16 Jul

An Ensemble Learning and Problem Solving Architecture for Airspace Management

Posted by cognitivecomputing in Learning. Tagged: , , , .

In this paper we describe the application of a novel learning and problem solving architecture to the domain of airspace management, where multiple requests for the use of airspace need to be reconciled and managed automatically. The key feature of our “Generalized Integrated Learning Architecture” (GILA) is a set of integrated learning and reasoning (ILR) systems coordinated by a central meta-reasoning executive (MRE). Each ILR learns independently from the same training example and contributes to problem-solving in concert with other ILRs as directed by the MRE. Formal evaluations show that our system performs as well as or better than humans after learning from the same training data. Further, GILA outperforms any individual ILR run in isolation, thus demonstrating the power of the ensemble architecture for learning and problem solving.

Read the paper:

An Ensemble Learning and Problem Solving Architecture for Airspace Management

by XS Zhang et al.

International Conference on Innovative Applications of Artificial Intelligence (IAAI-09), Pasadena, CA, July 2009
www.cc.gatech.edu/faculty/ashwin/papers/er-09-03.pdf

15 Jul

Goal-Driven Learning in the GILA Integrated Intelligence Architecture

Posted by cognitivecomputing in Learning. Tagged: , , , .

Goal Driven Learning (GDL) focuses on systems that determine by themselves what has to be learned and how to learn it. Typically GDL systems use meta-reasoning capabilities over a base reasoner, identifying learning goals and devising strategies. In this paper we present a novel GDL technique to deal with complex AI systems where the meta-reasoning module has to analyze the reasoning trace of multiple components with potentially different learning paradigms. Our approach works by distributing the generation of learning strategies among the different modules instead of centralizing it in the meta-reasoner. We implemented our technique in the GILA system, that works in the airspace task orders domain, showing an increase in performance.

Read the paper:

Goal-Driven Learning in the GILA Integrated Intelligence Architecture

by Jai Radhakrishnan, Santi Ontañón, Ashwin Ram

International Joint Conference on Artificial Intelligence (IJCAI-09), Pasadena, CA, July 2009
www.cc.gatech.edu/faculty/ashwin/papers/er-09-02.pdf

12 Jul

Learning from Human Demonstrations for Real-Time Case-Based Planning

Posted by cognitivecomputing in Game AI, Learning. Tagged: , , , , .

One of the main bottlenecks in deploying case-based planning systems is authoring the case-base of plans. In this paper we present a collection of algorithms that can be used to automatically learn plans from human demonstrations. Our algorithms are based on the basic idea of a plan dependency graph, which is a graph that captures the dependencies among actions in a plan. Such algorithms are implemented in a system called Darmok 2 (D2), a case-based planning system capable of general game playing with a focus on real-time strategy (RTS) games. We evaluate D2 with a collection of three different games with promising results.

Read the paper:

Learning from Human Demonstrations for Real-Time Case-Based Planning

by Santi Ontañón, Kane Bonnette, Praful Mahindrakar, Marco Gómez-Martin, Katie Long, Jai Radhakrishnan, Rushabh Shah, Ashwin Ram

IJCAI-09 Workshop on Learning Structural Knowledge from Observations, Pasadena, CA, July 2009
www.cc.gatech.edu/faculty/ashwin/papers/er-09-04.pdf

1 Jul

On Similarity Measures based on a Refinement Lattice

Posted by cognitivecomputing in Agents. Tagged: , .

Retrieval of structured cases using similarity has been studied in CBR but there has been less activity on defining similarity on description logics (DL). We present an approach that allows us to present two similarity measures for feature logics, a subfamily of DLs, based on the concept of “refinement lattice”. The first one is based on computing the anti-unification (AU) of two cases to assess the amount of shared information. The second measure decomposes the cases into a set of independent “properties”, and then assesses how many of these properties are shared between the two cases. Moreover, we show that the defined measures are applicable to any representation language for which a refinement lattice can be defined. We empirically evaluate our measures comparing them to other measures in the literature in a variety of relational data sets showing very good results.

Read the paper:

On Similarity Measures based on a Refinement Lattice

by Santi Ontañón and Enric Plaza

in ICCBR 2009, LNAI 5650, pp 240 – 255
www.cc.gatech.edu/faculty/ashwin/papers/er-09-11.pdf

10 Jun

Using Content Analysis to Investigate The Research Paths Chosen by Scientists over Time

Posted by cognitivecomputing in Language. Tagged: , .

We present an application of a clustering technique to a large original dataset of SCI publications which is capable at disentangling the different research lines followed by a scientist, their duration over time and the intensity of effort devoted to each of them. Information is obtained by means of software-assisted content analysis, based on the co-occurrence of words in the full abstract and title of a set of SCI publications authored by 650 American star-physicists across 17 years. We estimated that scientists in our dataset over the time span contributed on average to 16 different research lines lasting on average 3.5 years and published nearly 5 publications in each single line of research. The technique is potentially useful for scholars studying science and the research community, as well as for research agencies, to evaluate if the scientist is new to the topic and for librarians, to collect timely biographic information.

Read the paper:

Using Content Analysis to Investigate The Research Paths Chosen by Scientists over Time

by Chiara Franzoni, Chris Simpkins, Baoli Li, Ashwin Ram

Scientometrics journal 83(1):321-335, April 2010. (Earlier version in 1th International Conference on Scientometrics and Infometrics (ISSI-07), Madrid, Spain, June 2007.)
www.springerlink.com/content/5462n515405715u2/?p=8344e997766b4ecdabee78f5e27a9faa&pi=18
www.cc.gatech.edu/faculty/ashwin/papers/er-07-06.pdf

22 May

Emotional Memory and Adaptive Personalities

Posted by cognitivecomputing in Agents, Learning, Robotics. Tagged: , , , , .

Believable agents designed for long-term interaction with human users need to adapt to them in a way which appears emotionally plausible while maintaining a consistent personality. For short-term interactions in restricted environments, scripting and state machine techniques can create agents with emotion and personality, but these methods are labor intensive, hard to extend, and brittle in new environments. Fortunately, research in memory, emotion and personality in humans and animals points to a solution to this problem. Emotions focus an animal’s attention on things it needs to care about, and strong emotions trigger enhanced formation of memory, enabling the animal to adapt its emotional response to the objects and situations in its environment. In humans this process becomes reflective: emotional stress or frustration can trigger re-evaluating past behavior with respect to personal standards, which in turn can lead to setting new strategies or goals.

To aid the authoring of adaptive agents, we present an artificial intelligence model inspired by these psychological results in which an emotion model triggers case-based emotional preference learning and behavioral adaptation guided by personality models. Our tests of this model on robot pets and embodied characters show that emotional adaptation can extend the range and increase the behavioral sophistication of an agent without the need for authoring additional hand-crafted behaviors.

Read the paper:

Emotional Memory and Adaptive Personalities

by Anthony Francis, Manish Mehta, Ashwin Ram

Handbook of Research on Synthetic Emotions and Sociable Robotics: New Applications in Affective Computing and Artificial Intelligence, IGI Global, 2009
www.cc.gatech.edu/faculty/ashwin/papers/er-08-10.pdf

15 May

Using First Order Inductive Learning as an Alternative to a Simulator in a Game Artificial Intelligence

Posted by cognitivecomputing in Game AI, Learning. Tagged: , , .

Currently many game artificial intelligences attempt to determine their next moves by using a simulator to predict the effect of actions in the world. However, writing such a simulator is time-consuming, and the simulator must be changed substantially whenever a detail in the game design is modified. As such, this research project set out to determine if a version of the first order inductive learning algorithm could be used to learn rules that could then be used in place of a simulator.

We used an existing game artificial intelligence system called Darmok 2. By eliminating the need to write a simulator for each game by hand, the entire Darmok 2 project could more easily adapt to additional real-time strategy games. Over time, Darmok 2 would also be able to provide better competition for human players by training the artificial intelligences to play against the style of a specific player. Most importantly, Darmok 2 might also be able to create a general solution for creating game artificial intelligences, which could save game development companies a substantial amount of money, time, and effort.

Read the thesis:

Using First Order Inductive Learning as an Alternative to a Simulator in a Game Artificial Intelligence

by Katie Long

Undergraduate Thesis, College of Computing, Georgia Institute of Technology, Atlanta, GA, 2009
www.cs.utexas.edu/users/katie/UgradThesis.pdf

11 May

Creating Behavior Authoring Environments for Everyday Users

Posted by cognitivecomputing in Game AI. Tagged: , , , .

The design of interactive experiences is increasingly important in our society. Examples include interactive media, computer games, and interactive portals. There is increasing interest in modes of interaction with virtual characters, as they represent a natural way for humans to interact. Creating such characters is a complex task, requiring both creative skills (to design personalities, emotions, gestures, behaviors) and programming skills (to code these in a scripting or programming language). There is little understanding of how the behavior authoring process can be simplified with easy-to-use authoring environments that can support the cognitive needs of everyday users and help them at every step to easily carry out this creative task.

Our research focuses on behavior authoring environments that not only make it easy for novices/everyday users to create characters but also provide them scaffolding in designing these interactive experiences. In this paper we present results from a user study with a paper prototype of an authoring environment that is aimed to allow everyday users to create virtual characters. The study aims at determining whether typical computer users are able to create character personalities in specific scenarios and think about characters’ mental states, and if so, then what kinds of user interfaces would be suitable for this authoring environment.

Read the paper:

Creating Behavior Authoring Environments for Everyday Users

by Manish Mehta, Christina Lacey, Iulian Radu, Abhishek Jain, Ashwin Ram

International Conference on Computer Games, Multimedia, and Allied Technologies (CGAT-09), Singapore, May 2009
www.cc.gatech.edu/faculty/ashwin/papers/er-09-01.pdf

9 Mar

Teaching Computational Thinking to Third-Graders

Posted by cognitivecomputing in Education. Tagged: , .

From the Paideia School Newsletter, March 2009:

Computational Thinking is a new way of solving problems that derives from computer science. It involves approaching problems in a systematic, step-by-step manner, and building up solutions to complex problems from smaller pieces. Nowadays, computational thinking is a fundamental skill for everyone, not just computer scientists. Many educators argue that it should be part of every child’s education along with reading, writing and arithmetic.

Assisted by Georgia Tech computer science professor Dr. Ashwin Ram, a group of students in Kelly and Tony’s 3rd grade class at The Paideia School explored computational thinking over a four-week affinities session. The students built familiar computer games, including Etch-A-Sketch, Pong, and Tennis, and shared them with each other on a web site. They built the games not only from scratch but in Scratch, a new tool from MIT that is designed to enable young children design and build interactive computer programs.

In teams of two, the students played with algorithmic concepts including scripts, conditionals, and loops. They customized the look-and-feel of their games, built game characters called sprites, designed behaviors for their sprites, and added sounds for extra effect. They decided how their games would respond to the player and how to keep score. They tested, revised, tested again, revised again.

And when they were done, they shared their games for their classmates and anyone else to enjoy. They received some nice comments! You can try out their games yourself at scratch.mit.edu/users/kellytony.

Says Ashwin: “I received a wonderful set of thank you cards from my affinities group. What a nice surprise. One of the children wrote, “Those weeks were the best weeks of my life.” Wow. It was a great experience. The kids got a lot out of it — they learned something new, grappled with a new way of thinking about problems, and had fun at the same time. I know some of them will want to continue doing this, at least until they find their next passion! And I had a blast as well. Thank you, Kelly and Tony, for giving me this opportunity.”

Ashwin and Preetha Ram are the proud parents of three Paideia students, Naveen (3rd grade), Maya (6th grade), and Nikhil (12th grade). Ashwin can be reached by email (ashwin AT cc.gatech.edu), Twitter (@ashwinram), and LinkedIn (www.linkedin.com/in/ashwinram). Scratch is freely available at scratch.mit.edu.


29 Jan

NLP: Not (Just) Language, People

Posted by cognitivecomputing in Language, Talks, Web / Web 2.0. Tagged: .

As consumers become producers and, now, participants in online social communities, there are new opportunities and challenges in the increasing amounts of textual information and interactions on the web, within enterprises, in government, and in new types of social media and virtual worlds.

Natural Language Processing (NLP) researchers have traditionally regarded language as the object of study. In this talk, I argue that NLP is as much a study of people as of language per se. Doing NLP well requires us to model and reason about Content (domain knowledge), Context (goals and tasks), and Community (social context). I discuss why modeling the three C’s is difficult, and illustrate some approaches to these problems using examples from my recent academic and commercial projects.

Invited talk at PARC (Palo Alto Research Labs), Palo Alto, CA, January 2009
« Older Entries
Newer Entries »