Transforming the Industry: Watson in Education

Watson, named after IBM founder Thomas J. Watson, was built by a team of IBM scientists with valuable help from research partners from Carnegie Mellon University, University of Texas, University of Southern California, University of Massachusetts, University of Trento (Italy), MIT, RPI, and the University of Albany. The team set out to accomplish a grand challenge—to build a computing system that rivals a human’s ability to answer questions posed in natural language with speed, accuracy and confidence. Watson passed its first test on Jeopardy! in February 2011, but the real test will be in applying the underlying systems, data management and analytics technology across different industries, especially in education.

Invited panel presentation at IBM Watson in Education: Transforming the Industry, IBM Almaden Research Center, November 16, 2011

Crowdsourcing: From Phenomenon to Business Model

Crowdsourcing is changing both the way we work, as well as the way Internet applications are designed and delivered. It’s no longer just the domain of technologists (who can now achieve breakthroughs together never before achievable); crowdsourcing is now ripe for enterprise professionals to understand and leverage the possibilities for their business goals.

From Wikipedia and YouTube, crowdsourcing has moved to a $5B “crowd worker” industry with applications proliferating for productivity, research, marketing, advertising, creative development, corporate workflow management, language translations, and much more — see a list of projects here.

I discuss social networks as a kind of crowdsourcing, with unique benefits and challenges.

Invited panel presentation at PARC Forum, Palo Alto, CA, November 10, 2011

 

View the panel discussion:

www.parc.com/event/1456/crowdsourcing-ceo-expert-panel.html

Case-Based Reasoning Research And Development

This book constitutes the refereed proceedings of the 19th International Conference on Case-Based Reasoning, held in London, UK, in September 2011. The 32 contributions presented together with 3 invited talks were carefully reviewd and selected from 67 submissions. The presentations and posters covered a wide range of CBR topics of interest both to practitioners and researchers, including CBR methodology covering case representation, similarity, retrieval, and adaptation; provenance and maintenance; recommender systems; multi-agent collaborative systems; data mining; time series analysis; Web applications; knowledge management; legal reasoning; healthcare systems and planning systems.

Find the book:

Case-Based Reasoning Research and Development | Lecture Notes in Artificial Intelligence, Vol. 6880

edited by Ashwin Ram and Nirmalie Wiratunga

Springer, October 20, 2011, ISBN 978-3-642-23290-9

www.springer.com/computer/ai/book/978-3-642-23290-9

Robust and Authorable Multiplayer Storytelling Experiences

Interactive narrative systems attempt to tell stories to players capable of changing the direction and/or outcome of the story. Despite the growing importance of multiplayer social experiences in games, little research has focused on multiplayer interactive narrative experiences. We performed a preliminary study to determine how human directors design and execute multiplayer interactive story experiences in online and real world environments. Based on our observations, we developed the Multiplayer Storytelling Engine that manages a story world at the individual and group levels. Our flexible story representation enables human authors to naturally model multiplayer narrative experiences. An intelligent execution algorithm detects when the author’s story representation fails to account for player behaviors and automatically generates a branch to restore the story to the authors’ original intent, thus balancing authorability against robust multiplayer execution.

Read the paper:

Robust and Authorable Multiplayer Storytelling Experiences

by  Mark Riedl, Boyang Li, Hua Ai, Ashwin Ram

in Seventh International Conference on AI and Interactive Digital Entertainment (AIIDE-2011).

www.cc.gatech.edu/faculty/ashwin/papers/er-11-06.pdf

Social Media for Health and Wellness 2.0

The Internet has surpassed physicians as the leading source of health information. With the advent of the social web, Health 2.0 is emerging as a strong segment with 34% of consumers using social resources such as blogs and forums to locate health information. Yet information overload leads to “search engine fatigue” that discourages users.

We advocate a consumer-centric approach to healthcare information access that increases engagement and improves health literacy.  Artificial Intelligence (AI) techniques can be used to support human effort, creating a new generation of “intelligent web” technologies. These technologies can combine the benefits of the “information web” (timely, relevant health information) with those of the “social web” (human interaction, support, comfort). Our vision is to promote well-being and prevention before illness, support and information during illness, and comfort to family and friends in a natural, social, yet private manner.

Invited talk at Humana Innovation Conference: Connect, Collaborate, Create (C3), Louisville, KY, September 23, 2011.

SoCS Computational Models and Techniques: A Case Study

I spoke today at the NSF Workshop on Social-Computational Systems (SoCS) on Mike Pazzani‘s Computational Models and Techniques panel with Tuomas Sandholm, Lise Getoor, and Tina Eliassi. We were asked to address the questions of what computation can teach us about socially intelligent systems, and what problems are encountered when applying existing technologies to such systems.

I focused on two key SoCS challenges : impedance mismatch, and research-at-scale. Let me explain.

What can computation teach us about SoCS? If we begin with technology, we’ll encounter the key challenge of “impedance mismatch” between people and technology. The technology, however good, may not address people’s needs. Instead, let’s reverse the question: What do socially intelligent systems teach us about computational technology?

Consider, as a case study, the problem of education: building a SoCS system to help students learn. Our first pass was a collaborative learning site with a state-of-the-art collaboration platform, a kind of “Google Docs meets WebEx meets Etherpad meets Skype on steroids”. While the site was useful, we learned that students didn’t use most of the features we had built. The issue was impedance mismatch: the technology did not address education problems from a student perspective.

What, then, are these problems? There are two: Access (scale) and engagement. To tackle the impedance mismatch, we need to design technology that provides the right affordances (in the Gibsonian sense) for student behaviors that address those problems.

We created a vision for Open Social Learning that blends, not Google Docs and WebEx, but Facebook and World of Warcraft. With funding from NSF, NIH, GRA, and Gates/Hewlett NextGenLC, and partnerships with MIT, Yale, NYU, and many others, we rethought the site from Education to SoCS to Learning Theories to Design Principles to Affordances to Architecture to User Experience (UX) to Mechanisms. (See slides and references below.) This process resulted in a fundamentally disruptive idea, one driven not by technology but by the SoCS it was to support.

Only then did it make sense to think about Computation: really real-time collaboration technologies for a highly interactive experience; intelligent recommender systems to help learners connect with relevant content and other learners; mining and analytics to assess learner outcomes; and reputation techniques to establish social capital.

The new OpenStudy.com is an Open Peer-to-Peer Social Learning Community, a place that matches learners studying the same things into live “massively multiplayer study sessions“. The problems of access (scale) and engagement are addressed through two mechanisms: A Luis von Ahn approach where the social community scales itself, and a kind of gamification in which everyone is on the same team.

Great idea—but how do we know it works? The education literature is full of great ideas that don’t work in practice. SoCS data research involves studying large-scale communities; the same applies to SoCS technology design. This is the research-at-scale challenge. Laboratory studies don’t prove much; the research fundamentally requires scale.

After 9 months, OpenStudy has grown into a vibrant community that both provides value to its users and serves as a “living lab” to study and validate the ideas. We’re continuing to research how new technologies can be combined to address the problem of education in a manner that is highly scalable yet interactive and engaging.

To understand what socially intelligent systems teach us about computation, then, requires a new methodology comprised of old ideas about design thinking brought into the new world of Social-Computational Systems at a massive scale.

READINGS

P Adams (2009). Designing for Social Interactions.

Terry Anderson (2007). Distance Learning: Social Software’s Killer App?

J Daniel (1996), cited in JS Brown (2007). Minds on Fire: Open Education, The Long Tail, and Learning 2.0.

RA DeMillo (2011). Abelard to Apple: The Fate of American Colleges and Universities in the Twenty-First Century.

R Friedrich, M Peterson, A Koster (2011). The Rise of Generation C.

Gates Foundation study: JM Bridgeland, JJ Dilulio Jr, KB Morrison (2006), The Silent Epidemic: Perspectives of High School Dropouts.

D Thomas & JS Brown (2011). A New Culture of Learning.

More readings at: Massively Multiplayer Online—Learning?

SLIDES

Open Social Learning Communities

With the advent of open education resources, social networking technologies and new pedagogies for online and blended learning, we are in the early stages of a significant disruption in current models of education. The disruption is fueled by a staggering growth in demand. It is estimated that there will be 100 million students qualified to enter universities over the next decade. To educate them, a major university would need to be created every week.

Universities have responded to this need with Open Education Resources—thousands of free, high quality courses, developed by hundreds of faculty, used by millions worldwide. Unfortunately, online courseware does not offer a supporting learning experience or the engagement needed to keep students motivated. Students read less when using e-textbooks; video lectures are boring; and retention and course completion rates are low.

Therein lies the core problem: How to engage a generation of learners who live on the Internet yet tune out of school, who seek interaction on Facebook yet find none on iTunes U, who need community yet are only offered content. We propose a new approach to this problem: open social learning communities, anchored with open content, providing an interactive online study group experience akin to sitting with study buddies on a world-wide campus quad.

This solution is enabled by state-of-the-art web technologies: really real-time collaboration technologies for a highly interactive experience; intelligent recommender systems to help learners connect with relevant content and other learners; mining and analytics to assess learner outcomes; and reputation techniques to establish social capital.  We will discuss these technologies and how they can be combined to address the problem of education in a manner that is highly scalable yet interactive and engaging.

This approach can be used for other types of learning communities. We will show an application to healthcare information access to help consumers learn about their healthcare questions and needs.

Keynote talk at SIPA Conference: Entrepreneurship—Idea Wave 3.0, Mountain View, CA, November 12, 2011.

 

Keynote talk at the International Conference on Web Intelligence, Mining and Semantics (WIMS-11), Sogndal, Norway, May 27, 2011.

 

View the talk:

videolectures.net/wims2011_ram_learning

Read the paper:

www.cc.gatech.edu/faculty/ashwin/papers/er-11-04.pdf

View the slides:

 

 

 

Towards A National Study Guild

New post on blog@CACM: My presentation to President Obama’s Science & Technology advisory council (PCAST) on Education.

“Imagine a Facebook where the point is to study together, not trade pictures and jokes. Imagine a World of Warcraft where students earn levels and points by helping each other learn. Not a video game that teaches physics; instead, let’s create an educational experience that is social and game-like.”

READ IT HERE:
cacm.acm.org/blogs/blog-cacm/109290-our-big-idea-open-social-learning/

SLIDES:

A Case Base Planning Approach for Dialogue Generation in Digital Movie Design

We apply case based reasoning techniques to build an intelligent authoring tool that can assist nontechnical users with authoring their own digital movies. In this paper, we focus on generating dialogue lines between two characters in a movie story. We use Darmok2, a case based planner, extended with a hierarchical plan adaptation module to generate movie characters’ dialogue acts with regard to their emotion changes. Then, we use an information state update approach to generate the actual content of each dialogue utterance. Our preliminary study shows that the extended planner can generate coherent dialogue lines which are consistent with user designed movie stories using a small case base authored by novice users. A preliminary user study shows that users like the overall quality of our system generated movie dialogue lines.

Read the paper:

A Case Base Planning Approach for Dialogue Generation in Digital Movie Design

by Sanjeet Hajarnis, Christina Leber, Hua Ai, Mark Riedl, Ashwin Ram

19th International Conference on Case-Based Reasoning (ICCBR-11), London. 

www.cc.gatech.edu/faculty/ashwin/papers/er-11-05.pdf

Open Social Learning aka Massively Multiplayer Online Learning

“Our Open Social Learning solution to these three problems of education is therefore elegantly simple. In this solution OpenCourseWare courses are augmented by a community of learners who help one another, support one another and learn together as they socialize and spend time together online. Not only is this solution validated by educational research, it is also eminently scaleable because you are not dependent on hiring tutors or teachers to spend time assisting self learners.  The community helps one another.  Open Social Learning also fits right in with what our digital millenials want to do: hang out online for hours!  Why not get them talking about math instead of … well, let’s not go there.” — Dr. Preetha Ram

P. Ram, A. Ram, C. Sprague (2011). Socializing OpenCourseWare with OpenStudy. OCWC Global 2011: Celebrating 10 years of OpenCourseWare, Cambridge.

View it here: preetharam.wordpress.com/2011/05/11/open-social-learning-aka-massively-multiplayer-online-learning