Fairness in Machine Learning


Presentation by Delip Rao: “…The models you create have power to get people arrested or vindicated, get loans approved or rejected, determine what interest rate should be charged for such loans, who should be shown to you in your long list of pursuits on your Tinder, what news do you read, who gets called for a job phone screen or even a college admission… the list goes on.

So what can you do about it?…

I have detailed notes for some of these slides. If you would like to follow those, try going directly to Google Slides.

 

Data protection laws around the world


Fifth edition Handbook by DLA Piper’s Data Protection and Privacy practice: “More than ever it is crucial that organisations manage and safeguard personal information and address their risks and legal responsibilities in relation to processing personal data, to address the growing thicket of applicable data protection legislation.

A well‑constructed and comprehensive compliance program can solve these competing interests and is an important risk‑management tool.

This handbook sets out an overview of the key privacy and data protection laws and regulations across nearly 100 different jurisdictions and offers a primer to businesses as they consider this complex and increasingly important area of compliance….(More)”

Infomediaries and accountability


Paper by Becky Carter: “A synthesis of what the existing evidence says (and where there are gaps) on:

1) What role might ‘infomediaries’, and specifically the media have in helping translate transparency into greater government accountability? In generating that accountability? In empowering citizens?

2) In what contexts or types of contexts do ‘infomediaries’ and media play such a facilitating role, and why?

3) What enabling factors contributed to success?

4) What role, if any, have donors had in supporting these sectors in this capacity?

5) What risks exist in this space?…(More)”

A ‘design-thinking’ approach to governing the future


Bronwyn van der Merwe at The Hill: “…Government organizations are starting to realize the benefits of digital transformation to reinvent the citizen experience in the form of digital services tailored to individual needs. However, public service leaders are finding that as they move further into the digital age, they need to re-orient their internal organizations around this paradigm shift, or their investments in digital are likely to fail. This is where Design Thinking comes into play.

Design Thinking has become a proven approach to reimagining complex service or organizational issues in the private sector. This approach of user research, rapid prototyping, constant feedback and experimentation is starting to take hold in leading business, like Citrix Systems, Ebay and Google, and is slowly spilling over into government bodies.

Challenges to Adopting a Design-Led Approach

Success in implementing Design Thinking depends on disrupting embedded organizational beliefs and practices, including cultural shifts, changing attitudes toward risk and failure, and encouraging openness and collaboration. Specifically, government bodies need to consider:

  • Top to bottom support – any change as wide-ranging as the shift to Design Thinking requires support from the top. Those at the top of design-led organizations need to be experimenters, improvisers and networkers who lead by example and set the tone for change on the ground.
  • Design skills gap – talent to execute innovation is in short supply and few governments are in a financial position to outbid private sector firms on pay. But the public sector does have something to offer that private companies most often do not: the ability to do meaningful work for the public good. Public sector bodies also need to upskill their current employees – at times partnering with outside design experts.
  • No risk, no reward – for government agencies, it can be challenging to embrace a culture of trial and error. But Design Thinking is useless without Design Doing. Agencies need to recognize the benefits of agile prototyping, iterating and optimizing processes, and that failings early on can save millions while costing little.

What Can Government Bodies Do to Change?

Digital has paved the way for governments and the private sector to occasionally partner to solve thorny challenges. For instance, the White House brought together the U.N. Refugee Agency and crowdfunding platform Kickstarter to raise money for the Syrian relief effort. The weeklong partnership raised nearly $1.8 million for more than 7,000 people in need.

But to effectively communicate with today’s digitally-enabled citizens, there are several key principals government bodies must follow:

  • Plain and simple – use simple language focused on content, structure, navigation, grouping and completion. Strip away the bureaucratic, government-speak and be transparent.
  • Take an outside-in design approach – by considering the entire ecosystem, and using research to uncover insights, service design reveals an outside-in view of the people in the entire ecosystem.
  • Be sensitive – too many government services, tools and processes are opaque and cumbersome when dealing with sensitive issues, such as immigration, making a tax submission, or adopting a child. Fjord recently took a human-centered design framework to the State of Michigan by designing a system that allowed caseworkers to convey the fairness of a child support order, while delivering excellent customer service and increasing transparency and accuracy to families in the midst of an emotionally-charged separation.
  • Work to digitize processes and services across departments – Governments should look to organize their digital services around the needs of the people – whether they are starting a business, retiring or having a child – rather than around their own departmental structures.
  • Address privacy concerns – The assurance of privacy and security is a critical step to encourage adoption of digital channels….(More)”

E-Government Strategy, ICT and Innovation for Citizen Engagement


Brief by Dennis Anderson, Robert Wu, Dr. June-Suh Cho, and Katja Schroeder: “This book discusses three levels of e-government and national strategies to reach a citizen-centric participatory e-government, and examines how disruptive technologies help shape the future of e-government. The authors examine how e-government can facilitate a symbiotic relationship between the government and its citizens. ICTs aid this relationship and promote transparencies so that citizens can place greater trust in the activities of their government. If a government can manage resources more effectively by better understanding the needs of its citizens, it can create a sustainable environment for citizens. Having a national strategy on ICT in government and e-government can significantly reduce government waste, corruption, and inefficiency. Businesses, CIOs and CTOs in the public sector interested in meeting sustainability requirements will find this book useful. …(More)”

CITI-SENSE: methods and tools for empowering citizens to observe acoustic comfort in outdoor public spaces


Itziar Aspuru et al in Noise Mapping: “The purpose of this research was to design and deploy tools that apply the concept of citizen observatories and empowering citizens in the assessment of acoustic comfort in public places. The research applies an iterative cycle of design and this article presents the results of a demonstrative experiment carried out in situ using the first products developed. This work was undertaken as part of the CITI-SENSE project. A viable technical and procedural solution was designed and tested in a field demonstration, where 53 people were engaged to provide 137 observations in the city of Vitoria-Gasteiz, using environmental sensors connected to a smartphone. The results have been analyzed and discussed in terms of the product’s attractiveness for engaging citizens in the evaluation of acoustic comfort in urban places, the accuracy of the noise levels measured by the acoustic app service integrated into the smartphone, and its ability to obtain simultaneous acoustic and perception data. The results presented in this article are considered a step forward in the research into developing solutions for assessing acoustic comfort. Limitations of the proposed solution are discussed, as are suggestions for further research….(More)”

Open Data and Beyond


Paper by Frederika Welle Donker, Bastiaan van Loenen and Arnold K. Bregt: “In recent years, there has been an increasing trend of releasing public sector information as open data. Governments worldwide see the potential benefits of opening up their data. The potential benefits are more transparency, increased governmental efficiency and effectiveness, and external benefits, including societal and economic benefits. The private sector also recognizes potential benefits of making their datasets available as open data. One such company is Liander, an energy network administrator in the Netherlands. Liander views open data as a contributing factor to energy conservation. However, to date there has been little research done into the actual effects of open data. This research has developed a monitoring framework to assess the effects of open data, and has applied the framework to Liander’s small-scale energy consumption dataset….(More)

Foundation Openness: A Critical Component of Foundation Effectiveness


Lindsay Louie at PhilanthroFiles: “We created the Fund for Shared Insight—a funder collaborative with diverse support from 30 different funders—to increase foundation openness. We believe that if foundations are more open—which we define as how they share about their goals and strategies; make decisions and measure progress; listen and engage in dialogue with others; act on what they hear; and share what they themselves have learned—they will be more effective.

WPhilanthropy Lessonse were so pleased to support Exponent Philanthropy’s video series featuring philanthropists being more open about their work: Philanthropy Lessons. To date, Exponent Philanthropy has released 5 of the total 9 videos, including:

Future video releases include:

  • Who Knows More? (expected 4/27/16)
  • Being Transparent (expected 4/27/16)
  • Value Beyond Dollars (expected 5/25/16)
  • Getting Out of the Office (expected 6/22/16)

We would love to see many more foundations make videos like these; engage in conversation with each other about these philanthropy lessons online and in person; share their experiences live at regional grantmaker association meetings or a national conferences like those Exponent Philanthropy hosts; and find other ways to be more open.

Why is this so important?

Recent research from the Center for Effective Philanthropy (report on CEP’s website here, full disclosure we funded this research) found that foundation CEOs see grantees, nonprofits that are considering applying for a grant, and other foundations working on similar issues as the top three audiences who benefit from a foundation being open about its work. Further, 86% of foundation CEOs who responded to the survey said they believe transparency is necessary for building strong relationships with grantees.

It was great to learn from this research that many foundations are open about their criteria for nonprofits seeking funding, their programmatic goals, and their strategies; and share about who makes decisions about the grantee selection process. Yet the research also found that foundations are not as open about sharing what they are achieving, how they assess their work, and their experiences with what has and hasn’t worked—and that foundation CEOs believe it would be beneficial for foundations to share more in these specific areas….(More)”

The Alberta CoLab Story: Redesigning the policy development process in government


Alex Ryan at Medium: “Alberta CoLab is an evolving experiment built on three counter-intuitive ideas:

1. Culture shifts faster through collaborative project work than through a culture change initiative.

2. The way to accelerate policy development is to engage more perspectives and more complexity.

3. The best place to put a cross-ministry design team is in a line ministry.

I want to explain what CoLab is and why it has evolved the way it has. We don’t view CoLab as a best practice to be replicated, since our model is tailored to the specific culture and context of Alberta. Perhaps you are also trying to catalyze innovation inside a large bureaucratic organization. I hope you can learn something from our journey so far,….

….Both the successes and frustrations of Alberta CoLab are consequences of the way that we have mediated some key tensions and tradeoffs involved with setting up a public sector innovation lab. Practitioners in other labs will likely recognize these tensions and tradeoffs, although your successes and frustrations will be different depending on how your business model reconciles them.

  1. Where should the lab be? Public innovation labs can exist inside, outside, or on the edge of government. Dubai The Model Centre and Alberta CoLab operate inside government. Inside labs have the best access to senior decision makers and the authority to convene whole of government collaborations, but may find it harder to engage openly with citizens and stakeholders. Unicef Innovation Labs and NouLab exist outside of government. Outside labs have more freedom in who they convene, the kind of container they can create, and timelines to impact, but find it harder to connect with and affect policy change. MindLab and MaRS Solutions Lab are examples of labs on the edge of government. This positioning can offer the best of both worlds. However, edge labs are vulnerable to fluctuations in their relationship with government. Surviving and thriving on the edge means continually walking a tightrope between autonomy and integration. Labs can change their positioning. Alberta CoLab began as an external consulting project. The Behavioural Insights Team is a social purpose company that was spun-off from a lab inside the U.K. government. The location of the lab is unlikely to change often, so it is an important strategic choice.
  2. How deep should the lab go? Here the tension is between taking on small, tactical improvement projects that deliver tangible results, or tackling the big, strategic systems changes that will take years to manifest. Public sector innovation labs are a reaction to the almost total failure of traditional approaches to move the needle on systems change.Therefore, most labs have aspirations to the strategic and the systemic. Yet most labs are also operating in a dominant culture that demands quick wins and measures success by linear progress against a simple logic model theory of change. We believe that operating at either extreme of this spectrum is equally misguided. We use a portfolio approach and a barbell strategy to mediate this tension. Having a portfolio of projects allows us to invest energy in systems change and generate immediate value. It allows us to balance our projects across three horizons of innovation: sustaining innovations; disruptive innovations; and transformative innovations. A barbell strategy means avoiding the middle of the bell curve. We maintain a small number of long-term, flagship initiatives, combined with a rapid turnover of quick-win projects. This allows us to remind the organization of our immediate value without sacrificing long-term commitment to systems change.
  3. What relationship should the lab have with government? Even an inside lab must create some distance between itself and the broader government culture if it is to provide a safe space for innovation. There is a tension between being separate and being integrated. Developing novel ideas that get implemented requires the lab to be both separate and integrated at the same time. You need to decouple from regular policy cycles to enable divergence and creativity, yet provide input into key decisions at the right time. Sometimes these decision points are known in advance, but more often this means sensing and responding to a dynamic decision landscape. Underneath any effective lab is a powerful social network, which needs to cut across government silos and stratas and draw in external perspectives. I think of a lab as having a kind of respiratory rhythm. It starts by bringing fresh ideas into the organization, like a deep breath that provides the oxygen for new thinking. But new ideas are rarely welcome in old organizations. When the lab communicates outwards, these new ideas should be translated into familiar language and concepts, and then given a subtle twist. Often labs believe they have to differentiate their innovations — to emphasize novelty — to justify their existence as an innovation lab. But the more the output of the lab resembles the institutional culture, the more it appears obvious and familiar, the more likely it will be accepted and integrated into the mainstream.
  4. What relationship should the lab have with clients? Alberta CoLab is a kind of in-house consultancy that provides services to clients across all ministries. There is a tension in the nature of the relationship, which can span from consulting problem-solver to co-design facilitator to teacher. The main problem with a consulting model is it often builds dependency rather than capacity. The challenge with an educational relationship is that clients struggle to apply theory that is disconnected from practice. We often use facilitation as a ‘cover’ for our practice, because it allows us to design a process that enables both reflective practice and situated learning. By teaching systemic design and strategic foresight approaches through taking on live projects, we build capacity while doing the work our clients need to do anyway. This helps to break down barriers between theory and practice, learning and doing. Another tension is between doing what the client says she wants and what she needs but does not articulate. Unlike a customer, who is always right, the designer has a duty of care to their client. This involves pushing back when the client demands are unreasonable, reframing the challenge when the problem received is a symptom of a deeper issue, and clearly communicating the risks and potential side effects of policy options. As Denys Lasdun has said about designers: “Our job is to give the client, on time and on cost, not what he wants, but what he never dreamed he wanted; and when he gets it, he recognizes it as something he wanted all the time.”

Lessons Learned

These are our top lessons learned from our journey to date that may have broader applicability.

  1. Recruit outsiders and insiders. Bringing in outside experts elevates the lab’s status. Outsiders are essential to question and challenge organizational patterns that insiders take as given. Insiders bring an understanding of organizational culture. They know how to move files through the bureaucracy and they know where the landmines are.
  2. Show don’t tell. As lab practitioners, we tend to be process geeks with a strong belief in the superiority of our own methods. There is a temptation to cast oneself in the role of the missionary bringing the good word to the unwashed masses. Not only is this arrogant, it’s counter-productive. It’s much more effective to show your clients how your approach adds value by starting with a small collaborative project. If your approach really is as good as you believe it is, the results will speak for themselves. Once people are envious of the results you have achieved, they will be curious and open to learning how you did it, and they will demand more of it.
  3. Be a catalyst, not a bottleneck. Jess McMullin gave us this advice when we founded CoLab. It’s why we developed a six day training course to train over 80 systemic designers across the government. It’s why we run communities of practice on systemic design and strategic foresight. And it’s why we publish about our experiences and share the toolkits we develop. If the innovation lab is an ivory tower, it will not change the way government works. Think instead of the lab as the headquarters of a democratic grassroots movement.
  4. Select projects based on the potential for reframing. There are many criteria we apply when we decide whether to take on a new project. Is it a strategic priority? Is there commitment to implement? Are the client expectations realistic? Can our contribution have a positive impact? These are useful but apply to almost any service offering. The unique value a social innovation lab offers is discontinuous improvement. The source of discontinuous improvement is reframing — seeing a familiar challenge with new eyes, from a different perspective that opens up new potential for positive change. If a project ticks all the boxes, except that the client is certain they already know what the problem is, then that already limits the kind of solutions they will consider. Unless they are open to reframing, they will likely be frustrated by a lab approach, and would be better served by traditional facilitation or good project management.
  5. Prototyping is just the end of the beginning. After one year, we went around and interviewed the first 40 clients of Alberta CoLab. We wanted to know what they had achieved since our co-design sessions. Unfortunately, for most of them, the answer was “not much.” They were very happy with the quality of the ideas and prototypes generated while working with CoLab and were hopeful that the ideas would eventually see the light of day. But they also noted that once participants left the lab and went back to their desks, they found it difficult to sustain the momentum and excitement of the lab, and easy to snap back to business as usual. We had to pivot our strategy to take on fewer projects, but take on a greater stewardship role through to implementation.
  6. Find a rhythm. It’s not useful to create a traditional project plan with phases and milestones for a non-linear and open-ended discovery process like a lab. Yet without some kind of structure, it’s easy to lose momentum or become lost. The best projects I have participated in create a rhythm: an alternating movement between open collaboration and focused delivery. The lab opens up every few months to engage widely on what needs to be done and why. A core team then works between collaborative workshops on how to make it happen. Each cycle allows the group to frame key challenges, make progress, and receive feedback, which builds momentum and commitment.
  7. Be a good gardener. Most of the participants of our workshops arrive with a full plate. They are already 100% committed in their day jobs. Even when they are enthusiastic to ideate, they will be reluctant to take on any additional work. If we want our organizations to innovate, first we have to create the space for new work. We need to prune those projects that we have kept on life support — not yet declared dead but not priorities. This often means making difficult decisions. The flip side of pruning is to actively search for positive deviance and help it to grow. When you find something that’s already working, you just need to turn up the good…..(More)”

The sharing economy comes to scientific research


 at the Conversation: “…to perform top-quality and cost-effective research, scientists need these technologies and the technical knowledge of experts to run them. When money is tight, where can scientists turn for the tools they need to complete their projects?

Sharing resources

An early solution to this problem was to create what the academic world calls “resource labs” that specialize in one or more specific type of science experiments (e.g., genomics, cell culture, proteomics). Researchers can then order and pay for that type of experiment from the resource lab instead of doing it on their own.

By focusing on one area of science, resource labs become the experts in that area and do the experiments better, faster and cheaper than most scientists could do in their own labs. Scientists no longer stumble through failed experiments trying to learn a new technique when a resource lab can do it correctly from the start.

The pooled funds from many research projects allow resource labs to buy better and faster equipment than any individual scientist could afford. This provides more researchers access to better technology at lower costs – which also saves taxpayers money, since many grants are government-backed….

Connecting people on a scientific Craigslist

This is a common paradox, with several efforts under way to address it. For example, MIT has created several “remote online laboratories” running experiments that can be controlled via the internet, to help enrich teaching in places that can’t afford advanced equipment. Harvard’s eagle-i system is a directory where researchers can list information, data and equipment they are willing to share with others – including cell lines, research mice, and equipment. Different services work for different institutions.

In 2011, Dr. Elizabeth Iorns, a breast cancer researcher, developed a mouse model to study how breast cancer spreads, but her institution didn’t have the equipment to finish one part of her study. My resource lab could complete the project, but despite significant searching, Dr. Iorns did not have an effective way to find labs like mine.

Actively connecting scientists with resource labs, and helping resource labs keep their equipment optimally busy, is a model Iorns and cofounder Dan Knox have developed into a business, called Science Exchange. (I am on its Lab Advisory Board, but have no financial interest in the company.) A little bit Craigslist and Travelocity for science rolled into one, Science Exchange provides scientists and expert resource labs a way to find each other to keep research progressing.

Unlike Starbucks, resource labs are not found on every corner and can be difficult for scientists to find. Now a simple search provides scientists a list of multiple resource labs that could do the experiments, including estimated costs and speed – and even previous users’ reviews of the choices.

I signed onto Science Exchange soon after it went live and Iorns immediately sent her project to my lab. We completed the project quickly, resulting in the first peer-reviewed publication made possible through Science Exchange….(More).