Syndicate content

Data Science

Conversations with Chatbots: Exploring AI’s Potential for Development

Haishan Fu's picture

Development work is getting more technologically sophisticated by the day. The World Bank’s Information and Technology Solutions (ITS) department recently started an Artificial Intelligence (AI) Initiative. At the launch event, we explored the role of AI in development and what it might mean for the work that we do here at the Bank. In short: AI is already here, international organizations have an important role to play, and we need to invest in our skills and expertise.

AI is already being incorporated into development projects

A growing family of Artificial Intelligence techniques are being employed in development. Using machine learning for classification and prediction tasks is becoming as routine as running regressions. Our team recently launched a data science competition on poverty prediction and has been evaluating the performance of different machine learning algorithms. This includes the use of automated machine learning where the machine itself helps to select and tune models in a way a data scientist ordinarily would.

Weekly wire: The global forum

Roxanne Bauer's picture

World of NewsThese are some of the views and reports relevant to our readers that caught our attention this week.

Transparency, Accountability, and Technology
Plan International
The recently launched Sustainable Development Goals have kicked off a renewed development agenda that features, among other things, a dedicated emphasis on peace, justice, and strong institutions. This emphasis, encapsulated in Goal #16, contains several sub-priorities, including reducing corruption; developing effective, accountable, and transparent institutions; ensuring inclusive, participatory, and representative decision-making; and ensuring access to information.  Indeed, the governance-related Goals merely stamp an official imprimatur on what have now become key buzzwords in development. Naturally, where there are buzzwords, there are “tools.” In many cases, those “tools” turn out to be information and communications technologies, and the data flows they facilitate. It’s no wonder, then, that technology has been embraced by the development community as a crucial component of the global accountability and transparency “toolkit.”

Freedom in the World 2016
Freedom House
The world was battered in 2015 by overlapping crises that fueled xenophobic sentiment in democratic countries, undermined the economies of states dependent on the sale of natural resources, and led authoritarian regimes to crack down harder on dissent. These unsettling developments contributed to the 10th consecutive year of decline in global freedom.

Just-in (New Year’s resolution)-time learning

Abir Qasem's picture

Mediated Reality running on Apple iPhoneHello readers,
 
In this season of making resolutions (and hopefully sticking to a few of them) we invite you to join us for a year long skills transfer discussion/blog series on technology aided gut (TAG) checks.
 
TAG is a term we have coined to describe the use of simple web programming tools and techniques to do basic gut checks on data - big and small. TAG does not replace data science, rather it complements it. TAG empowers you - the development professionals - who rely on the story the data tells to accomplish your tasks. It does so by giving a you good enough idea about the data before you delve into the sophisticated data science methods (here is a good look at the last 50 years of data science from Stanford’s Dr. Donoho). In many cases it actually allows you to add your own insights to the story the data tells. As the series progresses we will talk a lot about TAGs.  For the eager-minded here’s an example of TAG usage in US politics.
 
In this series, we will use a just-in-time learning strategy to help you learn to do TAG checks on your data.  Just in time learning, as the name implies, is all about providing only the right amount of information at the right time. It is the minimum, essential information needed to help a learner progress to the next step. If the learner has a specific learning objective, just-in-time learning can be extremely efficient and highly effective. A good example of just in time information is the voice command a GPS gives you right before a turn. Contrast this with the use of maps before the days of GPS. You were given way more information than you needed and in a format that is not conducive to processing when you are driving.