 top authors in artificial intelligence, by citation, in all years, according to Microsoft Academic Search in May 2013. These people mostly do technical AI work, and tend to be relatively old and based in the US.
Source: AI Impacts
To get a more accurate assessment of the opinion of leading researchers in the field, I turned to the Fellows of the American Association for Artificial Intelligence, a group of researchers who are recognized as having made significant, sustained contributions to the field.
In early March 2016, AAAI sent out an anonymous survey on my behalf, posing the following question to 193 fellows:
“In his book, Nick Bostrom has defined Superintelligence as ‘an intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.’ When do you think we will achieve Superintelligence?”
…In essence, according to 92.5 percent of the respondents, superintelligence is beyond the foreseeable horizon.
See also Oren Etzioni on Building Intelligent Machines
From Oren Etzioni’s presentation at the O’Reilly AI conference, September 2016:
As recently as 2013, the [deep learning] space saw fewer than 10 deals. But since January 2015, deep learning startups have raised over 70 equity deals in aggregate, and over $600M in equity funding since 2012…
Computer Vision: Startups here are using deep learning for image recognition, analytics, and classification. Aerial image analytics startup Terraloupe was seed-funded this year by Germany-based Bayern Kapital. New York-based Calrifai — backed by investors including Google Ventures, Lux Capital, and NVidia — entered the R/GA accelerator this year, after raising $10M in Series A in Q2’15. Captricity, which extracts information from hand-written data, has raised $49M in equity funding so far from investors including Social Capital, Accomplice, White Mountains Insurance Group, and New York Life Insurance Company.
Speech analytics/conversational interface: Google made news last year when it entered the Chinese market with its investment — a $60M Series C round — in Shanghai-based Mobvoi. The smart watch maker’s core tech includes speech recognition, text-to-speech conversion, and semantic analysis. Another startup that has recently been in the news is Viv Labs, which demonstrated its Siri-like AI assistant earlier this year. The company has raised $30M in equity funding from investors including Horizons Ventures and Pritzker Group Venture Capital.
Core AI: Startups here are developing algorithms that can be applied across multiple industries like finance, healthcare, and e-commerce. To name a few, Japan-based LeapMindraised $3.4M from Archetype Ventures, ITOCHU Technology Ventures, and Visionnaire Ventures in Q3’16; Teradeep, a neural network startup that “accelerates” deep learning via field programmable gate arrays (FPGAs) — integrated circuits that can be programmed for customer-specific applications — received funding in Q1’16 from the corporate venture arm of XILINX.
Auto & Robotics: 5 out of 6 startups in this category raised their first equity funding rounds this year. Machine vision company Netradyne, which has developed a driver-safety platform called Driveri, received $16M in Series A from Reliance Industries in India. Two other auto tech startups, Andreessen Horowitz-backed Comma.ai and Oriza Ventures-based Drive.ai, raised $3M and $12M in early-stage funding, respectively. China-based Turing Robot, which was initially focused on voice technologies, is now expanding into the consumer robotics market. It raised $7.6M in corporate minority from Alpha Animation & Culture. Another China-based startup, Rokid, which also has an office in the US, is developing a social robot. You can read more about robotics startups in China in our post here.
BI, Sales & CRM: Applications here include voice analytics to extract information from calls, automated customer response solutions, business data analytics, and sales targeting. To name a few, Palo Alto-based Mariana raised $2M in seed money from investors including Blumberg Capital; London-based True AI, previously seed funded by Entrepreneur First, entered the Microsoft Ventures Accelerator in Q3’16; another UK-based startup, Ripjar, raised funds from Winton Ventures in Q2’16.
Healthcare: As we discussed in our webinar recently, healthcare is the hottest area of investment compared to other industry-specific applications of artificial intelligence. Deep learning startups here include drug discovery platforms Insilico and Atomwise, IBM-backed precision medicine startup Pathway Genomics, and diagnostics companies Butterfly Networkand Enlitic.
Security: Israel-based Deep Instinct, which claims to be the first startup to bring deep learning to cybersecurity, is backed by investors including Blumberg Capital, UST Global, and U.S. News & World Report. Other startups here include Seattle-based SignalSense, which applies deep learning to IT security and smart camera startup Umbo CV.
E-Commerce: Deep learning in e-commerce was spotlighted recently by Etsy’s acquisition of Blackbird Technologies. Three startups in the private sector using AI in e-commerce raised funding rounds this year: Reflektion raised $18M in Q1’16 from investors including Intel Capital, Battery Ventures, and Marc Benioff; ViSenze raised $10.5M in Series B from investors including Rakuten Ventures, Enspire Capital, and Phillip Private Equity; India-based Staqu raised angel funds in Q2’16.
See also
Faster Artificial Intelligence: Baidu Benchmarks Hardware For Deep Learning
Deep Learning Is Still A No-Show In Gartner 2016 Hype Cycle For Emerging Technologies
AI And Machine Learning Take Center Stage At Intel Analytics Summit
Things are looking up for the Internet of Things. 80% of organizations have a more positive view of IoT today compared to a year ago, according to a survey of 512 IT and business executives by CompTIA. “This reflects greater levels of attention from the C-suite and a better understanding of how the many different elements of the IoT ecosystem are starting to come together,” says CompTIA. Here are the highlights from this and other recent surveys:
How big is the IoT and how fast is it growing? The number of connected things, from computers to household monitors to cars, is projected to grow at an annual compound rate of 23.1% between 2014 to 2020, reaching 50.1 billion things in 2020.
What is the IoT? In the minds of the business and IT executives surveyed, the IoT is associated with “ever-greater levels of connectivity; more intelligence built into devices, objects, and systems; and a strong data and applied learning orientation.” These views “sync-up well with the macro trends of more powerful and pervasive computing and storage, the further blurring of the physical and the virtual and the harnessing of big data for real-world functional activities.”
What is the current level of IoT adoption? 60% of organizations have started an IoT initiative, 45% of which were funded by a new budget allocation. An additional 23% of companies plan to start an IoT initiative within a year. About 90% of the 500 executives Bain surveyed remain in the planning and proof-of-concept stage, and only about 20% expect to implement solutions at scale by 2020.
What is the perceived impact of the IoT compared to other new technologies? The IoT leads other much-discussed technologies, including robotics and artificial intelligence, as the technology that is having the most impact on the business.
What are the expected benefits from IoT and how do they relate to existing activities and operations? The top 5 expected benefits are:
While the expected benefits are roughly split between existing operations and new products or revenue streams, a majority of businesses (61%) report having their IoT initiative as “enabling and extending” technology as opposed to regarding it as a separate and distinct activity (37%).
Bain also found high expectations of the potential benefits of the IoT, including improving the quality of products or services, improving the productivity of the workforce, and increasing the reliability of operations.
Are they too optimistic or too pessimistic? 57% of respondents believe their organization is very well equipped or mostly well equipped to manage the security component of IoT. “Given the number of security unknowns with IoT,” says CompITA, “especially in areas that may be beyond the control of the operator, this confidence may be misplaced.” Indeed, Bain found security at the top of the list of concerns about IoT, with 45% of respondents citing it as one of the top three barriers to IoT implementation. Similarly, when Forrester surveyed 232 companies developing IoT products it found that 38% anticipated security to be the biggest challenge to IoT implementation, more than any other issue and 64% cited data and device security as the most important capability for their IoT product. Finally, a Tripwire survey of 220 security professionals found that only 30% felt their organizations were prepared for security threats related to IoT devices.
Originally published on Forbes.com
