Augmenting Human Innovation with Social Cognition

Social Media is everywhere: photos, videos, news, blogs, art, music, games… even business, finance, healthcare, government, design, and other serious applications are going social. These social media gave given rise to Social Cognition. What began with sharing has moved to creation. Consumers have become producers, and commerce has become a conversation.

Due to these conversations, individuals are no longer alone; whether you’re making a life decision, solving a critical business problem, or merely looking for a restaurant, your social graphs are available to augment your decision making process. These graphs have no geographic boundaries; professional networks are worldwide, and information streams from far corners of the globe into the palm of your hand.

Beyond media and commerce, the next big disruption is innovation. Humans everywhere want to innovate, and Social Cognition can augment human innovation in many everyday and expert domains.

I discuss three human capabilities that are amenable to social augmentation: problem solving, learning, and creativity. I illustrate them with challenge problems from my work: 1) healthcare: helping consumers find relevant health information without search; 2) energy: helping experts troubleshoot complex turbine failures; 3) learning: scaling education to a hundred million people; and 4) creativity: enabling average users to create artificial intelligence agents without programming, and 2) learning: scaling education to a hundred million people.

These technologies blend Cognitive Systems (artificial intelligence) and Cognitive Science (human cognition) in products that both exhibit and support cognition in large-scale social communities. This research not only provides scientific insight but also creates disruptive business opportunities.

Invited talk at PARC, Palo Alto, CA, April 7, 2011.
 
Invited talk at Wright State University, Center of Excellence in Human-Centered Innovation, Dayton, OH, October 24, 2010.
 

View the slides:

CBArch: A Case-Based Reasoning Framework for Conceptual Design of Commercial Buildings

The paper describes the first phase of development of a Case-Base Reasoning (CBR) system to support early conceptual design of buildings. As specific context of application, the research focuses on energy performance of commercial buildings, and the early identification of energy-related features that contribute to its outcomes. The hypothesis is that bringing knowledge from relevant precedents may facilitate this identification process, thus offering a significant contribution for early analysis and decision-making.

The paper introduces a proof-of-concept for such a system, proposing a novel integration of Case-Based Reasoning, Parametric Modeling (Building Information Modeling), and Ontology Classification. While CBR provides a framework to store and retrieve cases at an instance level, Parametric Modeling offers a framework for rule-based geometric adaptation and evaluation. The ontology is intended to provide a semantic representation, so that new design concepts can be created, classified and retained for further reuse. Potential advantages and limitations of this three-level integration approach are discussed along with recommendations for future development.

CBArch: A Case-Based Reasoning Framework for Conceptual Design of Commercial Buildings

by Andrés Cavieres, Urjit Bhatia, Preetam Joshi, Fei Zhao, Ashwin Ram

AAAI-11 Spring Symposium on Artificial Intelligence and Sustainable Design
www.cc.gatech.edu/faculty/ashwin/papers/er-11-07.pdf

Case-Based Reasoning and User-Generated AI for Real-Time Strategy Games

Creating AI for complex computer games requires a great deal of technical knowledge as well as engineering effort on the part of game developers. This paper focuses on techniques that enable end-users to create AI for games without requiring technical knowledge by using case-based reasoning techniques.

AI creation for computer games typically involves two steps: a) generating a first version of the AI, and b) debugging and adapting it via experimentation. We will use the domain of real-time strategy games to illustrate how case-based reasoning can address both steps.

Read the paper:

Case-Based Reasoning and User-Generated AI for Real-Time Strategy Games

by Santi Ontañón and Ashwin Ram

In P. Gonzáles-Calero & M. Gomez-Martín (ed.), AI for Games: State of the Practice, 2011.
www.cc.gatech.edu/faculty/ashwin/papers/er-11-02.pdf

Intentional analysis of medical conversations for community engagement

With an explosion in the proliferation of user-generated content in communities, information overload is increasing and quality of readily available online content is deteriorating. There is an increasing need for intelligent systems that make use of implicit user-generated knowledge in communities for community engagement. We describe our approach based on modeling user utterances in communities to proactively target the community for exchange of questions and answers. We envision a system that automatically encourages user engagement and participation by routing relevant conversations to users based on individual and community activity levels.

In this paper, we analyze health forum conversations from WebMD, a popular health portal consumer site, and classify them in different acts of speech using Verbal Response Modes (VRM) theory. We describe our approach for modeling an intelligent community recommender to engage participants based on observations from our analysis.

Read the paper:

Intentional analysis of medical conversations for community engagement

by Saurav Sahay, Hua Ai, Ashwin Ram

FLAIRS-11 International Conference on Artificial Intelligence
www.cc.gatech.edu/faculty/ashwin/papers/er-11-01.pdf

Learning Opponent Strategies through First Order Induction

In a competitive game it is important to identify the opponent’s strategy as quickly and accurately as possible so that an effective response can be staged. In this vein, this paper summarizes our work in exploring the use of the first order inductive learning (FOIL) algorithm for learning rules which can be used to represent opponent strategies. Specifically, we use these learned rules to perform plan recognition and classify an opponent strategy as one of multiple learned strategies. Our experiments validate this novel approach in a simple real-time strategy game.

Read the paper:

Learning Opponent Strategies through First Order Induction

by Kathryn Genter, Santiago Ontañón, Ashwin Ram

FLAIRS-11 International Conference on Artificial Intelligence
www.cc.gatech.edu/faculty/ashwin/papers/er-11-03.pdf

A is for Apple—The New Millennial Edition

A is for Apple, iGadgets at home
B is for Blackberry, your daddy’s smart phone
C is for CheckIn, place your own marks
D is for Disney, makes movies and parks
E is for Email, which old people use
F is for Facebook, it feeds you your news
G is for Google, search what it knows
H is for Hulu, catch up on shows
I is for iPod, your musical hits
J is for Java, coffee and bits
K is for Kindle, books in the cloud
L is for LOL, laughing out loud
M is for Microsoft, all hold it dear
N is for Napster, share with your peers
O is for Open, global and free
P is for Pandora, radio for me
Q is for QWERTY, tap to make words
R is for Remix, mashups absurd
S is for Skype, use it to talk
T is for Tweet, keep it real short
U is for Unix, it lives in /bin
V is for Valley, where startups begin
W is for WiFi, ‘net in the air
X is for Xbox, games with a flair
Y is for YouTube, its videos don’t end
Z is for Zynga, games with your friends
Copyright (c) 2010, Ashwin Ram

I’m looking for a collaborator to illustrate this in the form of a children’s book. And, a collaborator to set it to music. If you’re interested, please let me know.

Related Posts: Here’s a ballad I wrote many years ago: cognitivecomputing.wordpress.com/1986/01/28/a-short-ballad-dedicated-to-the-growth-of-programs

Make the World your Study Group

CNN Chalk Talk: A new website called OpenStudy allows students to share resources and learn with one another from all over the world.

Click the image to watch the video (3 min.)

Read the transcript: CNN Chalk Talk, October 1, 2010

T.J. HOLMES, CNN ANCHOR: Well, coming up, calling all college students. There’s now a group online that allows you to study in a unique way. You can get help from across the globe. You don’t even need a passport.

(BREAK)

HOLMES: Well, we turn to “Chalk Talk” today, now.

We are checking out a new study group that’s geared toward helping college students succeed. It’s an online study group called OpenStudy, and it’s linking students from around the world, helping them pass some tough courses.

Joining me now is Ashwin Ram. He’s the director of Georgia Tech’s Cognitive Computing Lab, one of the founders of OpenStudy.

Sir, thank you for being here.

OpenStudy, this is a worldwide study group. Do I kind of have that right?

ASHWIN RAM, DIRECTOR, GEORGIA TECH’S COGNITIVE COMPUTING LAB: That’s right. Open Study is a match.com for studying. It’s a social learning network that enables students to connect and study together, and get help when they need it.

HOLMES: Now, you said you’ve all been thinking about this for a while, for the past couple of years. What were you trying to work out, make sure there was a market for it, or is there some complicated technology you had to work out as well?

RAM: It was actually both. We wanted to get the value proposition right for students. We spent a lot of time researching the core need that students have, and that resulted in OpenStudy.

HOLMES: What did you determine was that core need? What did you find that students out there needed?

RAM: So, students all over the world are hitting their textbooks late at night cramming for exams. Maybe they’re working on review problems, watching video lectures on iTunes or MIT.

When these students need help, who can they turn to? The core need was to be able to find someone who can help them and give them help right there, right then, no matter what time they needed that help.

HOLMES: All right. And this is, again, supposed to link students with students. Essentially a study group like at the library.

RAM: It’s a worldwide study group. Our mantra is “We want to make the entire world your study group.” So there’s always someone who can help you.

HOLMES: How does this thing work? It looks like a social network page almost here.

RAM: It does. So let’s say that you are a student, and you’re one of 10,000 students studying computer science on MIT’s web site. And you’re working on video lectures or problem sets, and you have a question.

HOLMES: OK.

RAM: What do you do? You join a study group. When you do that, you get dropped into the MIT OpenStudy Group.

As you can see, we have over 2,200 people out there. Think of them as your classmates that can help you any time you want.

I noticed that we’ve just had someone join us from Kenya.

HOLMES: Oh, wow.

RAM: We actually have students from 138 countries from around the world. That’s 71 percent of the world’s countries.

HOLMES: Now, does this cost the kids anything to sign up for?

RAM: No, it’s completely free.

HOLMES: I’ll be danged. So you can pretty much — as well, you’re talking about kids up all hours of the night. No matter — somewhere in the world somebody is going to be up, somebody’s going to be logged on, somebody’s going to be studying.

RAM: Someone will always help you. And so if you have — you can go in and help somebody, but if you have a question, or you want to just study together with someone, you click on “Ask a Question,” type some question in that you want help with, and say, “Ask Now.”

The question is posted. Everything updates in real time. And you go back to the site, and then someone will be available to start answering you.

HOLMES: Will start answering you.

All right. Are you ready for growth? Because this might catch on. Are you ready for what might come?

RAM: We are ready for growth.

HOLMES: OK.

RAM: We’ve had remarkable growth already. We’ve only been live two weeks. We have over 6,000 people already using the site.

HOLMES: All right. This is going to be the next Facebook, 500 million. Come back when you get 500 million members in there. All right?

RAM: Thank you.

HOLMES: All right.

Ashwin Ram from Georgia Tech.

Thank you so much. Cool concept.

RAM: Thank you. It was a pleasure.

OpenStudying the Classics

I recently met Dr. Diana E. E. Kleiner, a distinguished professor at my alma mater and director of the Open Yale Courses initiative. We were talking about “OpenStudying the Classics”—to my knowledge, the first use of “OpenStudy” as a verb.[1] This made me think—what does it mean to “OpenStudy” something?

Some background first. In collaboration with Dean Preetha Ram of Emory University, our former student Chris Sprague from Georgia Tech’s HCI program, and experienced internet entrepreneur Phil Hill, and with funding from the National Science Foundation, the National Institutes of Health, and the Georgia Research Alliance, I’ve been working on a system called OpenStudy[2] which embodies a new way of studying. In the new millennial world of social networking, where social graphs have no geographical boundaries, professional networks are world wide, and entertainment streams from the far corners of the globe into the palm of your hand, it has always seemed odd to me that education is bounded by school walls, class interactions are limited to one teacher and a few dozen students who happened to register at the same time as you, and studying is largely a solitary activity circumscribed by so-called “collaboration policies” that typically require students to learn alone. Even “open learning” initiatives offer little more than a solitary experience watching instructional videos in your home, albeit from world famous experts.

OpenStudy, in contrast, whole-heartedly embraces the idea of “social learning”. The world is your study group, we claim. Connect with others studying the same things you are. Give and get help. The world learns as one.

But what is the “OpenStudy experience”? What will it mean, as Prof. Kleiner wonders, to “OpenStudy the Classics”? I don’t have the final answer (sic) but I do want to share my observations from a pilot with MIT OpenCourseware (OCW). For the past month, learners in three OCW courses have been given an option to “Join a study group”.[3] OCW reports their study groups are growing at a “blistering pace”—by our metrics, by about 10% a day. Learners are demanding more OpenStudy groups; if we don’t respond quickly, they create their own. What’s going on?

It’s too early for hard metrics, but permit me to share some anecdotes. MIT pilot courses include Intro CS, Calculus, and Chinese, and there are certainly interesting interactions around those topics.[4] But users are also exploring other interests. For example, there’s an active conversation about Greek Classics. What’s interesting are the participants:

  • a classics librarian at an exclusive four-year college in New England
  • a young woman considering a PhD in social sciences
  • an international student at a community college student in Georgia
  • a professor from the MIT Physics group
  • someone studying Chinese
  • a mid-career Math/CS geek from Michigan

These people did not know each other prior to their OpenStudy encounter. OpenStudy is described as the Match.com for studying together[5]—if so, this certainly seems to be working. A student from Peru came online recently, introduced himself, and apologized for his poor English. A student in the US responded in Spanish, and they struck up a conversation around their mutual study interests. Then a user from Mexico City jumped into the conversation, and off they went studying together with two users from Costa Rica. This is the new world of OpenStudying—social learning without geographical boundaries.

A homeschooled teenager recently joined OpenStudy and said “I’m new. How do you OpenStudy?” 15 minutes later, she had connected with students in an all-girls private school. She initiated a discussion on World Religions which, less than a day later, has nearly 20 participants. Half of them have contributed and half are listening. A Hindu undergraduate from India, an Orthodox Jew from Texas, and a Muslim student from Turkey are talking about what “real” Islam is like. A teenager who can’t drive, doesn’t go to school, and does not have traditional teachers or schoolmates has answered her own question. This is how you OpenStudy. You study with the world.

Every educator knows the challenge of keeping students engaged. Studying together not only improves learning, it is a lot more fun. One of the users recently emailed us saying: “Personally, I’ve come further in my development as a programmer in the month of being on OpenStudy than the previous few years struggling on my own. Being able to see how other people approach problems and considering their questions is absolutely wonderful.” A GSU professor says she is seeing 400% increase in student engagement in her required lower-division biology class due to OS.

So this, Prof. Kleiner, is how we will be able to “OpenStudy the Classics”. Students connecting with students studying the same things they are. I call it “massively multiplayer online learning[6], a wordplay on the MMO experience we’re seeing in the gaming world. Here it has value beyond entertainment; the diversity adds to the richness of the online study group, the globalness broadens access beyond elite institutional walls, the interactivity engages today’s millennials in—of all things—study.

On average, we’re seeing 5.3 participants per “studypad” (a real-time interaction tool that facilitates conversation, discussion, or simply question answering). About 30% of the interactions occur synchronously in real time. This is quite different from a typical question site, where you post a question and wait—an hour? a day? who knows when someone might answer. OpenStudying is like a conversation in a university library or the local Starbucks, instant real-time interaction with peers—except that these peers might be halfway around the globe. The world is, after all, your social network, your professional rolodex, and, now, your study group.

Ashwin Ram
September 15, 2010

If you’ve OpenStudied and would like to share your experience, I’d love to hear about it. Please add a comment below.


[1] I’ve always been interested in the origins of words, especially new ones. Who coined the term “WebLog”, and who first shortened it to “blog? Who first used “Google” as a verb? In this day and age, surely there must be a record somewhere. To this end and with her permission, I’d like to credit Prof. Diana Kleiner with the first use of “OpenStudy” as a verb.

[2] OpenStudy is free and publicly available at OpenStudy.com, a for-profit spinoff from Georgia Tech and Emory University created via the university’s commercialization program. Our objective is to create not just an interesting research project but a sustainable product that will make a difference to thousands of learners everywhere. To accomplish this, we need to grapple with the realities of business models, lest our project die the way countless other good ideas do when their research funding runs out.

[3] Update: MIT OCW has expanded its program to several more courses: web.mit.edu/newsoffice/2010/ocw-openstudy.html

[4] Click “Join a Study Group” on the course page to see the corresponding study group.

[5] See Marc Parry’s article in The Chronicle’s Wired Campus, Start-Up Aspires to Make the World ‘One Big Study Group’, September 8, 2010: wiredcampus.chronicle.com/blogPost/Start-Up-Aspires-to-Make-the/26780

[6] My talk at the Knowledge Futures: Disrupting the University forum at Emory University, entitled Massively Multiplayer Online—Learning? aka Are social networks disrupting models of education? cognitivecomputing.wordpress.com/2010/04/26/massively-multiplayer-online—learning/

MMPM: A Generic Platform for Case-Based Planning Research

Computer games are excellent domains for research and evaluation of AI and CBR techniques. The main drawback is the effort needed to connect AI systems to existing games. This paper presents MMPM, a middleware platform that supports easy connection of AI techniques with games. We will describe the MMPM architecture, and compare with related systems such as TIELT.

Read the paper:

MMPM: A Generic Platform for Case-Based Planning Research

by Pedro Pablo Gómez-Martín, David Llansó, Marco Antonio Gómez-Martín, Santiago Ontañón, Ashwin Ram

ICCBR-2010 Workshop on Case-Based Reasoning for Computer Games
www.cc.gatech.edu/faculty/ashwin/papers/er-10-03.pdf

Real-Time Case-Based Reasoning for Interactive Digital Entertainment

(Click image to view the video – it’s near the bottom of the new page.)

User-generated content is everywhere: photos, videos, news, blogs, art, music, and every other type of digital media on the Social Web. Games are no exception. From strategy games to immersive virtual worlds, game players are increasingly engaged in creating and sharing nearly all aspects of the gaming experience: maps, quests, artifacts, avatars, clothing, even games themselves. Yet, there is one aspect of computer games that is not created and shared by game players: the AI. Building sophisticated personalities, behaviors, and strategies requires expertise in both AI and programming, and remains outside the purview of the end user.

To understand why authoring Game AI is hard, we need to understand how it works. AI can take digital entertainment beyond scripted interactions into the arena of truly interactive systems that are responsive, adaptive, and intelligent. I will discuss examples of AI techniques for character-level AI (in embedded NPCs, for example) and game-level AI (in the drama manager, for example). These types of AI enhance the player experience in different ways. The techniques are complicated and are usually implemented by expert game designers.

I propose an alternative approach to designing Game AI: Real-Time CBR. This approach extends CBR to real-time systems that operate asynchronously during game play, planning, adapting, and learning in an online manner. Originally developed for robotic control, Real-Time CBR can be used for interactive games ranging from multiplayer strategy games to interactive believable avatars in virtual worlds.

As with any CBR technique, Real-Time CBR integrates problem solving with learning. This property can be used to address the authoring problem. I will show the first Web 2.0 application that allows average users to create AIs and challenge their friends to play them—without programming. I conclude with some thoughts about the role of CBR in AI-based Interactive Digital Entertainment.

Keynote talk at the Eighteenth Conference on Pattern Recognition and Artificial Intelligence (RFIA-12), Lyon, France, February 5, 2012.
Slides and video here: rfia2012.liris.cnrs.fr/doku.php?id=pub:ram
 
Keynote talk at the Eleventh Scandinavian Conference on Artificial Intelligence (SCAI-11), Trondheim, Norway, May 25, 2011.
 
Keynote talk at the 2010 International Conference on Case-Based Reasoning (ICCBR-10), Alessandria, Italy, July 22, 2010.
 
GVU Brown Bag talk, October 14, 2010. Watch the talk here: www.gvu.gatech.edu/node/4320 
 
Try it yourself:
Learn more about the algorithms:
View the talk:
www.sais.se/blog/?p=57

View the slides: