A method is being developed that provides for an artificial-intelligence system to learn a user’s preferences for sets of objects and to thereafter automatically select subsets of objects according to those preferences. The method was originally intended to enable automated selection, from among large sets of images acquired by instruments aboard spacecraft, of image subsets considered to be scientifically valuable enough to justify use of limited communication resources for transmission to Earth. The method is also applicable to other sets of objects: examples of sets of objects considered in the development of the method include food menus, radio-station music playlists, and assortments of colored blocks for creating mosaics.
The method does not require the user to perform the often-difficult task of quantitatively specifying preferences; instead, the user provides examples of preferred sets of objects. This method goes beyond related prior artificial-intelligence methods for learning which individual items are preferred by the user: this method supports a concept of set-
based preferences, which include not only preferences for individual items but also preferences regarding types and degrees of diversity of items in a set. Consideration of diversity in this method involves recognition that members of a set may interact with each other in the sense that when considered together, they may be regarded as being complementary, redundant, or incompatible to various degrees. The effects of such interactions are loosely summarized in the term “portfolio effect.”
The learning method relies on a preference representation language, denoted DD-PREF, to express set-based preferences. In DD-PREF, a preference is represented by a tuple that includes quality (“depth”) functions to estimate how desired a specific value is, weights for each feature preference, the desired diversity of feature values, and the relative importance of diversity versus depth. The system applies statistical concepts to estimate quantitative measures of the user’s preferences from training examples (preferred subsets) specified by the user. Once preferences have been learned, the system uses those preferences to select preferred subsets from new sets.
The method was found to be viable when tested in computational experiments on menus, music playlists, and rover images. Contemplated future development efforts include further tests on more diverse sets and development of a submethod for (a) estimating the parameter that represents the relative importance of diversity versus depth, and (b) incorporating background knowledge about the nature of quality functions, which are special functions that specify depth preferences for features.
This work was done by Kiri L. Wagstaff of Caltech and Marie desJardins and Eric Eaton of the University of Maryland, Baltimore County, for NASA’s Jet Propulsion Laboratory. For more information, download the Technical Support Package (free white paper) at www.techbriefs.com/tsp under the Information Sciences category.
The software used in this innovation is available for commercial licensing. Please contact Daniel Broderick of the California Institute of Technology at
This Brief includes a Technical Support Package (TSP).
Algorithms for Learning Preferences for Sets of Objects
(reference NPO-43828) is currently available for download from the TSP library.
Don't have an account?
Overview
The document titled "Algorithms for Learning Preferences for Sets of Objects" (NPO-43828) from NASA's Jet Propulsion Laboratory outlines innovative methods for understanding and learning user preferences regarding sets of objects, rather than just individual items. This research addresses the need for users to select groups of items that may interact in complex ways, such as being complementary or incompatible, a phenomenon referred to as the "portfolio effect."
A practical example provided is the scenario of a rover collecting images, where it can only transmit a subset back to Earth. The technology developed aims to enable the rover to autonomously determine which images best align with user-defined preferences, particularly in terms of scientific priorities.
The solution proposed consists of two main components. The first is a specification language called DD-PREF, which allows users to articulate their preferences for sets of objects. The second component is a technique for automatically learning these preferences based on example sets provided by users. This approach is particularly beneficial because it simplifies the process for users, allowing them to give examples of desired sets rather than requiring them to quantitatively define their preferences for each individual feature.
The novelty of this work lies in its focus on set-based preferences, marking a significant advancement over traditional methods that only consider individual objects. Users can specify not only the items they wish to include but also how these items interact, emphasizing desired characteristics such as "depth" and "diversity" within the selected set.
The document references a related publication, “Learning User Preferences for Sets of Objects,” presented at the 23rd International Conference on Machine Learning in 2006, which further elaborates on the methodologies and findings of this research.
Overall, this work represents a significant step forward in automated subset selection, providing a framework that aligns with user preferences in complex scenarios, particularly in aerospace applications. The research has broader implications for various technological, scientific, and commercial applications, showcasing NASA's commitment to advancing technology for wider use beyond aerospace.
