An inquiry into computer understanding

This essay addresses a number of issues centered around the question of what is the best method for representing and reasoning about common sense (sometimes called plausible inference). Drew McDermott has shown that a direct translation of commonsense reasoning into logical form leads to insurmountable difficulties, from which McDermott concluded that we must resort to procedural ad hocery. This paper shows that the difficulties McDermott described are a result of insisting on using logic as the language of commonsense reasoning. If, instead, (Bayesian) probability is used, none of the technical difficulties found in using logic arise. For example, in probability, the problem of referential opacity cannot occur and nonmonotonic logics (which McDermott showed don't work anyway) are not necessary. The difficulties in applying logic to the real world are shown to arise from the limitations of truth semantics built into logic–probability substitutes the more reasonable notion of belief. In Bayesian inference, many pieces of evidence are combined to get an overall measure of belief in a proposition. This is much closer to commonsense patterns of thought than long chains of logical inference to the true conclusions. Also it is shown that English expressions of the “IF A THEN B” form are best interpreted as conditional probabilities rather than universally quantified expressions. Bayesian inference is applied to a simple example of linguistic information to illustrate the potential of this type of inference for AI. This example also shows how to deal with vague information, which has so far been the province of fuzzy logic. It is further shown that Bayesian inference gives a theoretical basis for inductive inference that is borne out in practice. Instead of insisting that probability is the best language for commonsense reasoning, a major point of this essay is to show that real inference is a complex interaction between probability, logic, and other formal representation and reasoning systems.     

Comments on Peter Cheeseman''s An inquiry into computer understanding

I agree that probability in some, possibly disguised, form is a necessary component of practical inference. Our goal, therefore, must be to find how to make probability practical. Normatively motivated thinking (which includes textbook Bayesianism) is, however, mute on two topics that are crucial to practicality: the choice of a limited but useful set of initial hypotheses and the choice of decision rules. The best way of reducing the computational demands of probability is to use the smallest and simplest set of hypotheses, and the crudest decision rules that are compatible with your application's goals. Learning how to make these choices is an empirical, not a normative, endeavour.     

ED-DESIGN: An engineering drawing based design model for computer understanding

Constructing a solid model through computer understanding of engineering drawings could resolve the input problem of solid models. For computer understanding, however, the engineering drawing has to be defined and represented with clear semantics. The conventional engineering drawing is usually described by showing examples and represented as a set of curve segments on a drawing plane, so that it can only be understood by human engineers. In this paper, an engineering drawing based design model, ED-DESIGN, which is much more computer understandable than the conventional one is proposed. In this model, an engineering drawing is represented in a drawing space rather than in a drawing plane. In other words, the scheme of engineering drawing is interpreted as pseudo-3D surface design scheme rather than a 2D design scheme. With this design model, it is possible to develop definite knowledge for constructing a solid model through computer understanding.     

An inquiry into human nature and the cost of the wealth of nations

Current economic ontology development has failed to confront two important errors associated with historicism. Embracing the linearity of economic value being directly attributed to the labor applied to natural resources taken together with efficiency arguments used to justify monetary policy on both the microlevel (transaction) and macrolevel (global trade), we know these legacies of the scientific method applied to economic systems have left the G-20 paralyzed to deal with structural failings evidenced from banking to business to economic policy. An exploration of the structural modalities that impair our current capacity for adaptation and alternative methods for accounting for value is the basis for this inquiry. Integral accounting is proposed as a more suitable method to transition from scarcity-based market models to abundance-based modes of sustainable engagement.     

Mining students’ inquiry actions for understanding of complex systems

This study lies at an intersection between advancing educational data mining methods for detecting students’ knowledge-in-action and the broader question of how conceptual and mathematical forms of knowing interact in exploring complex chemical systems. More specifically, it investigates students’ inquiry actions in three computer-based models of complex chemical systems when their goal is to construct an equation relating physical variables of the system. The study’s participants were 368 high-school students who interacted with the Connected Chemistry (CC1¹) curriculum and completed identical pre- and post-test content knowledge questionnaires. The study explores whether and how students adapt to different mathematical behaviors of the system, examines how these explorations may relate to prior knowledge and learning in terms of conceptual and mathematical models, as well as components relating to understanding systems. Students’ data-collection choices were mined and analyzed showing: (1) In about half the cases, mainly for two out of the three models explored, students conduct mathematically-astute (fit) explorations; (2) A third of the students consistently adapt their strategies to the models’ mathematical behavior; (3) Fit explorations are associated with prior conceptual knowledge, specifically understanding of the system as complex, however, the three explorations’ fitness is predicted by the understanding of distinct sets of systems’ components; (4) Fit explorations are only somewhat associated with learning along complementary dimensions. These results are discussed with respect to 1) the importance of a conceptual understanding regarding individual system elements even when engaged in large-scale quantitative problem solving, 2) how distinct results for the different models relate to previous literature on conceptual understanding and particular affordances of the models, 3) the importance of engaging students in creating mathematical representations of scientific phenomena, as well as 4) educational applications of these results in learning environments.     

Teaching, learning and inquiry strategies using computer technology

This paper reports a case study in a Singapore school where a group of 'at risk' secondary students were given a series of linked computer-based business and consumer projects involving word processing, survey design and analysis, and presentation of results with computer graphics. There were clear benefits in terms of the group's motivation, behaviour and performance on tests, particularly in English. Both quantitative and qualitative data to support the findings are provided. The theoretical underpinnings of constructivism, collaborative inquiry and situated learning provide the conceptual framework for this case study.     

The interstitiality of IT risk: An inquiry into information systems development practices

Information systems development (ISD) has been part of the core of information systems for over 40 years. Throughout its history, the issue of risk has been closely related to ISD projects, and significant efforts have been made by researchers and practitioners to improve their quality. While important headway has been made in assessing and resolving ISD risk, the literature shows that new and salient risks emerge outside the scope of extant risk management regimes. As a consequence, organizations still struggle with leveraging new technology as projects continue to fail at almost the same rate, albeit for different reasons. Understood as the distinction between reality and possibility, risk is inherently intertwined with practice and rooted in the knowledge, goals, power, and values of specific actors in particular contexts. Hence, to understand how risks emerge, we present a longitudinal case study in which we trace the origin and locus of risks in contemporary ISD practices. We draw on these insights to theorize information technology risk as increasingly interstitial, originating from sources positioned in between established practices and therefore outside the scope of conventional risk analyses. In conclusion, we discuss interstitial risks as an important form of emergent risk with implications for both research and practice.     

An inquiry dialogue system

The majority of existing work on agent dialogues considers negotiation, persuasion or deliberation dialogues; we focus on inquiry dialogues, which allow agents to collaborate in order to find new knowledge. We present a general framework for representing dialogues and give the details necessary to generate two subtypes of inquiry dialogue that we define: argument inquiry dialogues allow two agents to share knowledge to jointly construct arguments; warrant inquiry dialogues allow two agents to share knowledge to jointly construct dialectical trees (essentially a tree with an argument at each node in which a child node is a counter argument to its parent). Existing inquiry dialogue systems only model dialogues, meaning they provide a protocol which dictates what the possible legal next moves are but not which of these moves to make. Our system not only includes a dialogue-game style protocol for each subtype of inquiry dialogue that we present, but also a strategy that selects exactly one of the legal moves to make. We propose a benchmark against which we compare our dialogues, being the arguments that can be constructed from the union of the agents’ beliefs, and use this to define soundness and completeness properties that we show hold for all inquiry dialogues generated by our system.     

Inference and the computer understanding of natural language

The notion of computer understanding of natural language is examined relative to inference mechanisms designed to function in a language-free deep conceptual base (Conceptual Dependency). The conceptual analysis of a natural language sentence into this conceptual base, and the nature of the memory which stores and operates upon these conceptual structures are described from both theoretical and practical standpoints. The various types of inferences which can be made during and after the conceptual analysis of a sentence are defined, and a functioning program which performs these inference tasks is described. Actual computer output is included.     

A computer method of understanding ocular fundus images

We have studied some fundamental problems towards the understanding of color ocular fundus images which are used in the mass diagnosis of adult diseases such as hypertension and diabetes.These problems are: the extraction of blood vessels from the retinal background; the recognition of arteries and veins; the detection and analysis of peculiar regions such as hemorrhages, exudates, optic discs and arterio-venous crossings.We propose a computer method for each of these problems and show some experimental results.