Specific Inference Research Articles

Assessing children's inference generation: What do tests of reading comprehension measure?

Previous research suggests that children with specific comprehension difficulties have problems with the generation of inferences. This raises important questions as to whether poor comprehenders have poor comprehension skills generally, or whether their problems are confined to specific inference types. The main aims of the study were (a) using two commonly used tests of reading comprehension to classify the questions requiring the generation of inferences, and (b) to investigate the relative performance of skilled and less-skilled comprehenders on questions tapping different inference types. The performance of 10 poor comprehenders (mean age 110.06 months) was compared with the performance of 10 normal readers (mean age 112.78 months) on two tests of reading comprehension. A qualitative analysis of the NARA II (form 1) and the WORD comprehension subtest was carried out. Participants were then administered the NARA II, WORD comprehension subtest and a test of non-word reading. The NARA II was heavily reliant on the generation of knowledge-based inferences, while the WORD comprehension subtest was biased towards the retention of literal information. Children identified by the NARA II as having comprehension difficulties performed in the normal range on the WORD comprehension subtests. Further, children with comprehension difficulties performed poorly on questions requiring the generation of knowledge-based and elaborative inferences. However, they were able to answer questions requiring attention to literal information or use of cohesive devices at a level comparable to normal readers. Different reading tests tap different types of inferencing skills. Lessskilled comprehenders have particular difficulty applying real-world knowledge to a text during reading, and this has implications for the formulation of effective intervention strategies.

Seminormalizing a default theory

Most of the work in default logic is about default theories that are completely specified. In this category are the proposals of appropriate semantics for default logic, the characterizations of the complexity of reasoning with a default theory, the algorithms for finding consequences of default theories, etc. Relatively little attention has been paid to the process of building a default theory, and most of the work on this topic is about translating knowledge bases from other formalisms (such as circumscription, autoepistemic logic, and action description languages) into default logic. This paper is about expressing knowledge in default logic. In particular, we assume that defaults are initially formulated as normal, and are then corrected using specific inference examples.

Utility of electroencephalogram in altered states of consciousness in intensive care unit patients

BACKGROUND: EEG is an investigative tool for assessing cerebral activity. Although certain EEG patterns may have a specific diagnostic or prognostic inference, they may not be precise for any sole etiology in majority of cases and may need clinical correlation. OBJECTIVE: Aim of this study was to assess the severity and prognosis of cerebral dysfunction in patients admitted to Intensive Care Unit (ICU) and to evaluate the incidence of non-convulsive status epilepticus (NCSE). DESIGN: A prospective study, wherein we analyzed EEG characteristics in a series of 70 patients. SETTING: A tertiary care hospital in Mumbai, India. PATIENTS: EEG characteristics of 70 patients admitted in ICU over a period of 9 months were comprehensively analyzed. These patients were clinically examined and a questionnaire was completed without knowledge of the EEG findings. EEGs were requested for by neurologist or intensivist and our inclusion criteria were (i) patients with altered sensorium of varying etiology, (ii) unconscious patients at risk for non-convulsive status epilepticus (those with a history of epilepsy), and (iii) unconscious patients with involuntary jerky eye movements. RESULTS: Of the various clinical presentations on ICU admission, there were 20 patients with seizures, 15 with metabolic disorders, 13 with infective causes, 9 with hypoxia, 9 with cerebro-vascular accident on presentation, 1 patient with alcohol/drug overdose, 2 with intra-cerebral space occupying lesion and 1 with ambiguous etiology on admission (there being an overlap among the presentation). Mean duration from presentation to performing EEG was 13 hours. 64 (91.42%) patients had abnormal EEGs. 32(50%) patients had EEG slowing and 4(6.25%) patient had electro cerebral inactivity. Eleven (21.87%) patients had epileptiform activity on the EEG of which seven did not have overt seizures (NCSE). Follow-up EEGs of these patients showed resolution of the epileptiform activity. CONCLUSIONS: EEG is useful in patients admitted to ICU in diagnosing NCSE and various other conditions. Emergent EEG study in obtunded patients provides valuable diagnostic and prognostic information.

Hardware/software codesign of configurable fuzzy control systems

Fuzzy inference techniques are an attractive and well-established approach for solving control problems. This is mainly due to their inherent ability to obtain robust, low-cost controllers from the intuitive (and usually ambiguous or incomplete) linguistic rules used by human operators when describing the control process. This paper focuses on the hardware/software codesign of configurable fuzzy control systems. Two prototype systems implemented on general-purpose development boards are presented. In both of them, hardware components are based on specific and configurable fuzzy inference architecture whereas software tasks are supported by a microcontroller. The first prototype uses an off-the-shelf microcontroller and a low-complexity Xilinx XC4005XL field programmable gate array (FPGA). The second one is implemented as a system on programmable chip (SoPC), integrating the microcontroller together with the fuzzy hardware architecture and its interface circuits into a Xilinx Spartan2E200 FPGA.

Hardware/software codesign of configurable fuzzy control systems

Abstract Fuzzy inference techniques are an attractive and well-established approach for solving control problems. This is mainly due to their inherent ability to obtain robust, low-cost controllers from the intuitive (and usually ambiguous or incomplete) linguistic rules used by human operators when describing the control process. This paper focuses on the hardware/software codesign of configurable fuzzy control systems. Two prototype systems implemented on general-purpose development boards are presented. In both of them, hardware components are based on specific and configurable fuzzy inference architecture whereas software tasks are supported by a microcontroller. The first prototype uses an off-the-shelf microcontroller and a low-complexity Xilinx XC4005XL field programmable gate array (FPGA). The second one is implemented as a system on programmable chip (SoPC), integrating the microcontroller together with the fuzzy hardware architecture and its interface circuits into a Xilinx Spartan2E200 FPGA.

Semantic formalization of interactive reasoning functionality

In this article a semantic framework is developed that provides a logical description of the functionality of an interactive reasoning process. The concept of functionality description defines the functionality of a reasoning process abstracted from specific inference relations or knowledge bases. Moreover, a domain description is formalized. A number of properties of a functionality description are identified and related to properties of the domain. It is established under which conditions a functionality can be implemented by an inference relation and a knowledge base. © 2002 Wiley Periodicals, Inc.

On generalized modus ponens with multiple rules and a residuated implication

We consider a multiple-rule, generalized modus ponens inference scheme, with an interpretation based on compositional rule of inference (CRI) and a residuated implication. We show that such a system is equivalent, as far as CRI is concerned, to a system that satisfies the “basic requirement for fuzzy reasoning”, proposed by Turksen and Tian (Fuzzy Sets and Systems 58 (1993) 3–40). We establish an analogous conclusion for an alternative interpretation (due to Magrez and Smets) of generalized modus ponens, which we call consequent dilation rule (CDR). This method is computationally faster than CRI, but it produces less specific inference results.

Knowledge-based spectroscopic assignment

Studying the energy level structure of different molecules by means of spectroscopy is one of the fields in physics and chemistry. In this paper, we describe a knowledge-based system, molecular spectroscopic assignment assistant (MOSAA), which can assist researchers in the assignment of the far-infrared and infrared absorption peaks of the molecules in question by using the spectral information provided, along with a knowledge base containing known ground-state energy levels of the given molecules and spectroscopic assignment rules provided by researchers. A total of 313 rules together with their associated components (i.e., parameters, functions and properties) compose the knowledge base for MOSAA. A specific inference engine was designed to handle the “trial and error” process required for spectroscopic assignment. An overview of MOSAA as well as the testing results are discussed in this paper. Testing of the CH 3 18OH and CD 3 16OH spectra showed that MOSAA successfully assigned 11 testing series for each spectrum in 1/8 of the time on average required to do a manual assignment. Two series of the CD 3 16OH spectrum assigned by MOSAA were missed in the original manual assignment.

Stochastic model of the overdispersion in the place cell discharge

The spontaneous firing activity of the place cells reflects the position of an experimental animal in its arena. The firing rate is high inside a part of the arena, called the firing field, and low outside. It is generally accepted concept that this is the way in which the hippocampus stores a map of the environment. This well known fact was recently reinvestigated [Fenton, A.A., Muller, R.U., 1998. Proc. Natl. Acad. Sci. USA 95, 3182–3187] and it was found that while the activity was highly reliable in position, it did not retain the same reliability in time. The number of action potentials fired during different passes through the firing field were substantially different (overdispersion). We present a mathematical model based on a doubly stochastic Poisson process which is able to reproduce the experimental findings. Further, it enables us to propose specific statistical inference on the experiments in aim to verify data and model compatibility. The model permits to speculate about the neural mechanisms leading to the overdispersion in the activity of the hippocampal place cells. Namely, the statistical variation of the intensity of firing can be achieved, for example, by introducing a hierarchical structure into the local neural network.

How Do Adolescents Process Advertisements? The Influence of Ad Characteristics, Processing Objective, and Gender

This study investigates the influences of print advertisements on the affective and cognitive responses of adolescents. Junior and senior high school males (n = 111) and females (n = 84) were randomly assigned to either a low- or high-elaboration condition to process primarily visual and primarily verbal print advertisements. The students then responded to questions measuring three dependent variables—memory of specific facts, inference, and emotional response. Three-way ANOVA results indicated that predominantly visual advertisements elicited memory of more facts, more inferencing, and more intense emotional responses than predominantly verbal ads. In addition, females remembered more facts, made more inferences, reported stronger emotional responses, and detected the explicit claim of the ad more frequently than males. Finally, students in the high-elaboration condition remembered more details than students in the low-elaboration condition. The results are discussed in terms of implications for advertising media literacy.

Intelligent GIS: Architectural Issues and Implementation Methods

The paper presents an intelligent GIS architecture that enables us to extend GIS func- tionality by using domain specific knowledge and inference engine. In this architecture, an intelli- gent agent monitors events, which occur in the GIS environment, and execute tasks depending on user's actions. The intelligent agent includes an expert system shell and knowledge base. A hybrid knowledge representation method is used that integrates rule-based, object-oriented, and procedural knowledge representations.

Parallel fuzzy information processing system

During the past decade, several specific hardwares for the fast fuzzy inference have been developed. However, most of them are dedicated to a specific inference method and thus cannot support fuzzy set operations. In this paper, we present a hardware architecture called KAFA (KAIST fuzzy accelerator) which provides various fuzzy inference methods and fuzzy set operations. The architecture has SIMD structure, which consists of two parts: a system control unit (MC: main controller) and an arithmetic unit (FPE: fuzzy processing element). The prototype of the proposed architecture was developed by using FPGA chips. Each FPE has a maximum speed of 5M FLOPS (fuzzy logic operations per second). As the KAFA contains 128 FPEs, we achieve a peak peformance of 640M FLOPS. The speed of the KAFA holds promise for the development of new fuzzy application systems.

A hybrid architecture for text comprehension

O'Brien et al. (1988) reported that readers generated elaborative inferences only when a text contained characteristics (a strong biasing context or a demand sentence) that made it easy to predict the specific inference that a reader would draw, and virtually eliminated the possibility of the inference being discon-firmed. Garrod et al. (1990), however, offered two qualifications to these conclusions. First, the two text characteristics manipulated may have produced different types of elaborative inferencing: a biasing context results in a passive form of elaborative inferencing, involving setting up a context of interpretation, whereas the presence of a demand sentence invites the reader to actively predict a subsequent expression. Secondly, clear evidence for either type of inference will be apparent only with truly anaphoric materials. This work describes how a passive form of elaborative inferencing, reported by Garrod et al, may be implemented in a connectionist manner. We take the connectionist model proposed by Shastri and Ajjanagadde (1993) in order to represent a text in the form of a network. Next we analyze and discuss how an attentional focus could operate with the proposed reasoner system in order to deal with inference control and anaphora resolution (i.e., antecedent activation), during text understanding. Our own suggestion is that the system proposed by Shastri and Ajjanagadde could be extended by incorporating focus in order to apply it to some open problems related to text processing. However, this extension presents implementational problems due to its local character. Distributed models seem to be in a better position to deal with the shift of the attentional states, given that these systems can learn. Particularly, the model proposed by Sun (1994) is close to our aims, and so we will discuss its implications in order to design a cognitive architecture for text comprehension.

Surface cues and robust inference as a basis for the early acquisition of subcategorization frames

How could children possibly acquire their first subcategorization frames? The hypothesis that they directly observe the syntactic structures of the utterances they hear raises two questions. First, how can children parse input utterances reliably without already knowing the syntactic properties of all the words in them? Second, how do children survive the ungrammatical or misconstrued utterances they inevitably encounter? This paper suggests a specific inference algorithm that substantially reduces the effects of ungrammatical or misconstrued input. Since children must have some such inference procedure, they can use approximate cues to determine syntactic structure. In particular, they can use string-local surface cues rather than global constraints. Such cues make it possible to discover relevant syntactic structure in an utterance without already knowing all the words in it. This paper suggests a possible set of cues for English subcategorization frames that assumes only the ability to detect the ends of utterances and knowledge of a few function morphemes and proper names. Simulation experiments on naturally occuring, child-directed English show that these cues, combined with the proposed inference mechanism, do surprisingly well at discovering subcategorization frames.

A mathematical framework to implement statistical databases

The objective of this paper is to define a mathematical framework for statistical Databases (SDB) and to show how it can be implemented. Matrix algebra is used as a tool in the representation of statistical databases. The matrix algebra operations are considered the building blocks for the implementation of inference control methods. The rigor of this approach can be very beneficial for further studies on this topic. The implementation of Inference Control Methods for SDBs can be greatly simplified if this approach is used. Inference Control methods release aggregate information from a SDB about large subsets of records in a form that attempts to protect individual privacy. The implementation of specific Inference Control Methods (ICM) will be presented in a future publication. Several examples of typical operations on SDBs are presented in the text. SAS/IML was used to implement the ideas proposed here. The samples were prepared and tested using a personal computer (486 IBM/PC compatible) and a mainframe (IBM 4381) available at CSU, Chico.

Dimensions of knowledge sharing and reuse

Many workers in medical informatics are seeking to reuse knowledge in new applications and to share encoded knowledge across software environments. Knowledge reuse involves many dimensions, including the reapplication of lexicons, ontologies, inference syntax, tasks, and problem-solving methods. Principal obstacles to all current work in knowledge sharing involve the difficulties of achieving consensus regarding what knowledge representations mean, of enumerating the context features and background knowledge required to ascribe meaning to a particular knowledge representation, and of describing knowledge independent of specific interpreters or inference engines. Progress in the area of knowledge sharing will necessitate more practical experience with attempts to interchange knowledge as well as better tools for viewing and editing knowledge representations at appropriate levels of abstraction. The PROTÉGÉ-II project is one attempt to provide a knowledge-base authoring environment in which developers can experiment with the reuse of knowledge-level problemsolving methods, task models, and domain ontologies.

A critique of two methods for assessing the nutrient adequacy of diets

The adequacy of diets can be assessed using several analytical approaches. This paper reviews two methods of assessment: a cutoff method, which estimates the percentage of the population having usual intakes below a given value; and a probability method, which assesses the percentage of the population whose usual intakes are below their individual requirements. First, the concept of usual nutrient intakes and the problems associated with estimating usual intake distributions are discussed. Next, the two methods of dietary assessment and their related assumptions are described and compared. The more specific inference of the probability method is shown to rely on its assumptions and data that are currently not available. While the cutoff method is simpler, its use may result in misclassification errors and its estimates are highly influenced by the cut-off standard selected.

Knowledge based Incipient Fault Diagnosis of Industrial Robots

A computer assisted incipient fault diagnosis system (CAFD) for industrial robots is intro-duced, with the objective to detect and diagnose faults in the mechanical part of the devices at a rela-tively early stage to prevent subsequent damages. For this purpose a suitable combination of analytical and heuristic tools was developed. First the analytical symptom generation procedure is described. It is comprised of the detailed matnematical modeling of the robot's different axes. The parameters of the models directly represent characteristic physical quantities (process coefficients) being identified by specific continuous-time parameter estimation algorithms. The estimated quantities then undergo a statistical evaluation. Deviation of coefficients from nominal values are considered as symptoms. The subsequent heuristic evaluation processes these symptoms based on specific fault-symptom-trees and knowledge of fault statistics and process history using a specific inference mechanism. The developed diagnosis system was realized on a PC with process interface. Experimental results obtained from an industrial robot show the feasibility and efficiency of the devised method. Practical issues concerning the applicability of the incipient fault diagnosis system are finally discussed.

Working‐memory capacity and the use of elaborative inferences in text comprehension

The purpose of this research was to determine if individual differences in working‐memory capacity are related to the ways readers use inferences to facilitate text comprehension. Two groups of subjects, who differed in working‐memory span, read difficult narrative passages a few sentences at a time. The subjects furnished “thinking out loud” protocols of their emerging interpretations. Idea units from the subjects’ protocols were categorized with particular attention to those idea units that expressed a general or specific elaborative inference. Several differences between the two groups of subjects emerged. Low‐memory‐span subjects produced significantly more specific elaborations than the high‐span readers. In addition, most of the specific elaborations that were produced by the high‐span readers were toward the end of a passage. Low‐span readers had a more even distribution of specific elaborations throughout their protocols. Thus, readers with adequate working‐memory capacity can keep their interpret...

Reasoning situated in time I: basic concepts*

The needs of a real-time reasoner situated in an environment may make it appropriate to view error-correction and non-monotonicity as much the same thing. This has led us to formulate situated (or step) logic, an approach to reasoning in which the formalism has a kind of real-time self-reference that affects the course of deduction itself. Here we seek to motivate this as a useful vehicle for exploring certain issues in commonsense reasoning. In particular, a chief drawback of more traditional logics is avoided: from a contradiction we do not have all wffs swamping the (growing) conclusion set. Rather, we seek potentially inconsistent, but nevertheless useful, logics where the real-time self-referential feature allows a direct contradiction to be spotted and corrective action taken, as part of the same system of reasoning. Some specific inference mechanisms for real-time default reasoning are suggested, notably a form of introspection relevant to default reasoning. Special treatment of ‘‘now’’ and of contradictions are the main technical devices here. We illustrate this with a computer-implemented real-time solution to R. Moore’s Brother Problem.