Daily Archives: February 24, 2017

Forbes

What Companies Are Winning The Race For Artificial Intelligence? Forbes ... general AI research, including traditional software engineers to build infrastructure and tooling, UX designers to help make research tools, and even ecologists (Drew Purves) to research far-field ideas like the relationship between ecology and ...

The rest is here:

What Companies Are Winning The Race For Artificial Intelligence? - Forbes

Forbes

College Student Uses Artificial Intelligence To Build A Multimillion-Dollar Legal Research Firm Forbes Lawyers spend years in school learning how to sift through millions of cases looking for the exact language that will help their clients win. What if a computer could do it for them? It's not the kind of question many lawyers would dignify with an answer.

Read the original post:

College Student Uses Artificial Intelligence To Build A Multimillion-Dollar Legal Research Firm - Forbes

When a technology hype flops, do you think the industry can use it as a learning experience? A time of self-examination? An opportunity to pause and reflect on making the next consumer or business tech hype a bit less stupid?

Don't be silly.

What it does is pile the next hype on to the last hype, and call it "Hype 2.0".

"With AI integration in wearables, we are entering 'wearable 2.0' era," proclaim analysts Counterpoint Research in one of the most optimistic press releases we've seen in a while.

It's certainly bullish for market growth, predicting that "AI-powered wearables will grow 376 per cent annually in 2017 to reach 60 million units."

In fact it's got a new name for these "hearables". Apple will apparently have 78 per cent of this hearable market.

The justification for the claim is that language-processing assistants like Alexa will be integrated into more products. Counterpoint also includes Apple Airpods and Beats headphones as "AI-powered hearables", which may be stretching things a little.

It almost seems rude to point out that the current wearables market a bloodbath for vendors is already largely "hearable". Android Wear has been obeying OK Google commands spoken by users since it launched in 2014:

Apple built Siri into its Apple Watch in 2015 with its first update, watchOS 2:

Microsoft's Band built in Cortana:

If a "smart" natural language interface had the potential to make wearables sell, surely we would know it by now. But we hardly need to tell you what sales of these devices are. Many vendors have hit paused, or canned their efforts completely. You could even argue that talking into a wearable may be one of the reasons why the wearable failed to be a compelling or successful consumer electronics story. People don't want to do it.

Sprinkling the latest buzzword machine learning or AI over something that isn't a success doesn't suddenly make that thing a success. But AI has always had a cult-like quality to it: it's magic, and fills a God-shaped hole. For 50 years, the divine promise of "intelligent machines" has periodically overcome people's natural scepticism as they imagine a breakthrough is close at hand. Then it recedes into the labs again. All that won't stop people wishing that this time AI has Lazarus-like powers.

We can't wait for our machine-learning powered Sinclair C5 the Deluxe Edition with added Blockchain.

Can you?

Continued here:

Artificial intelligence 'will save wearables'! - The Register

From Jeopardy winners and Go masters to infamous advertising-related racial profiling, it would seem we have entered an era in which artificial intelligence developments are rapidly accelerating. But a fully sentient being whose electronic brain can fully engage in complex cognitive tasks using fair moral judgement remains, for now, beyond our capabilities.

Unfortunately, current developments are generating a general fear of what artificial intelligence could become in the future. Its representation in recent pop culture shows how cautious and pessimistic we are about the technology. The problem with fear is that it can be crippling and, at times, promote ignorance.

Learning the inner workings of artificial intelligence is an antidote to these worries. And this knowledge can facilitate both responsible and carefree engagement.

The core foundation of artificial intelligence is rooted in machine learning, which is an elegant and widely accessible tool. But to understand what machine learning means, we first need to examine how the pros of its potential absolutely outweigh its cons.

Simply put, machine learning refers to teaching computers how to analyse data for solving particular tasks through algorithms. For handwriting recognition, for example, classification algorithms are used to differentiate letters based on someones handwriting. Housing data sets, on the other hand, use regression algorithms to estimate in a quantifiable way the selling price of a given property.Machine learning, then, comes down to data. Almost every enterprise generates data in one way or another: think market research, social media, school surveys, automated systems. Machine learning applications try to find hidden patterns and correlations in the chaos of large data sets to develop models that can predict behaviour.

Machine learning, then, comes down to data. Almost every enterprise generates data in one way or another: think market research, social media, school surveys, automated systems. Machine learning applications try to find hidden patterns and correlations in the chaos of large data sets to develop models that can predict behaviour.

Data have two key elements samples and features. The former represents individual elements in a group; the latter amounts to characteristics shared by them.

Look at social media as an example: users are samples and their usage can be translated as features. Facebook, for instance, employs different aspects of liking activity, which change from user to user, as important features for user-targeted advertising.

Facebook friends can also be used as samples, while their connections to other people act as features, establishing a network where information propagation can be studied.

Outside of social media, automated systems used in industrial processes as monitoring tools use time snapshots of the entire process as samples, and sensor measurements at a particular time as features. This allows the system to detect anomalies in the process in real time.

All these different solutions rely on feeding data to machines and teaching them to reach their own predictions once they have strategically assessed the given information. And this is machine learning.

Any data can be translated into these simple concepts and any machine-learning application, including artificial intelligence, uses these concepts as its building blocks.

Once data are understood, its time to decide what do to with this information. One of the most common and intuitive applications of machine learning is classification. The system learns how to put data into different groups based on a reference data set.

This is directly associated with the kinds of decisions we make every day, whether its grouping similar products (kitchen goods against beauty products, for instance), or choosing good films to watch based on previous experiences. While these two examples might seem completely disconnected, they rely on an essential assumption of classification: predictions defined as well-established categories.

When picking up a bottle of moisturiser, for example, we use a particular list of features (the shape of the container, for instance, or the smell of the product) to predict accurately that its a beauty product. A similar strategy is used for picking films by assessing a list of features (the director, for instance, or the actor) to predict whether a film is in one of two categories: good or bad.

By grasping the different relationships between features associated with a group of samples, we can predict whether a film may be worth watching or, better yet, we can create a program to do this for us.

But to be able to manipulate this information, we need to be a data science expert, a master of maths and statistics, with enough programming skills to make Alan Turing and Margaret Hamilton proud, right? Not quite.

We all know enough of our native language to get by in our daily lives, even if only a few of us can venture into linguistics and literature. Maths is similar; its around us all the time, so calculating change from buying something or measuring ingredients to follow a recipe is not a burden. In the same way, machine-learning mastery is not a requirement for its conscious and effective use.

Yes, there are extremely well-qualified and expert data scientists out there but, with little effort, anyone can learn its basics and improve the way they see and take advantage of information.

Going back to our classification algorithm, lets think of one that mimics the way we make decisions. We are social beings, so how about social interactions? First impressions are important and we all have an internal model that evaluates in the first few minutes of meeting someone whether we like them or not.

Two outcomes are possible: a good or a bad impression. For every person, different characteristics (features) are taken into account (even if unconsciously) based on several encounters in the past (samples). These could be anything from tone of voice to extroversion and overall attitude to politeness.

For every new person we encounter, a model in our heads registers these inputs and establishes a prediction. We can break this modelling down to a set of inputs, weighted by their relevance to the final outcome.

For some people, attractiveness might be very important, whereas for others a good sense of humour or being a dog person says way more. Each person will develop her own model, which depends entirely on her experiences, or her data.

Different data result in different models being trained, with different outcomes. Our brain develops mechanisms that, while not entirely clear to us, establish how these factors will weighout.

What machine learning does is develop rigorous, mathematical ways for machines to calculate those outcomes, particularly in cases where we cannot easily handle the volume of data. Now more than ever, data are vast and everlasting. Having access to a tool that actively uses this data for practical problem solving, such as artificial intelligence, means everyone should and can explore and exploit this. We should do this not only so we can create useful applications, but also to put machine learning and artificial intelligence in a brighter and not so worrisome perspective.

There are several resources out there for machine learning although they do require some programming ability. Many popular languages tailored for machine learning are available, from basic tutorials to full courses. It takes nothing more than an afternoon to be able to start venturing into it with palpable results.

All this is not to say that the concept of machines with human-like minds should not concern us. But knowing more about how these minds might work will gives us the power to be agents of positive change in a way that can allow us to maintain control over artificial intelligence and not the other way around.

This article was originally published on The Conversation. Read the original article.

If you liked this article, you may also want to read:

See allthe latest robotics newson Robohub, orsign up for our weekly newsletter.

More:

Artificial intelligence: Understanding how machines learn - Robohub

News concerning Artificial Intelligence (AI) abounds again. The progress with Deep Learning techniques are quite remarkable with such demonstrations of self-driving cars, Watson on Jeopardy, and beating human Go players. This rate of progress has led some notable scientists and business people towarn about the potential dangers of AI as it approaches a human level. Exascale computers are being considered that would approach what many believe is this level.

However, there are many questions yet unanswered on how the human brain works, and specifically the hard problem of consciousness with its integrated subjective experiences. In addition, there are many questions concerning the smaller cellular scale, such as why some single-celled organisms can navigate out of mazes, remember, and learn without any neurons.

In this blog, I look at a recent review that suggests brain computations being done at a scale finer than the neuron might mean we are far from the computational, power both quantitatively and qualitatively. The review is by Roger Penrose (Oxford) and Stuart Hameroff (University of Arizona) on their journey through almost three decades of investigating the role of potential quantum aspects in neurons microtubules. As a graduate student in 1989, I was intrigued when Penrose, a well-known mathematical physicist, published the book, The Emperors New Mind, outlining a hypothesis that consciousness derived from quantum physics effects during the transition from a superposition and entanglement of quantum states into a more classical configuration (the collapse or reduction of the wavefunction). He further suggested that this process, which has baffled generations of scientists, might occur only when a condition, based on the differences of gravitational energies of the possible outcomes, is met (i.e., Objective Reduction or OR). He then went another step in suggesting that the brain takes advantage of the this process to perform computations in parallel, with some intrinsic indeterminacy (non-computability), and over a larger integrated range by maintaining the quantum mix of microtubule configurations separated from the noisy warm environment until this reduction condition was met (i.e., Orchestrated Objective Reduction or Orch OR).

As an anesthesiologist, Stuart Hameroff questioned how relatively simple molecules could cause unconsciousness. He explored the potential classical computational power of microtubules. The microtubules had been recognized as an important component of neurons, especially in the post synaptic dendrites and cell body, where the cylinders lined up parallel to the dendrite, stabilized, and formed connecting bridges between cylinders (MAPs). Not only are there connections between microtubules within dendrites but there are also interneuron junctions allowing cellular material to tunnel between neuron cells. One estimate of the potential computing power of a neurons microtubules (a billion binary state microtubule building blocks , tubulins, operating at 10 megahertz) is the equivalent computing power of the assumed neuronnet of the brain (100 billion neurons each with 1000 synapses operating at about 100 Hz). That is, the brains computing power might be the square of the standard estimate (10 petaflops) based on relatively simple neuron responses.

Soon after this beginning, Stuart Hameroff and Roger Penrose, found each others complementary approach and started forming a more detailed set of hypotheses. Much criticism was leveled about this view. Their responses included modifying the theory, calling for more experimental work, and defending against general attacks. Many experiments await to be done, including whether objective reduction occurs but this experiment cannot be done yet with the current resolution of laboratory instruments. Other experiments on electronic properties of microtubules were done in Japan in 2009 which discovered high conductance at certain frequencies from kilohertz to gigahertz frequencies. These measurements, which also show conductance increasing with microtubule length, are consistent with conduction pathways through aligned aromatic rings in the helical and linear patterns of the microtubule. Other indications of quantum phenomena in biology include the recent discoveries quantum effects in photosynthesis, bird navigation, and protein folding

There are many subtopics toexplore. Often the review discusses potential options without committing (or claiming) a specific resolution. These subtopics include interaction of microtubule with associated protein and transport mechanisms, the relationship of microtubules to diseases such as Alzheimers, the frequency of the collapse from the range of megahertz to hertz, memory formation and processing with molecules that bind to microtubules, the temporal aspect of brain activity and conscious decisions, whether the quantum states are spin (electron or nuclear) or electrical dipoles, the helical pattern of the microtubule (A or B), the fraction of microtubules involved with entanglement, the mechanism for environmental isolation, and the way that such a process might be advantageous in evolution. The review ends not with a conclusion concerning the validity of the hypothesis but instead lays a roadmap for the further tests that could rule out or support their hypothesis.

As I stated at the beginning, the progress in AI has been remarkable. However, the understanding of the brain is still very limited and the mainstream expectation that computers are getting close to equaling computing potential may be far off both qualitatively and quantitatively. While in the end it is unclear how much of this hypothesis will survive the test of experiments, it is very interesting to consider and follow the argumentative scientific process.

Stuart Hameroffs Web Site: http://www.quantumconsciousness.org/

Review Paper site: http://smc-quantum-physics.com/pdf/PenroseConsciouness.pdf

See original here:

Artificial Intelligence or Artificial Expectations? - Science 2.0

Google said it will begin offering media groups an artificial intelligence tool designed to stamp out incendiary comments on their websites.

The programming tool, called Perspective, aims to assist editors trying to moderate discussions by filtering out abusive "troll" comments, which Google says can stymie smart online discussions.

"Seventy-two per cent of American internet users have witnessed harassment online and nearly half have personally experienced it," said Jared Cohen, president of Google's Jigsaw technology incubator.

"Almost a third self-censor what they post online for fear of retribution," he added in a blog post on Thursday titled "When computers learn to swear."

Perspective is an application programming interface (API), or set of methods for facilitating communication between systems and devices, that uses machine learning to rate how comments might be regarded by other users.

The system, which will be provided free to media groups including social media sites, is being tested by The Economist, The Guardian, The New York Times and Wikipedia.

Many news organizations have closed down their comments sections over lack of sufficient human staff to monitor the postings for abusive content.

"We hope we can help improve conversations online," Cohen said.

Google has been testing the tool since September with The New York Times, which wanted to find a way to maintain a "civil and thoughtful" atmosphere in reader comment sections.

Perspective's initial task is to spot toxic language in English, but Cohen said the goal was to build tools for other languages, and to identify when comments are "unsubstantial or off-topic."

Twitter said earlier this month that it also would start rooting out hateful messages, which are often anonymous, by identifying the authors and prohibiting them from opening new accounts, or hiding them from internet searches.

Last year, Google, Twitter, Facebook and Microsoft signed a "code of good conduct" with the European Commission, pledging to examine most abusive content signalled by users within 24 hours.

See the original post here:

Google rolls out artificial intelligence tool for media companies to combat online trolls - CTV News

Its difficult to diagnose infants with autism due to trouble determining whether any behavioral traits common to autism are present. This difficulty is most pronounced before the age of two, and especially before the age of one, resulting in delayed diagnoses. All that may be changing, though, thanks to artificial intelligence and its ability to predict with high accuracy which infants will be diagnosed with autism by their second year.

As detailed by a new study funded by the US National Institute of Health, predictions of future Autism Spectrum Disorder can be made based on MRI scans of an infants brain. The technology heavily relies on brain scans taken of infants that are at high risk of being diagnosed with autism by their second birthday.

Using 106 brain scans of high-risk infants, researchers determined that certain aspects of the brains maturation may be early indicators of autism. One potential indicator is a quickly growing brain, one that grows faster than normal during the age spanning 12 to 24 months. As well, the cortex may grow faster than average during this time period.

Based on this information and using brain scans taken at 6 and 12 months, a customized algorithm was able to predict which infants would end up with an autism diagnoses with an 81-percent accuracy. Knowing whether a baby is likely to be autistic may help, in certain cases, early intervention and the preparation of parents.

SOURCE: PBS

View post:

Artificial intelligence used to detect very early signs of autism in infants - SlashGear