Category: data availability

How Access to Data Is a Key Issue for Start-Ups

Following on our previous post ‘How Data Really is the New Oil, and Better‘, there are two issues for data like for oil: access to the data, and then the ability to process it to produce value.

It is easier for startups to find new ways of processing the data to extract value, because the way it can be processed is heavily dependent on the utilization and hence niches can be produced to provide the user with specific benefits. The way to massage the data to provide value requires a lot of user interaction.

I observe that access to data is an issue for startups and the usual internet suspects have of course a substantial advance on that aspect.

In my limited experience as Business Angel with start-ups I observe that access to the data is often the most frustrating part. Governments’ public cata availability initiatives (open data) have not yet delivered in all countries and areas of public life. Private data can be difficult to get to help develop a product without entering legal issues, and they are often difficult to use technical as well because of their inconsistency or historical aspects. In that context data cleansing and preparation is becoming a real trade.

Hence for start-ups, access to relevant data is often the bottleneck and this aspect must be carefully vetted before launching a venture, because it is where you may fail or have to surrender to some data hoarding giant.

Share

How People Have Changed Their Discretionary Entertainment Expenditure Towards Streaming Platforms

In this interesting post by Frederic Filloux ‘The Consumer Trends That Destroyed Media’s Business Model‘, some interesting trends are exposed about our consumption habits in the area of entertainment (thus, discretionary expenditure). Of course those expenditures only represent a small share of our total budgets, but they have great influence on our understanding of the world.

The curves and statistics show that most expenditures are taken by subscriptions to internet and mobile services, followed closely by a fast growing video-on-demand trend, led by Netflix. This leaves only limited space in the budgets for other expenditures such as events and tickets, literature, or press subscriptions. This translates to more and more time spent on those services to the detriment of written books or media, or even live shows or visiting museums.

At the same time that there is an increasing development of somewhat addictive series instead of feature films, this trend is at the same time not too surprising but also concerning.

It definitely shows a substantial shift as many of those expenditures did not exist 20 years ago, therefore, they have necessarily replaced other activities and expenditures and those (including cable TV, the press, cinemas or generally live shows) are thus suffering. The trends give all of us much more choices in what we are watching while giving us potentially less opportunity to discover new things.

Anyway, it is quite illuminating to observe how our expenditure habits have dramatically changed in the last years, giving us more opportunity to choose our entertainment, with a large domination of online and video entertainment. This is certainly a trend that will continue.

Share

How Natural Language Processing Still Has Strong Limitations

I am getting quite interested these days about Artificial Intelligence (AI) and its actual applications. Most is about Deep Learning of course, and an essential element is Natural-Language Processing (NLP) or making sense for the machine of texts or words. This is an essential first stage to allow the machine to then perform statistical analysis of the data and produce all sorts of useful analysis.

I observe that NLP now performs very well in fields where expression is quite standardized and normalized, such as in legal or scientific fields. In particular, applications of AI to legal aspects is really becoming amazing. However, it still has limitations when it comes to analyzing informal correspondence and longer texts. This makes it harder to use AI to make sense of informal messages and data and to use those datasets as a basis for further analysis.

Of course it has improved substantially in the last months and years as any user of Google Translate or equivalent can witness: translations are now more to the point thanks to AI. Still it does not appear to be sufficient to deal with large sets of informal exchanges such as messages, email and other informal communication channels.

When this aspect will be overcome – which will take time and may not be immediately transferable between languages – the power of AI will be much more visible and dramatic than it is now. Let’s watch for progress in this area!

Share

How Dismantling Facebook Is an Increasingly Strong Idea

In the past months the pressure on dismantling the Facebook empire has increased dramatically: from the NY times pledge by a former co-founder Chris Hughes ‘It’s Time to Break Up Facebook‘, calls to revitalize the old anti-trust laws (such as in ‘Steering with the Windshield Wipers‘ by Cory Doctorow). And all of this under increasing evidence that more or less voluntarily social networks such as Facebook have strongly influenced elections (listen to the TED talk about ‘Facebook role in Brexit and the threat to democracy‘).

One particular aspect that appears to be concerning and specific to Facebook is the unrestricted power of its founder Mark Zuckerberg (owning 60% of the voting shares) and thus the associated governance issues (things are a bit different in other internet companies). This is a particular strong point made by Chris Hughes in his column, and is quite noticeable.

Another strong aspect coming out of all this literature is the fact that certain free market ideology developing since the 1970s have made acting against monopolies less straightforward than before, and that it would be legally harder today to apply those principles.

I truly believe that sooner or later, some action will be taken against internet behemoths in the field of anti-monopoly. They might defend themselves and hire superb lawyers, one day will come where their dominance must be kept in check. Let’s hope it is not too far away.

Share

How to Overcome the Science Reproducibility Crisis

Following up on our previous post on “How Fake Science is Strongly on the Rise and Endangers Us“, the issue or reproducibility in science is also coming up strongly: even paper and findings recognized as legitimate for a long time are put in question by the inability to reproduce results, and more specifically in human sciences. This post ‘Why Your Company Needs Reproducible Research‘ provides a good summary of the issues at stake.

Recent efforts are reproducing psychology results lead to “Only about 40% of the findings could be successfully replicated, while the rest were either inconclusive or definitively not replicated.” Similar proportions are obtained in business-related research.

While this may be due to very human bias like the need to show some results from research, and the inherent complexity of the environment around some experiments, there is definitely a need for more thorough replication requirements prior to confirming results. This puts more challenge on researchers but is probably a need in a world that sees increasingly fake science.

Science will always progress by invalidating previous results or restraining the boundaries of validity of previous results. This is a normal process, still we need to be wary to ensure reproducibility of results before they are spread as invariant truths.

Share

How Fake Science is Strongly on the Rise and Endangers Us

In this excellent paper ‘The Rise of Junk Science‘ the issue of increasing number of improperly reviewed publications is described. They now substantially dominate in number more professional scientific publications. Some even believe that there is a “global “epidemic” of scams by academic journals that was corrupting research and, in effect, endangering the public“.

With much easier access to publishing, there is an increasing number of scam scientific journals that are easily accessible, do not practice rigorous peer reviews. There are also an increasing number of fake conferences, that even name recognized scientists that don’t even attend in an attempt to attract other reputable scientists.

While this issue is probably not new – it was always possible to do vanity publishing – it spreads to an unprecedented level, and in a context where it is increasingly difficult to distinguish between legitimate and fake science. “for the first time in history, scientists and scholars worldwide are publishing more fraudulent and flawed studies than legitimate research—maybe ten times more. Approximately 10,000 bogus journals run rackets around the world, with thousands more under investigation, according to Cabell’s International, a publishing-services company. “We’re publishing mainly noise now,” Franco laments. “It’s nearly impossible to hear real signals, to discover real findings.”

Luckily there seem to be an increasing awareness of the issue and modern technology also allows to develop powerful tools to distinguish between junk science and legitimate science. Still we need to be careful that something looking like a scientific paper might also be fake or junk!

Share

How to Design a Font to Better Remember What you Read

I love this initiative from an Australian University: Sans Forgetica, the font that is scientifically designed to help remember courses: see the Sans Forgetica page.

“Sans Forgetica is a font designed using the principles of cognitive psychology to help you to better remember study notes. It was created by a multidisciplinary team of designers and behavioural scientists.”

The interesting part I find is that “Sans Forgetica is more difficult to read than most typefaces – and that’s by design. The ‘desirable difficulty’ you experience when reading information formatted in Sans Forgetica prompts your brain to engage in deeper processing.”

By making harder and longer to read and understand, it seems that we remember better. I find that is quite an interesting insight. It shows that it is important to spend sufficient time reading and processing what has been read to better remember.

Share

How Mark Zuckerberg Needs and Requests Increased Regulation

In an amazing turn of events Mark Zuckerberg published simultaneously in several languages on mainstream newspaper an opinion end March that calls for increased regulation of social network activities (see the Washington post version here: ‘Mark Zuckerberg: The Internet needs new rules. Let’s start in these four areas.’)

First let’s note that this move is probably due to the underlying observation that lack of regulation may spell doom on his empire. What are the four areas he is calling for increased regulation?

  • managing better harmful content and getting external input on removal rules, as well as standardizing rules
  • better rules around political advertisement
  • a globally recognized framework around privacy and data protection aligned with the EU rules
  • guaranteeing data portability from one service to the next

Mark Zuckerberg concludes: “The rules governing the Internet allowed a generation of entrepreneurs to build services that changed the world and created a lot of value in people’s lives. It’s time to update these rules to define clear responsibilities for people, companies and governments going forward.”

I observe that the topics raised are mostly about making sure all competitors will be subject to the same scrutiny as Facebook and avoid competitors that would avoid the complication of those basic requirements. The request for regulation could go beyond this minimum.

Still it is a start, and with major social network players now asking for more regulation, the door is opened for a serious regulatory review of what has become a major tool in the world’s hands.

Share

How Social Networks and Modern Media Isolate Large Parts of the Population

Specific groups of people, based on their condition and beliefs, have always preferred different means of information: politically-orientated newspapers have existed for a long time. Still the extent of isolation of certain groups of people has reached an unprecedented level. In the USA it would seem that “one third of the American electorate has been isolated in an information loop of its own. For this group, which mistrusts the mainstream press on principle, and as a matter of political identity, Trump has become the major source of information about Trump, along with Fox News, which has slowly been merging with the Trump government.” The detail is given in the Pressthink post ‘Hating on journalists the way Trump and his core supporters do is not an act of press criticism. It’s a way of doing politics.’

We have long noted that social networks tend to close one’s source of information to what is getting broadcast by his or her immediate connections. The extent of the situation in the US is however broader as some mainstream media, in particular the Fox News network, seem to have turned frankly one-sided; and because fake news are being disseminated from the highest governmental levels, with strong hate terminology being used as well. “These are people beyond the reach of journalism, immune to its discoveries. Trump is their primary source of information about Trump. The existence of a group this size shows that de-legitimizing the news media works. The fact that it works means we will probably see more of it.”

That this happens in a country where free speech and free press laws are among the strongest in the world show the extent of the transformation of news broadcasting our societies undergo. I do not believe the change is as dramatic as mentioned, but we need to be wary about it and track fake news and fake statements with more urgency.

Share

How Our Usage of Maps Has Changed Dramatically

It is amazing how our usage of maps has changed dramatically with the advent of portable electronic maps. The interesting part is that beyond the map itself as a tool, our relationship with our geographical environment is also changing dramatically too.

Dwindling Road Maps

I am of a generation that used paper maps and atlases to navigate the world, including for driving around. I notice how this mundane task (which was traditionally a highly contentious activity for couples when the significant other was reading the map) has changed.

What I find to be the most startling change of functionality is the possibility to zoom as we want, using maps as an actual multi-scale tool, in addition to knowing at any one time where one is located thanks to the GPS.

Moreover, the amount of data super-imposed on the map has increased tremendously, allow new usage of maps, and they also are increasingly updated more or less real time. We tend more and more to interpret our immediate environment not only with what we can observe, but with the added data available in our pockets, from the location of services and businesses to their rating and more detail about their operation. It is augmented reality in its infancy. And it is only the beginning.

It is amazing how with portable multi-scale, multi-layer maps our perception of the physical world around us is changing. I still hope we will able to keep relying on actual observation of reality!

Share

How Automated Sentiment Analysis Brings Emotions Into Automated Decision-Making

We know that emotions is a substantial source of decisions and actions for humans. Being able to identify that emotion is an essential technology that currently becomes mature, and I was not really aware of the extent of progress in that area. An excellent summary is given in this Quartz Obsession Post ‘Sentiment analysis‘.

More and more companies use sentiment analysis to drive their decisions about communication. The first applications were sentiment analysis based on social network posts, crudely based on the types of words people were using. Today automated sentiment analysis applies to images and videos – emotions can be detected from selfies and pictures and even from micro-expressions in videos, adding a lot of context and more subtle ways to detect the emotional condition of people.

It goes further – for example affective computing is “Affective Computing is computing that relates to, arises from, or deliberately influences emotion or other affective phenomena” (see the affective computing MIT group page). It clearly states here that we are at the stage where computing will start influencing emotions. Or at least, detecting our emotional state and propose to change it. This can be potentially dangerous, and this space needs to be watched closely. Social networks and brands will now not only inadvertently influence our emotions, but probably also voluntarily!

Share

How Newspaper Paywalls would be a Paradox regarding Spreading Quality News

Following up on our investigation of the economics of writing (see previous post ‘How Writers’ Income Sources Are Changing‘), an interesting perspective is given in this article ‘How Paywalls are Making Us Dumber‘. The thesis of this paper revolves around the paradox that great journalistic content needs to be compensated -hence paywalls and subscriptions- but that this prevents many people from accessing it, leaving space open for much less reliable news and even fake news, which are all freely accessible.

I believe it is useful to come back to a historical perspective. Since the beginning, journalism has been financed either by tycoons – sometimes well meaning, sometimes with the objective to manipulate – or by advertisement. Sometimes by both. In certain countries there are also subsidies by government, with all the problems related to free press. In general, the income from readers or subscriptions has always remained limited due to the costs of physical printing and distribution.

Therefore, it does not seem to me that the overall balance of financing of journalism is shifting dramatically. Jeff Bezos – a tycoon – had to take over the Washington Post recently; same happens in France for the main newspapers. Nor that fake news and opinionated papers have been around for lots of time. The current newspaper crisis is mainly linked to the fact that advertisement dollars have migrated elsewhere, and newspapers are struggling to compensate.

Of course, the reach of internet and the easier spreading of fake news is a concern, as well as the difficulty to effectively regulate it. There are still many sources of quality information sufficiently open to benefit from it. Financing a news outlet only with income from readers has always been an illusion. At the same time it does not seem to me that paywalls are really a problem, in particular if they let read a few key articles per month for free if one investigates a specific subject. What do you think?

Share