We propose a method that learns a discriminative yet semantic space for object categorization, where we also embed auxiliary semantic entities such as supercat-egories and attributes. Contrary to prior work, which only utilized them as side information, we explicitly embed these semantic entities into the same space where we embed categories, which enables us to represent a category as their linear combination. By exploiting such a unified model for semantics, we enforce each category to be generated as a supercategory + a sparse combination of attributes, with an additional exclusive regularization to learn discriminative composition. The proposed reconstructive regularization guides the discriminative learning process to learn a model with better generalization. This model also generates compact semantic description of each category, which enhances interoperability and enables humans to analyze what has been learned.

In recent years, we have witnessed the success of autonomous agents applying machine-learning techniques across a wide range of applications. However, agents applying the same machine-learning techniques in online applications have not been so successful. Even agent-based hybrid recommender systems that combine information filtering techniques with collaborative filtering techniques have been applied with considerable success only to simple consumer goods such as movies, books, clothing, and food. Yet complex, adaptive autonomous agent systems that can handle complex goods such as real estate, vacation plans, insurance, mutual funds, and mortgages have emerged. To a large extent, the reinforcement learning methods developed to aid agents in learning have been more successfully deployed in offline applications.

For large collections of images, such as those resulting from astronomy sky surveys, the typical useful product is an online database cataloging entries of interest. We focus on the automation of the cataloging effort of a major sky survey and the availability of digital libraries in general. For the primary scientific analysis of these data, it is necessary to detect, measure, and classify every sky object. The learning algorithms are trained to classify the detected objects and can classify objects too faint for visual classification with an accuracy level exceeding 90 percent. This accuracy level increases the number of classified objects in the final catalog threefold relative to the best results from digitized photographic sky surveys to date.

He points out that one of the key features these systems lack is "a suitable verification methodology or a technique for testing the consistency and completeness of a rule set." It is precisely this feature that we address here. LES is a generic rule-based expert system building tool (Laffey, Perkins, and Nguyen 1986) similar to EMYCIN (Van Melle 1981) that has been used as a framework to construct expert systems in many areas, such as electronic equipment diagnosis, design verification, photointerpretation, and hazard analysis. LES represents factual data in its frame database and heuristic and control knowledge in its production rules. LES allows the knowledge engineer to use both data-driven and goaldriven rules.

The traditional approach involving protracted interaction between a knowledge engineer and a domain expert is viable only to the extent that both these resources are available; this approach will not meet the apparently exponential growth in demand for expert systems. A solution to this dilemma requires rethinking the way knowledge-based products are built. An example of this reappraisal of methodology appears in Michie (1983), and is based on the principle of formalizing and refining the knowledge implicit in collections of examples or data bases. Dietterich and Michalski (1983) give an overview of methods for learning from examples. There are many such, all based on the idea of inductive generalization. One of the simplest of these methods dates back to work by Hunt in the late fifties (Hunt et al., 1966). Each given example, described by measuring certain fixed properties, belongs to a known class and the'learning' takes the form of developing a classification rule that can then be applied to new objects. Simple though it may be, derivatives of this method have achieved useful results; Kononenko et al. (1984), for example, have managed to generate five medical diagnosis systems with minimal reference to diagnosticians.

Programs Based on the Blackboard Model 8 2.2.1.

The Intent of the protocol is to show AGE from the user's point of view. The main body describes the motivations for and the description of the facilities in AGE.

AGE is a system designed to aid knowledge engineers build knowledge-based programs. We have built a laboratory of building blocks which the user can assemble in various ways to fit a particular problem. This paper describes the facilities available to the user and descibes an example program built with AGE.

Psychological work on memory, in particular by Bartlett (1932), has led the conclusion that people faced with a new situation use large amounts of highly structured knowledge acquired from previous experience. Bartlett used the word schema to refer to this phenomenon. Minsky (1975), his famous paper, proposed the notion of a frame as a fundamental structure used in natural language understanding, as well as in scene analysis. I will use the former term in the rest of this chapter, in spite of its general connotation. The main thesis defended by Bartlett was that the phenomena of memorization and remembering are both constructive and selective. The hypothesis has more recently been revived by psychologists working on discourse structure (Collins, 1978; Bransford and Franks, 1971; Kintsch, 1976). Various experiments performed on subjects who were told stories and then asked to describe what they remembered showed that people not only forget facts but add some. Moreover, they are unable to distinguish between what they have actually heard and what they have inferred. People hearing a story make assumptions, which they might revise or refine as more information comes in, either confirmatory or contradictory. Making such assumptions entails building (or retrieving) models of the expected text contents. A corollary of this process is that if the story adequately fits the model people have in mind, the story will be understood more easily. This chal)ter is based on a technical memo (HPP-79-25) from the Heuristic Programming lh( iect, l)cparmlent of Computer Science, Stanford University.

If it is true that good problems produce good science, then it will be worthwhile to identify good problems, and even more worthwhile to discover the attributes that make them good problems. This discovery process is necessarily empirical, so we examine several challenge problems, beginning with Turing's famous test, and more than a dozen attributes that challenge problems might have. We are led to a contrast between research strategies -- the successful "divide and conquer" strategy and the promising but largely untested "developmental" strategy -- and we conclude that good challenge problems encourage the latter strategy.