Subword Variation in Text Message Classification data sheet 1892 Views
Author:
Robert Munro and Christopher Manning
Publication Date:
Jun 2010
Publication Type:
Journal article
Abstract:
For millions of people in less resourced regions of the world, text messages (SMS) provide the only regular contact with their doctor. Classifying messages by medical labels supports rapid responses to emergencies, the early identification of epidemics and everyday administration, but challenges include text-brevity, rich morphology, phonological variation, and limited training data. We present a novel system that addresses these, working with a clinic in rural Malawi and texts in the Chichewa language. We show that modeling morphological and phonological variation leads to a substantial average gain of F=0.206 and an error reduction of up to 63.8% for specific labels, relative to a baseline system optimized over word-sequences. By comparison, there is no significant gain when applying the same system to the English translations of the same texts/labels, emphasizing the need for subword modeling in many languages. Language independent morphological models perform as accurately as language specific models, indicating a broad deployment potential.
During the last FailFaire (well, actually the first one to date) we were reminded by a guest rather sternly that NGOs often forget who their users are and, more importantly, what the needs of these users are. When we, collectively as a field, implement mobile deployments with constituents or groups, do we tend to forget user needs and capabilities, getting too enchanted with the tech (and ourselves) and then fail when, not surprisingly, there isn't any uptake? We have seen many a project fail for precisely this reason. Our grouchy attendee had a point even if he did not deliver it very gracefully.
To this end, we are reposting here a recent report from Zambia. Project Mwana is UNICEF Innovation project that is "working with the Zambian Ministry of Health, UNICEF Zambia, the Malawi Ministry of Health, UNICEF Malawi and many implementing and technical partners to find appropriate, scalable and impactful ways that mobile technologies can strengthen health services for mothers and infants in rural health clinics."
As part of a "Mobile Telemedicine" initiative undertaken by the Millennium Villages Project in Ghana, I have been researching and documenting existing software platforms that enable and support remote consultation activities.
Telemedicine is the use of medical information exchanged from one site to another via electronic communications to improve patients’ health status or for educational purposes. It includes consultative, diagnostic, and treatment services.
Mobile health information technology (mHealth) typically refers to portable devices with the capability to create, store, retrieve, and transmit data in real time between end users for the purpose of improving patient safety and quality of care.
On Wednesday evening I was lucky enough to attend the first ever "Failfaire", organized by MobileActive.org where several brave souls agreed to present their failed "Information Technology for Development" projects, explaining why they failed and what they learned from them.
I work on knowledge management in UNICEF, and have a strong interest in improving how we learn from our experience. This event (which was certainly not a failure!) was interesting to our work from at least two points of view:
1. The lessons learned from the projects themselves
2. The idea for the event itself and whether this might be something we could try ourselves.
There were four presentations during the meeting:
Bradford Frost presented on Mobileimpact.org a project to recycle old cellphones and donate them to Africa.
MobileActive hosted the inaugural FAILfaire last night, bringing together mobile technologists and NGOs to talk about failed projects in M4D and ICT4D. Presenters talked about their failed projects, answering the questions: "What was the project? What was the failure? Why did it fail? And what would you do differently next time?”
The event was filled to capacity with more than 70 people. The five presenters made us think (and laugh), and the audience asked some great questions. For those of you who couldn’t be there, here’s a quick look at the failed projects presented at the first (of what we hope will be many) FAILfaire.
Bradford Frost: MobileImpact.org? Not exactly...
Starting off the evening was Bradford Frost, who told the story of his failed non-profit venture, MobileImpact.org. The goal of his project was to bridge the gap between people trying to recycle used phones and developing countries. He felt he had a strong idea and a strong brand with the tagline “One phone. Change the World,” and that there was enough of an untapped phone recycling market (the current cell phone recycling market only captures about 25% of reusable devices) for the project to succeed.
However, the project didn’t work out as Frost had hoped. He used Facebook ads in order to target a younger, social media-savvy audience. He spent 1,000 dollars to launch an ad campaign and $5000 in a partnership with a phone recycling company. In the end, the non-profit gathered 131 phones valued at a sum total of …$252. And many of those phones were donated through word-of-mouth connections (friends and family) rather than people who saw the Facebook ads.
As we here at MobileActive.org have been covering ICT and mobiles for development now for more than five years, we have seen our fair share of failures. For every great project that changes how a community benefits from technology to improve the lives of its people, there seem to be twice as many projects that fail, and end up wasting time, money, and maybe worst, goodwill.
Too often in our field, we talk up our successes, overhype and overestimate the value of our projects, and sweep the failures under the rug. But, if we don’t talk about what didn’t work (and, perhaps more importantly, why it didn’t work), others will keep repeating the same mistakes.
That is why we invented FailFaire, a gathering that is happening tonight in New York City and that we hope will take place in other cities around the world. FailFaire is a place where it's ok to talk about what didn't work to learn from for the next project using mobiles for social change and development.
Spring has us ready to get out and share new ideas, so it’s time for our monthly event roundup. Below are some of the mobile events happening around the world that we think might be of interest to the MobileActive.org community. If you know of any great M4D events happening near you, let us know over email (info at mobileactive dot org) or in the comments or on our Facebook page.
FAILfaire, April 14 New York, NY: Of course, we have to start off with promoting our own FAILfaire, which is taking place on April 14th, 5:00-7:00 p.m., in New York City. At FAILfaire, we want to learn from each others’ failures in ICT and mobiles for development – speaking openly about what didn’t work is the best way to improve for the future. Gather round with some wine and a sense of humor, and get ready to talk about FAIL! Sign up here.
Calling all failures! On April 14 we're hosting the first FAILfaire featuring a close look at ICT and mobile development projects that have crashed, burned, and simply FAILED.
While we often focus on highlighting successes and gains in this field, it's no secret that many projects just don't work - some aren't scalable, some aren't sustainable, some can't get around bureaucratic hoops, and many fail due to completely unanticipated barriers. FAILfaire is a platform to openly and honestly discuss failures so that we can learn from what hasn't worked in the past in order to make our future projects stronger and better.
This is where YOU come in. Have you been a part of a project that flopped? Maybe the project used the wrong technology for its region. Maybe it didn't engage the intended community. Did not take culture, people, or both into consieration. Or maybe the rollout was too rushed. Whatever the reason, we want to hear from you.
GUIDE: Getting Medical Information into the Hands of Community Health Workers data sheet 4903 Views
For community health workers who are far from medical libraries and urban centers, staying up-to-date on the latest medical advancements and disease treatments can be difficult. GUIDE, developed by AED-Satellife, is a conversion and content management system that transforms medical literature into HTML forms that are easily accessible over a mobile phone.
GUIDE currently runs on smartphones (Samsung I780) in a pilot program in Port Elizabeth, South Africa. For the last year, 50 nurses from three different branches of local hospitals have used mobiles to stay up to date on medical developments.
Can text messages help stop the spread of HIV/AIDS? Text to Change (TTC), a Ugandan NGO, thinks the
answer is yes. To prove it they have teamed up with Zain to launch an innovative project that utilizes a technology that has been rapidly expanding to make sure that HIV/AIDS does not do the same.
Text at Work and Stay Healthy: HIV/AIDS Prevention in the Workplace data sheet 1134 Views
Author:
Hajo van Beijma
Publication Date:
Feb 2010
Publication Type:
Report/White paper
Abstract:
Through a pilot partnership with the Text to Change (TTC) initiative, HIPS is using an innovative tool to support companies in providing key prevention messages and education on HIV/AIDS to their employees and community networks – text messages. Early results are promising: the launch of the program resulted in a 40% increase in demand for sexual and reproductive health services from participant company clinics.
Getting the Word Out About HIV: Imbizo Men's Health Program data sheet 4232 Views
(This case study is reprinted with permission from Glen Thompson of BulkSMS.com)
HIV support programme uses SMS to keep men connected
This case study illustrates that the use of appropriate mobile technology has a positive impact on maintaining men’s participation in a health programme focusing on HIV testing and education.
Recent research findings on the Imbizo Men’s Health Programme indicate that its male participants responded positively to the use of SMS communications. SMS was used to remind participants to go for HIV testing and inform them of upcoming events promoting AIDS and gender awareness. Taken together, this use of mobile technology has played an important part in ensuring that men have remained involved in this public health initiative.
ChildCount: Monitoring Children's Health Through SMS data sheet 7164 Views
Many mobile projects struggle with scale and impact. While a mobile health project may run well with a small number of patients in one hospital, expanding the scope of a project until it is large enough to have real impact takes money, time, and widespread support of key stakeholders in a given community. ChildCount is well on its way to show scale and, so we hope, significant health impacts using mobile technology for patient support.
In a little over eight months, ChildCount has enrolled nearly 10,000 children under five in their catchment area into the ChildCount health monitoring system – an acceptance rate of more than 95%.
Enabling Data-Driven Decisions with the Open Data Kit (ODK) data sheet 5465 Views
A research group at the University of Washington has done what few others manage – turn a research project into a real-world application. Open Data Kit (ODK) is a collection of tools that allows organizations to collect and send data using mobile phones. The system, in operation for about a year, has already been used for projects such as counseling and testing HIV patients in Kenya to monitoring forests in the Brazilian Amazon.
What is ODK?
The project began when University of Washington (UW) professor Gaetano Borriello began a sabbatical at Google to build a mobile data collection system. He brought along some of his PhD students from UW’s Computer Science and Engineering program to work on the idea as their intern project, and ODK was born.
The project is two-fold. One service targets what Tim Wood, director of the Grameen Foundation’s Mobile Health Innovation and ICT Innovation divisions calls “pregnant parents,” and another targets community health workers.
Mobile Phones' Potential to Address Information and Communication Needs of Healthcare Workers in Isolated Rural Areas in Peru data sheet 2630 Views
Author:
David Crespo
Publication Date:
Jan 2009
Publication Type:
Other
Abstract:
The aim of this paper is to present the main findings of the interviews and observation undertaken in health locations in rural areas of Peru, aimed at exploring the application of mobile technology for health in isolated areas, and thereby to provide wider reflections about the use of mobile phones to improve health delivery in isolated areas in developing countries. It also reflects on methodological complexities, particularly those related to the application of online research methods, through the experience gained during the research.
The great growth of the penetration of mobile phones in developing countries has attracted the attention of the ICT4D community and numerous publications reflect about their impact for development (see for example Donner, 2005; Vodafone, 2004; Sullivan, 2007 or Scott et al., 2004). A part of this literature is focused on the application of mobile phones for health delivery (bridges.org, 2005; Katz, 2005 or Voxiva, 2005). Nevertheless, publications about the application of mobile phones for health focused in rural areas of developing countries are still scarce. This paper focuses on this gap of the literature, focusing on rural areas of Peru and wider reflecting about the behavior of mobile phones in this specific context.
Given the evolution and forecasts of penetration of mobile phones in developing countries (see Gartner, 2006 or Donner, 2005) as well as the appealing of these devices for the population (Castells et al.,2007), it is important to remark that Health Ministries in developing countries and development agencies have not paid enough attention to the possibilities that mobile technology brings for health delivery.
This paper reports on the interviews to healthcare workers undertaken in rural areas of Peru, aimed at exploring their perception about how mobile phones can contribute to the resolution of their information and communication needs, as well as provide wider reflections about the general debate about the benefits of mobile phones for health for development. The workers were chosen as participants in two innovative and contrasting ICT4D projects: CellPREVEN and EHAS, seeking at obtaining valuable feedback about their experience on one side, and their opinion as healthcare workers with no previous experience on the use of mobile telephony within an ICT4D initiative, on the other. Additionally, the paper aims at exploring the complexities involved in the methodology, particularly the difficulties in the appliance of online research methods during the research.
The article is structured as follows: first, the context of the research is presented; after that, some methodological issues about the field work in Peru are discussed. Next, the most remarkable findings of the interviews are presented, linked to wider reflections about the benefits of mobile phones in rural areas of developing countries. These results highlight three important advantages provided by mobile phones: the value of the immediacy in the communication for emergencies, the potential of multimedia capabilities of mobile phones and the improved reliability comparing to other technologies, but they also reflect on the recent limitations for their adoption and use in isolated communities. The article ends concluding that the application of mobile phones for health delivery is clear and of great impact in isolated areas of developing countries, but limited at this moment, with great potential in the short term.
Safe drinking water is a necessity for life. But according to a 2005 report published by the World Health Organization and UNICEF, 1.1 billion people worldwide lack access to quality drinking water.
In South Africa, a current project is monitoring water quality with SMS in a push to bring safer water to the area. Run by the University of Bristol and the University of Cape Town, the four year project is two-fold: 1) develop a new means of testing water quality and 2) develop a new means of reporting the results of these water quality tests.
Aquatest, the water quality testing system, is still under development, but the Water Quality Reporter is up and running – on mobile phones with reporting via SMS. The application allows field workers to cheaply and effectively transfer data about water quality to a centralized database, while receiving feedback about how to handle local water problems.
Says Melissa Loundon, a researcher at the University of Cape Town who worked with the development of the Water Quality Reporter,
“The main part of the project is to develop the water test. But the original project team at the University of Bristol realized that if you’ve got a water test that can be used by people who aren’t in the field, or people who aren’t specialists, it doesn’t really help them if they get a result and see that their water is not safe to drink. They may not have a whole lot of resources to do anything about it. So the point of the cell phone application is that once somebody has a result, they can communicate it to a central database and also to somebody in the area who can provide support.”
The Vodafone Americas Foundation is announcing the last call for nominations for the second annual Wireless Innovation Project, a competition to identify and reward the most promising advances in wireless related technologies that can be used to solve critical problems around the globe. Proposals will be accepted through February 1, 2010, with the final winners announced on April 19, 2010 at the annual Global Philanthropy Forum in Redwood City, California.
By using local numbers and relaying them through VOIP Rebtel can offer people low-cost international calls. The only cost for the user will be for calling a local number. Rebtel is busy launching SMS functions for free or low cost also.
Tool Category:
Is a web-based application/web service
Key Features :
Enables low-cost international calls.
Main Services:
Mobile Social Network/Peer-to-peer
Display tool in profile:
Yes
Tool Maturity:
Currently deployed
Release Date:
2010-01
Platforms:
Mac/Apple
Windows
Windows Mobile
All phones -- SMS
All phones -- Voice
All phones/Mobile Browser
Program/Code Language:
Java/Android
Objective C/iPhone
Organizations Using the Tool:
We are currently working with Hand-in-Hand to be able to develop better communications for people in need.
Information and Communication Technologies (ICTs) for social and economic development in emerging economies have long been a focus of governments, the private sector, and most certainly donors and international development agencies. Yes, despite all the attention garnered on this field, we are seeing a checkered history of ICTs as a tool for development, with both successes and significant failures littering the landscape.
With the phenomenal growth of mobile technology in the last ten years, the attention of donors, governments, and multi-lateral and international agencies has now turned to the telecommunications sector and mobile technologies as channels to deliver services and products to citizens at the bottom of the economic pyramid.
"Innovations in Mobile Data Collection for Social Action," a workshop co-hosted by MobileActive.org and UNICEF in Amman, Jordan, featured Ignite Talks -- five minute presentations by inspiring people who are using mobiles for social action in the Middle East -- and interviews with key participants. Jacob Korenblum describes the work of Souktel in Palestine, and Erica Kochi from UNICEF Innovation, the co-host of the event, illustrates why data collected by mobiles is so important for their work in Iraq.
We will be blogging and twittering this week from a workshop we are co-hosting on Innovations in Mobile Data Collection for Social Action in Amman, Jordan.
Co-hosted by UNICEF’s country office in Iraq, UNICEF Innovation, and MobileActive.org, this three-day gathering is bringing invited experts from around the world together to explore some of the key issues related to using mobiles for data collection and analysis of some of the toughest social issues.
Why are we hosting this event?
With the ubiquity of mobile technology, data collection and monitoring of key indicators from the ground up by affected populations is now possible. Mobile technology in the hands of people can now be more than a person-to-person communication medium but can be used for capturing, classifying and transmitting image, audio, location and other data, interactively or autonomously.
Two of the fastest-growing and popular mobile data collection tools have recently seen some exciting upgrades in newly released versions.
Open Data Kit recently released v1.1 of ODK Collect. Open Data Kit (ODK) is a suite of tools to help organizations collect, aggregate and visualize their data. ODK Collect is powerful phone-based replacement for paper forms. Collect is built on the Android platform and can collect a variety of form data types: text, location, photos, video, audio, and barcodes. ODK Collect can be downloaded in the Android marketplace or here. The developers also have a demo video that describes the new features of the release. Open Data Kit is a member of the Open Mobile Consortium of which MobileAtive.org is a founding member.
Some of the new features of ODK Collect include barcode scanning, image/audio/video capture and playback, editing of saved forms, and device metadata (phone number, IMEI, IMSI) support. GPS acquisition and form processing is a faster, and the developers added review data entry. The user interface has been field tested and reworked to make training and use much easier. ODK Collect also supports question grouping, repeats, constraints, complex logic, and multiple languages.
ODK is currently deployed for HIV counseling with AMPATH in Kenya, user feedback gathering for Grameen's AppLab in Uganda, war crime documentation with the Berkeley Human Rights Center in the Central African Republic, and forest monitoring with the Brazilian Forest Service.
Meanwhile, our friends over at Datadyne have released version 2.0 of their popular mobile data collection platform EpiSurveyor. For some of the very cool GPS features of that, see the video below. EpiSurveyor is a free, user-friendly mobile-phone-and-web-based data collection system. Version 2.0 has many new features such as GPS (users with GPS-enabled phones (like the Nokia E71) can automatically create a "GPS stamp" for every record collected AND automatically see the results on a Google map, all within EpiSurveyor.org), advanced logic, including skip logic; numeric range limits for data entry; and a much better user interface for the web-backend.
EpiSurveyor is used by organizations around the world. One organization, TulaSalud in Guatemala, uses EpiSurveyor for maternal health. The video below (en Espanol) explains how the organization is using the tool.
Video informativo de TulaSalud, sobre la aplicación del sistema de monitoreo epidemiológico aplicado con la tecnología de EpiSurveyor, el cual pretende tener a tiempo real el reporte epidemiológico de las comlunidades.
Mobile phones are becoming prolific in society, both in industrialized and non-industrialized countries. Of all of the developments in new media, mobile is the only one that seems to be narrowing the digital divide rather than widening it, making it a crucial tool in improving the health of poor and hard-to-reach populations. Organizations around the world are beginning to implement mobile technology into their health behavior interventions and are seeing rising success. In diabetes and other chronic disease management, mobile phones are being used for medication alerts and health reporting. The technology is being used to send sexual health and HIV/AIDS information in places like downtown San Francisco and rural Africa, as part of ongoing testing, counseling and prevention services. In the US, applications involving GPS technology, video games, and persuasive technology are all being tested for their effectiveness in influencing health behavior and applicability on mobile phones.
Through published research and interviews with key leaders in the mobile health field, this paper outlines the benefits of mobile technology and the barriers to integrating mobile fully into public health campaigns. Despite the enormous work already being done and a host of new technologies on the rise, most experts will agree that mobile phones on their own will not serve as primary tools for influencing health behavior. A comprehensive and integrated campaign that incorporates mobile will be most effective, but mobile alone faces too many barriers to fostering social norms on the wide-scale. In fact, many argue that the ways in which public health advocates are able to place health messages in broadcast media cannot be applied on cell phones because of society’s growing resistance to mobile marketing.
The paper asks: What are some examples of how organizations around the world have harnessed mobile technology to improve the health behavior of disease-specific and target populations? What lies in the future for mobile technology and health? Is mobile the answer we have been looking for?
Mobile Web for Social Development Roadmap data sheet 3183 Views
Author:
Mobile Web For Social Development
Publication Date:
Nov 2009
Publication Type:
Other
Abstract:
This document aims to help readers understand the current challenges of deploying development-oriented services on mobile phones, evaluate existing technologies, and identify the most promising directions to lower the barriers of developing, deploying and accessing services on mobile phones and thereby creating an enabling environment for more social-oriented services to appear.
This document is divided into two major parts. The first part presents the major challenges today for both developing and accessing mobile services, potential ways to bridge them with existing tools, technologies and infrastructure, and potential research directions to follow to provide a more comprehensive resolution or solution. The second part focuses on presenting the major technologies and the major options existing today to deploy content and applications on mobile phones. For each of these technologies, the document presents a short analysis of the technology's potential and the requirements in terms of infrastructure, devices, targeted end-users, and costs associated with implementation and delivery.
