Category: data availability

How Artificial Intelligence Still Requires Tutoring in the Real World

Complementing our optimistic post ‘How Artificial Intelligence Progresses by Leaps and Bounds‘ on the incredible advances of AI when applied to rigorous logical situations and combinatorial problems, we have had recently (2016) on an other hand a dramatic result about AI applied to real life conversation.

A Microsoft AI chatbot went mad in less than 24h. The details are in this post from The Verge ‘Twitter taught Microsoft’s AI chatbot to be a racist asshole in less than a day‘.

This shows that AI is not yet ready to learn by itself without a tutor when it comes to the real world. Which also shows that a lot of our filters have to do with sound judgment and de facto social feedback. That will be quite difficult to emulate in the world of AI!

Share

How Brain-Computer Interfaces and Brain Implants Are Already Effective

The very interesting Bloomberg post ‘Brain-Computer Interfaces Are Already Here‘ describes graphically and in a compelling manner the progress of brain-computer interfaces today.

The most promising technology, it seems, involves an implant in the brain that can be plugged in the computer (instead of non-invasive technologies). The results that are mentioned in the post are quite amazing. “The past year has been particularly impressive. Researchers at the University of Pittsburgh Medical Center connected touch sensors from a robot’s fingertips to a paralyzed man’s sensory cortex so he could feel what it was touching. At Case Western, scientists linked a paralyzed man’s motor cortex to a computer that electrically stimulated muscles in his arm, enabling him to bring a forkful of food from a dish to his mouth. At Brown, Borton’s team implanted electrodes and a wireless transmitter in a monkey’s motor cortex and connected it to a receiver wired to the animal’s leg, restoring its walking motion.”

Although confined today to remediation of people with motorsensory problems, the technology might well soon become mainstream. The potential is quite difficult to envisage, in particular if that can enhance some of our physical and brain capabilities.

Share

How Wearables Will Change our Lives

In this excellent article (in French) ‘and if tomorrow your watch was to replace your shrink‘, the potential benefits of wearables and ongoing monitoring to deal medically with psychological issues is highlighted.

In particular for people with borderline psychological troubles, feedback from smartwatches and other wearables depending on the location, situation, ambient noise and physiological factors such as heart rate could be a nice way to deal with excessive anxiety and other difficult situations.

And it would provide 24/7 monitoring, not just more or less frequent encounters with a psychologist.

When I searched for a picture for this post about smartwatches I could almost only find pictures of people doing some form of sport. But in the very next future, these devices and their monitoring capability will find their way in our daily lives and may be coupled with personalized feedback systems that might intervene in certain situations.

Of course on some aspects this might seem a bit annoying to be monitored continuously, on the other hand as any new technology we will need some time to tame it and learn how to live with it, taking the benefits and avoiding the shortcomings.

Share

How Availability of More Data and AI Capability Might not be Entirely Related

In this interesting column of WIRED ‘AI and “enormous data’ could make tech giants harder to topple‘, the results of a study by Google is exposed, which shows that machine-learning AI algorithms can become significantly better if exposed to very large amounts of data.

This is of course an interesting finding because intuitively, the learning capability must become asymptotic at the end; it seems that the asymptote would be actually somewhat further away than thought. It exists though: “Crunching Google’s giant dataset of 300 million images didn’t produce a huge benefit—jumping from 1 million to 300 million images increased the object detection score achieved by just 3 percentage points“.

The paper continues to discuss on the fact that this finding would prove an advantage for those companies that currently amass huge amounts of data such as Google or Facebook. However there is something to be examined as to whether this scale of ‘enormous data’ really means a definite advantage of the AI algorithms that are derived. I personally believe that there must be also something in the quality of the dataset – in particular, how many deviant data is available that is properly categorized – how much like real life it looks. And it is this quality of data which is important for proper learning.

I am thus not entirely convinced by the argument that ‘enormous data’ is better than ‘big data’ when it comes to the value to be derived from AI. Maybe real-life data would be more discriminating.

Share

How the Computer Disappears as a Computer

In the Verge’s column ‘The disappearing computer – Tech was once always in your way. Soon, it will be almost invisible‘, Walt Mossberg makes the point that with the advent of wearables and the pervasion of computing power in our lives, computers will become invisible.

Instead of the old fashioned desktop with its keyboard and screen, or tablet with its finger-touch interface; computers will be where we won’t see them anymore. And they will have even more influence in our lives.

“ I expect that one end result of all this work will be that the technology, the computer inside all these things, will fade into the background. In some cases, it may entirely disappear, waiting to be activated by a voice command, a person entering the room, a change in blood chemistry, a shift in temperature, a motion. Maybe even just a thought. Your whole home, office and car will be packed with these waiting computers and sensors. But they won’t be in your way, or perhaps even distinguishable as tech devices.

This is ambient computing, the transformation of the environment all around us with intelligence and capabilities that don’t seem to be there at all.”

This of course gives quite some food for thought. There will be advantages and drawbacks from this situation. And I guess more advantages than we can envisage today.

Share

How We Can Protect our Business from Competition in the Future (the Moat)

In ‘The New Moats: Why Systems of Intelligence™ are the Next Defensible Business Model‘, the author Jerry Chen makes the point that traditional defensive moats around businesses are getting obsolete and that the new moat is around intelligent technology.

A Moat Around My Castle

“Companies that focus too much on technology without putting it in context of a customer problem will be caught between a rock and a hard place?—?or as I like to say, between open source and a cloud place”

Jerry Chen goes on explaining that he believes the new competition defences will be built around systems of intelligence that can combine several data sources to create substantial value.”In all of these markets, the battle is moving from the old moats, the sources of the data, to the new moats, what you do with the data“.

Personally I agree half-way, in particular because Jerry Chen places a lot of expectations on Artificial Intelligence. It might evolve that way, but for the moment, from my experience trying to create value-added applications for organizations, it is the engagement of the users around the data that creates value. The meaning is given by the experience of the users (although this might need to be facilitated and supported to properly define those items of value).

Yes, future defences to competition will be in clever data meaning development. But let’s not forget the engagement of the people around the data-set and the softer component of value creation. Here lies the real value of the future.

Share

How Digital Can Change Justice – Predictive Justice in Practice

As a Business Angel I see quite a few startup ideas and the concept of Predictice has struck me as a major experiment of digital applied to public service (disclosure: I liked it so much they got some investment from me too, just for the sake of following up their evolution more closely!).

The concept is to use all available judicial decisions for France (soon to be open-source but elsewhere available on some professional databases) to analyze the data using syntax associations and thus be able to give statistics on the expectations of a decision concerning a particular case. To be relevant, the algorithms have to be built with professionals using a specific process; so algorithms are being progressively developed in several judicial fields (family, private property etc).

The results are quite astounding as the system can give the average chance and compensation amount in fine grained detail based on the details of the situation and even the geographic location of the jurisdiction (which will certainly raise some eyebrows as it appears that there are regional differences).

The interesting thing is that it brings justice to a new level. It does not replace lawyers or judges, but it brings the debate to another level. As a support to the work of humans it brings substantial value. It gives valuable feedback to jurisdictions. More generally, with this startup Big Data transforms the way justice is being practiced.

There are certainly many fields of public service that are ripe to be disrupted in the same manner.

Share

How Facebook Makes Us Feel Bad

There is no doubt that social networks like Facebook become addictions because of the feelings they generate. The unfortunate thing is that it appears now proven that they make us feel bad, as explained in this serious Harvard Business Review paper ‘A New, More Rigorous Study Confirms: The More You Use Facebook, the Worse You Feel‘.

The study was using statistics and does not explain what could be the probable cause of this conclusion. However personally I have a suspect, described in our post ‘Why Who We Are Is Not What We Post‘: what we see from other people on social networks is much more positive than real life (whereas in real life interaction, we get the full version or at least something closer thanks to the entire context).

Definite conclusion: keep those real life interactions running so that we do not believe all what our online network tells us… and let’s keep our positivity!

Share

How Professional Book Publishing Fact-Checking Remains an Illusion

When we buy a professionally published book we tend to believe that it has been more seriously fact-checked than a self-published book. But that is not quite the case as explained in this interesting paper ‘Gay Talese Isn’t Alone: Why Aren’t More Books Factchecked?‘.

This must-read article narrates the story of an author seeking a thorough fact-checking of this book, after a few very public fiascos of the publishing industry. And he finds it very hard to do in a comprehensive manner.

This is quite the same issue of accuracy of Wikipedia against traditional encyclopedia. Traditional encyclopedia are only checked by a limited number of people, which can introduce bias and limits the means available for thorough fact-checking. Wikipedia is as accurate, in a different way: less bias but also more smaller mistakes.

The conclusion of the experiment was clear: “Yet even with all these eyes, not just diligently looking and parsing the words of the book, but some specifically hired to find errors, we missed something.” Not something big, but inaccuracies are always likely to enter a book volume.

So don’t always believe that professionally published books are 100% accurate!

Share

How the Internet of Things Winners Will Be the Ones Overcoming Hacking Risks

The Internet of Things is spreading, multiplying the number of clever devices and intruding deeper in our privacy. Those who will succeed in that market are those that will master the technologies that avoid fraud and excessive privacy intrusion.

HackedThe Internet of Things faces huge hacking risks. For two reasons:

  • IoT devices are relatively easier to hack because they do not usually include software upgrade and because they are based on standard chips with many more functionalities.
  • The consequences of hacking can also be much more visible, being devices that control the physical space.

Lately there have been some attacks launched from networks of connected IoT devices, that have been greater than anything recorded yet – for example read ‘A massive attack that may have hijacked online cameras will soon be “the new normal’.

At the start of internet, many services collapsed due to the issue of managing spam and fraud. For example, lots of paypal competitors died of this scourge, and paypal survived by having, from the start, implemented strong anti-fraud features.

The same will happen in the IoT: the survivors will be those that will develop and implement the technology that will avoid as much as possible hacking and subornation of their devices. This should be a key research angle for those that want to succeed in this field.

Share

How Google is Becoming an AI-centric company

After search-centric companies, and then mobile-centric companies, here come AI-centric companies! Following the trend such as at IBM, The new strategic impetus at Google is the inclusion of Artificial Intelligence in all its services, with dramatic quality improvements.

google_aiThis interesting NYTimes article ‘the Great AI awakening‘ is worth reading. It hightlights in particular the work of a particular division at Google called “Google Brain” with a focus on the usage of neural networks for deep machine learning and outcome quality improvements. According to the paper, in particular for the ‘Translate’ application, “the AI system has demonstrated overnight improvements roughly equal to the total gains the old one had accrued over its entire lifetime” (i.e. since 2006).

The paper also interestingly gives an account of the historical moves that have made machine learning based on neural networks mainstream in the past few years.

Let’s brace for similar improvements in a bunch of similar services that we are increasingly using in our daily life!…

Share

Why We Need to Connect With People to Get More Often Below Superficial

Following up from our post ‘How Everyone you Will Ever Meet Knows Something that You Don’t‘, the issue then becomes how to share this knowledge. How can we be in a position to better exchange knowledge through our daily interactions with people?

coffee connectionThe reality is that 95%+ of our daily interactions with people remain at a too superficial level to figure out what it is they know we don’t know. The issue is then to figure out how to setup those conversations in a way to enrich our experience and their experience.

It all comes down to connecting in the right manner, demonstrating interest to the person, its interests and aspirations. It also come down to a benevolent attitude that does not seek immediate advantage or profit from the relationship.

Of course that takes time so we can’t do that for everyone we meet, but we can certainly do better.

Benevolence is important. I had written first the first sentence of this post “how to benefit from this knowledge”. But the point is not to benefit, but to share!

Let’s try to learn more about the world by connecting better with more people, learning exciting new stuff we did not even know existed and sharing our knowledge too!

Share