Special Issue on Structured Knowledge Transfer

By Shapiro, Daniel G.; Munoz-Avila, Hector et al. | AI Magazine, Spring 2011 | Go to article overview

Special Issue on Structured Knowledge Transfer


Shapiro, Daniel G., Munoz-Avila, Hector, Stracuzzi, David, AI Magazine


Transfer learning is the problem of acquiring knowledge in one domain and using it to improve performance in another. While the field of psychology has studied transfer learning in people for many years, AI has only recently taken up the challenge. The topic received initial attention with work on inductive transfer in the 1990s, while the number of workshops and conferences has noticeably increased in the last five years. This special issue represents the state of the art in the subarea of transfer learning that focuses on the acquisition and reuse of structured knowledge. Its goal is to capture, in a general form, knowledge about the objects, relations, strategies, and processes used to solve tasks drawn from a source domain and exploit it in tasks taken from a target domain.

This special issue is motivated largely, but not entirely, by the completion of the DARPA Transfer Learning program, which funded the lion's share of work on structured knowledge transfer and has helped popularize the broader subject of transfer learning as a whole. The issue contains five articles sponsored by the program, plus two outside it, along with three opinion pieces authored by prominent figures in the field. Taken as a whole, the articles paint a picture of structured knowledge transfer as an emerging discipline that is just arriving at an understanding of the problems, methods, and techniques involved.

The articles in this special issue address transfer learning within a three-stage framework: (1) knowledge acquisition in a source task, (2) knowledge mapping between the source and target tasks, and (3) exploitation of the transferred knowledge in the target setting. We include work on near transfer, where the source and target tasks are quite similar (for example, drawn from the same domain), plus instances of far transfer where the tasks come from dissimilar domains that do not even share a common representational vocabulary. The papers describe transfer within, and between, analysis and synthesis tasks, and they do so by communicating multiple types of knowledge. The objects of transfer include declarative structure (for example, model parameters and relational predicates), procedural knowledge (for example, plans), and more abstract problem-solving strategies (for example, search control heuristics). One of our goals for this special issue was to display this breadth of metaphor.

"An Introduction to Intertask Transfer for Reinforcement Learning," by Matthew E. Taylor and Peter Stone illustrates transfer in reinforcement learning (RL) settings. It presents multiple methods of mapping actions, states, preferences, control knowledge, and specific value functions from source to target tasks. The work includes examples of both near and far transfer and contributes a broad categorization of transfer tasks. It also provides an insightful analysis of future research paths.

"Automatic Discovery and Transfer of Task Hierarchies in Reinforcement Learning," by Neville Mehta, Soumya Ray, Prasad Tadepalli, and Thomas Dietterich, also explores transfer in an RL setting but focuses on a deep causal analysis of successful strategies. Here, the object of transfer is a hierarchical solution procedure that generalizes source behavior for use in the target task. Importantly, this work isolates the benefit of structure versus value transfer through lesion experiments.

"The Case for Case-Based Transfer Learning," by Matthew Klenk, David W. Aha, and Matt Molineaux illustrates the application of case-based reasoning to transfer tasks. It examines work by multiple authors to show how case similarity and case reuse can perform transfer as a whole or address the component problems of transfer (source knowledge acquisition, knowledge mapping, and knowledge exploitation with learning in the target task). This work advances the characterization of transfer problems and documents the clear but underappreciated relation between case-based reasoning and transfer tasks. …

The rest of this article is only available to active members of Questia

Already a member? Log in now.

Notes for this article

Add a new note
If you are trying to select text to create highlights or citations, remember that you must now click or tap on the first word, and then click or tap on the last word.
One moment ...
Default project is now your active project.
Project items
Notes
Cite this article

Cited article

Style
Citations are available only to our active members.
Buy instant access to cite pages or passages in MLA 8, MLA 7, APA and Chicago citation styles.

(Einhorn, 1992, p. 25)

(Einhorn 25)

(Einhorn 25)

1. Lois J. Einhorn, Abraham Lincoln, the Orator: Penetrating the Lincoln Legend (Westport, CT: Greenwood Press, 1992), 25, http://www.questia.com/read/27419298.

Note: primary sources have slightly different requirements for citation. Please see these guidelines for more information.

Cited article

Special Issue on Structured Knowledge Transfer
Settings

Settings

Typeface
Text size Smaller Larger Reset View mode
Search within

Search within this article

Look up

Look up a word

  • Dictionary
  • Thesaurus
Please submit a word or phrase above.
Print this page

Print this page

Why can't I print more than one page at a time?

Help
Full screen
Items saved from this article
  • Highlights & Notes
  • Citations
Some of your highlights are legacy items.

Highlights saved before July 30, 2012 will not be displayed on their respective source pages.

You can easily re-create the highlights by opening the book page or article, selecting the text, and clicking “Highlight.”

matching results for page

    Questia reader help

    How to highlight and cite specific passages

    1. Click or tap the first word you want to select.
    2. Click or tap the last word you want to select, and you’ll see everything in between get selected.
    3. You’ll then get a menu of options like creating a highlight or a citation from that passage of text.

    OK, got it!

    Cited passage

    Style
    Citations are available only to our active members.
    Buy instant access to cite pages or passages in MLA 8, MLA 7, APA and Chicago citation styles.

    "Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn, 1992, p. 25).

    "Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn 25)

    "Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences." (Einhorn 25)

    "Portraying himself as an honest, ordinary person helped Lincoln identify with his audiences."1

    1. Lois J. Einhorn, Abraham Lincoln, the Orator: Penetrating the Lincoln Legend (Westport, CT: Greenwood Press, 1992), 25, http://www.questia.com/read/27419298.

    Cited passage

    Thanks for trying Questia!

    Please continue trying out our research tools, but please note, full functionality is available only to our active members.

    Your work will be lost once you leave this Web page.

    Buy instant access to save your work.

    Already a member? Log in now.

    Search by... Author
    Show... All Results Primary Sources Peer-reviewed

    Oops!

    An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.