“I’d really rather work alone. . .” Most of us have heard that from a student (or several students) when we assign a group project, particularly one that’s worth a decent amount of the course grade. It doesn’t matter that the project is large,…
In 2019, does it matter where do we perform our work?
When I asked in 2009 to use e-conferencing tools (Skype, Adobe Connect back then) to allow better attendance at faculty meetings, there was a mountain of arguments why NOT to. Such attitude was clearly expressed during the slow and painful advent of “online” education, which still leans more to “correspondence course” mentality rather then synchronous and interactive modern education.
The worst part is that in 2019 the attitude still persists.
large data is inherently noisy. \In general, the more “democratic” the production channel, the dirtier the data – which means that more effort has to be spent on its cleaning. For example, data from social media will require a longer cleaning pipeline. Among others, you will need to deal with extravagancies of self-expression like smileys and irregular punctuation, which are normally absent in more formal settings such as scientific papers or legal contracts.
The other major challenge is the labeled data bottleneck
crowd-sourcing and Training Data as a Service (TDaaS). On the other hand, a range of automatic workarounds for the creation of annotated datasets have also been suggested in the machine learning community.
Algorithms: a chain of disruptions in Deep Learning
Neural Networks are the workhorse of Deep Learning (cf. Goldberg and Hirst (2017) for an introduction of the basic architectures in the NLP context). Convolutional Neural Networks have seen an increase in the past years, whereas the popularity of the traditional Recurrent Neural Network (RNN) is dropping. This is due, on the one hand, to the availability of more efficient RNN-based architectures such as LSTM and GRU. On the other hand, a new and pretty disruptive mechanism for sequential processing – attention – has been introduced in the sequence-to-sequence (seq2seq) model by Sutskever et al. (2014).
Consolidating various NLP tasks
the three “global” NLP development curves – syntax, semantics and context awareness
the third curve – the awareness of a larger context – has already become one of the main drivers behind new Deep Learning algorithms.
A note on multilingual research
Think of different languages as different lenses through which we view the same world – they share many properties, a fact that is fully accommodated by modern learning algorithms with their increasing power for abstraction and generalization.
Spurred by the global AI hype, the NLP field is exploding with new approaches and disruptive improvements. There is a shift towards modeling meaning and context dependence, probably the most universal and challenging fact of human language. The generalisation power of modern algorithms allows for efficient scaling across different tasks, languages and datasets, thus significantly speeding up the ROI cycle of NLP developments and allowing for a flexible and efficient integration of NLP into individual business scenarios.
Facebook’s board works more like an advisory committee than an overseer, because Mark controls around 60 percent of voting shares. Mark alone can decide how to configure Facebook’s algorithms to determine what people see in their News Feeds, what privacy settings they can use and even which messages get delivered. He sets the rules for how to distinguish violent and incendiary speech from the merely offensive, and he can choose to shut down a competitor by acquiring, blocking or copying it.
We are a nation with a tradition of reining in monopolies, no matter how well intentioned the leaders of these companies may be. Mark’s power is unprecedented and un-American.
It is time to break up Facebook.
America was built on the idea that power should not be concentrated in any one person, because we are all fallible. That’s why the founders created a system of checks and balances.
More legislation followed in the 20th century, creating legal and regulatory structures to promote competition and hold the biggest companies accountable.
Starting in the 1970s, a small but dedicated group of economists, lawyers and policymakers sowed the seeds of our cynicism. Over the next 40 years, they financed a network of think tanks, journals, social clubs, academic centers and media outlets to teach an emerging generation that private interests should take precedence over public ones. Their gospel was simple: “Free” markets are dynamic and productive, while government is bureaucratic and ineffective.
From our earliest days, Mark used the word “domination” to describe our ambitions, with no hint of irony or humility.
Facebook’s monopoly is also visible in its usage statistics.About 70 percent of American adults use social media, and a vast majority are on Facebook products. Over two-thirds use the core site, a third use Instagram, and a fifth use WhatsApp. By contrast, fewer than a third report using Pinterest, LinkedIn or Snapchat. What started out as lighthearted entertainment has become the primary way that people of all ages communicate online.
The F.T.C.’s biggest mistake was to allow Facebook to acquire Instagram and WhatsApp. In 2012, the newer platforms were nipping at Facebook’s heels because they had been built for the smartphone, where Facebook was still struggling to gain traction. Mark responded by buying them, and the F.T.C. approved.
The News Feed algorithm reportedly prioritized videos created through Facebook over videos from competitors, like YouTube and Vimeo. In 2012, Twitter introduced a video network called Vine that featured six-second videos. That same day, Facebook blocked Vine from hosting a tool that let its users search for their Facebook friends while on the new network.The decision hobbled Vine, which shut down four years later.
unlike Vine, Snapchat wasn’t interfacing with the Facebook ecosystem; there was no obvious way to handicap the company or shut it out. So Facebook simply copied it. (opyright law does not extend to the abstract concept itself.)
As markets become more concentrated, the number of new start-up businesses declines. This holds true in other high-tech areas dominated by single companies, like search (controlled by Google) and e-commerce (taken over by Amazon). Meanwhile, there has been plenty of innovation in areas where there is no monopolistic domination, such as in workplace productivity (Slack, Trello, Asana), urban transportation (Lyft, Uber, Lime, Bird) and cryptocurrency exchanges (Ripple, Coinbase, Circle).
The choice is mine, but it doesn’t feel like a choice. Facebook seeps into every corner of our lives to capture as much of our attention and data as possible and, without any alternative, we make the trade.
Just last month, Facebook seemingly tried to bury news that it had stored tens of millions of user passwords in plain text format, which thousands of Facebook employees could see. Competition alone wouldn’t necessarily spur privacy protection — regulation is required to ensure accountability — but Facebook’s lock on the market guarantees that users can’t protest by moving to alternative platforms.
Mark used to insist that Facebook was just a “social utility,” a neutral platform for people to communicate what they wished. Now he recognizes that Facebook is both a platform and a publisher and that it is inevitably making decisions about values. The company’s own lawyers have argued in court that Facebook is a publisher and thus entitled to First Amendment protection.
As if Facebook’s opaque algorithms weren’t enough, last year we learned that Facebook executives had permanently deleted their own messages from the platform, erasing them from the inboxes of recipients; the justification was corporate security concerns.
Mark may never have a boss, but he needs to have some check on his power. The American government needs to do two things: break up Facebook’s monopoly and regulate the company to make it more accountable to the American people.
We Don’t Need Social Media
The push to regulate or break up Facebook ignores the fact that its services do more harm than good
Hughes joins a growing chorus of former Silicon Valley unicorn riders who’ve recently had second thoughts about the utility or benefit of the surveillance-attention economy their products and platforms have helped create. He is also not the first to suggest that government might need to step in to clean up the mess they made
Nick Srnicek, author of the book Platform Capitalismand a lecturer in digital economy at King’s College London, wrotelast month, “[I]t’s competition — not size — that demands more data, more attention, more engagement and more profits at all costs
Microcredentials, or short-form online learning programs, is the latest buzzword that higher education providers are latching onto. They come with diminutive names such as Micromasters (by several universities working with edX) and nanodegrees (by Udacity). But they have the potential to shake up graduate education, potentially reducing demand for longer, more-traditional professional programs. At the core of the trend is the idea that professionals will go “back to school” repeatedly over their lifetimes, rather than carving out years at a time for an MBA or technical degree.
Credential Engine, a nonprofit funded by the Lumina Foundation, Microsoft and JPMorgan Chase, today launched its Credential Registry, a digital platform where institutions can upload degrees and credentials so prospective students can search for and compare credentials side-by-side.
Udacity won a trademark for Nanodegree last year. And in April, the nonprofit edX, founded by MIT and Harvard University to deliver online courses by a consortium of colleges, applied for a trademark on the word MicroMasters. And MicroDegree? Yep, that’s trademarked too, by yet another company.
colleges and universities that seek to meet corporate needs must move beyond monolithic programs and think in terms of competencies, unbundling curriculum, modularizing and “microlearning.” Many institutions are already pioneering efforts in this direction, from the certificate- and badge-oriented University of Learning Store (led by the Universities of Wisconsin, California, Washington and others) to Harvard Business School’s HBX, and the new “iCert” that we developed at Northeastern University. These types of shorter-form, competency-oriented programs can better fit corporate demands for targeted and applied learning.
Despite its name, the Learning Management System (LMS) is not about learning. The LMS was originally the CMS—Course Management System.
The LMS succeeds as a core productivity tool for educators because it allows institutions to extend their academic capacity and transcend the constraints of time and space. However, the Learning Management System was never able to deliver on the promise of its new name because it was created for a completely different purpose: course management. Learning doesn’t happen within the digital space of the LMS; it happens beyond its borders.
Today’s generation of students is deeply social and collaborative. They rely on real-time online interaction, collaboration and sharing to feel supported, confident and successful. Having grown up on iPhone, Snapchat and Instagram, this generation expects seamless experiences that are deeply social and collaborative.
In the post-LMS world, learning technology is student-centric in its design because today’s students are vocal, creative and eager to share their blue sky ideals and ideas.
The post-LMS world is also social by nature. in the post-LMS world, learning technology is simple, modern and mobile.