Emergence of analogy from relation learning.
Ontology highlight
ABSTRACT:
SUBMITTER: Lu H
PROVIDER: S-EPMC6410800 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:People are habitual explanation generators. At its most mundane, our propensity to explain allows us to infer that we should not drink milk that smells sour; at the other extreme, it allows us to establish facts (e.g., theorems in mathematical logic) whose truth was not even known prior to the existence of the explanation (proof). What do the cognitive operations underlying the inference that the milk is sour have in common with the proof that, say, the square root of two is irrational? Our ability to generate explanations bears striking similarities to our ability to make analogies. Both reflect a capacity to generate inferences and generalizations that go beyond the featural similarities between a novel problem and familiar problems in terms of which the novel problem may be understood. However, a notable difference between analogy-making and explanation-generation is that the former is a process in which a single source situation is used to reason about a single target, whereas the latter often requires the reasoner to integrate multiple sources of knowledge. This seemingly small difference poses a challenge to the task of marshaling our understanding of analogical reasoning to understanding explanation. We describe a model of explanation, derived from a model of analogy, adapted to permit systematic violations of this one-to-one mapping constraint. Simulation results demonstrate that the resulting model can generate explanations for novel explananda and that, like the explanations generated by human reasoners, these explanations vary in their coherence.
| S-EPMC4222223 | biostudies-literature
Project description:Analogical reasoning has been hypothesized to critically depend upon working memory through correlational data, but less work has tested this relationship through experimental manipulation. An opportunity for examining the connection between working memory and analogical reasoning has emerged from the growing, although somewhat controversial, body of literature suggests complex working memory training can sometimes lead to working memory improvements that transfer to novel working memory tasks. This study investigated whether working memory improvements, if replicated, would increase analogical reasoning ability. We assessed participants' performance on verbal and visual analogy tasks after a complex working memory training program incorporating verbal and spatial tasks. Participants' improvements on the working memory training tasks transferred to other short-term and working memory tasks, supporting the possibility of broad effects of working memory training. However, we found no effects on analogical reasoning. We propose several possible explanations for the lack of an impact of working memory improvements on analogical reasoning.
| S-EPMC4154731 | biostudies-literature
Project description:Mie scattering is an optical phenomenon that appears when electromagnetic waves, in particular light, are elastically scattered at a spherical or cylindrical object. A transfer of this phenomenon onto electron states in ballistic graphene has been proposed theoretically, assuming a well-defined incident wave scattered by a perfectly cylindrical nanometer scaled potential, but experimental fingerprints are lacking. We present an experimental demonstration of an electrical analogue to Mie scattering by using graphene as a conductor, and circular potentials arranged in a square two-dimensional array. The tabletop experiment is carried out under seemingly unfavourable conditions of diffusive transport at room-temperature. Nonetheless, when a canted arrangement of the array with respect to the incident current is chosen, cascaded Mie scattering results robustly in a transverse voltage. Its response on electrostatic gating and variation of potentials convincingly underscores Mie scattering as underlying mechanism. The findings presented here encourage the design of functional electronic metamaterials.
| S-EPMC5052640 | biostudies-literature
Project description:To behave adaptively, animals must learn to predict future reward, or value. To do this, animals are thought to learn reward predictions using reinforcement learning. However, in contrast to classical models, animals must learn to estimate value using only incomplete state information. Previous work suggests that animals estimate value in partially observable tasks by first forming "beliefs"-optimal Bayesian estimates of the hidden states in the task. Although this is one way to solve the problem of partial observability, it is not the only way, nor is it the most computationally scalable solution in complex, real-world environments. Here we show that a recurrent neural network (RNN) can learn to estimate value directly from observations, generating reward prediction errors that resemble those observed experimentally, without any explicit objective of estimating beliefs. We integrate statistical, functional, and dynamical systems perspectives on beliefs to show that the RNN's learned representation encodes belief information, but only when the RNN's capacity is sufficiently large. These results illustrate how animals can estimate value in tasks without explicitly estimating beliefs, yielding a representation useful for systems with limited capacity.
| S-EPMC10513382 | biostudies-literature
Project description:To better understand the benefits of using analogy and explicit instructions, the underlying cognitive mechanism remains to be explored. The concept of chunking provides a promising approach to the cognitive mechanism of instructions and can be approximated by analyzing athletes' mental representations. The purpose of this study was to investigate the influence of analogy and explicit instructions on performance and the cognitive representations of the tennis serve in intermediate participants over the course of a 5-week training period. Junior tennis players (N = 44; M = 11.5 years) were tested on their tennis serve and, based on their initial performance and their individual error patterns, assigned to one of three groups: an analogy group (N = 15), an explicit group (N = 15), or a control group (N = 14). Their performance and their mental representation structures were assessed prior to and after the 5-week training period and again after a retention period of 14 days. Independent of group, findings demonstrated higher velocity from pretest to posttest. Participants in both the analogy and the explicit group showed enhanced accuracy over time and more functional mental representation structures. Thus, both analogy instruction and explicit instruction helped to structure mental representations in their long-term memory.
| S-EPMC7019697 | biostudies-literature
Project description:BACKGROUND:How to treat a disease remains to be the most common type of clinical question. Obtaining evidence-based answers from biomedical literature is difficult. Analogical reasoning with embeddings from deep learning (embedding analogies) may extract such biomedical facts, although the state-of-the-art focuses on pair-based proportional (pairwise) analogies such as man:woman::king:queen ("queen = -man +king +woman"). OBJECTIVE:This study aimed to systematically extract disease treatment statements with a Semantic Deep Learning (SemDeep) approach underpinned by prior knowledge and another type of 4-term analogy (other than pairwise). METHODS:As preliminaries, we investigated Continuous Bag-of-Words (CBOW) embedding analogies in a common-English corpus with five lines of text and observed a type of 4-term analogy (not pairwise) applying the 3CosAdd formula and relating the semantic fields person and death: "dagger = -Romeo +die +died" (search query: -Romeo +die +died). Our SemDeep approach worked with pre-existing items of knowledge (what is known) to make inferences sanctioned by a 4-term analogy (search query -x +z1 +z2) from CBOW and Skip-gram embeddings created with a PubMed systematic reviews subset (PMSB dataset). Stage1: Knowledge acquisition. Obtaining a set of terms, candidate y, from embeddings using vector arithmetic. Some n-gram pairs from the cosine and validated with evidence (prior knowledge) are the input for the 3cosAdd, seeking a type of 4-term analogy relating the semantic fields disease and treatment. Stage 2: Knowledge organization. Identification of candidates sanctioned by the analogy belonging to the semantic field treatment and mapping these candidates to unified medical language system Metathesaurus concepts with MetaMap. A concept pair is a brief disease treatment statement (biomedical fact). Stage 3: Knowledge validation. An evidence-based evaluation followed by human validation of biomedical facts potentially useful for clinicians. RESULTS:We obtained 5352 n-gram pairs from 446 search queries by applying the 3CosAdd. The microaveraging performance of MetaMap for candidate y belonging to the semantic field treatment was F-measure=80.00% (precision=77.00%, recall=83.25%). We developed an empirical heuristic with some predictive power for clinical winners, that is, search queries bringing candidate y with evidence of a therapeutic intent for target disease x. The search queries -asthma +inhaled_corticosteroids +inhaled_corticosteroid and -epilepsy +valproate +antiepileptic_drug were clinical winners, finding eight evidence-based beneficial treatments. CONCLUSIONS:Extracting treatments with therapeutic intent by analogical reasoning from embeddings (423K n-grams from the PMSB dataset) is an ambitious goal. Our SemDeep approach is knowledge-based, underpinned by embedding analogies that exploit prior knowledge. Biomedical facts from embedding analogies (4-term type, not pairwise) are potentially useful for clinicians. The heuristic offers a practical way to discover beneficial treatments for well-known diseases. Learning from deep learning models does not require a massive amount of data. Embedding analogies are not limited to pairwise analogies; hence, analogical reasoning with embeddings is underexploited.
| S-EPMC7441383 | biostudies-literature
Project description:We present an interactive graphical approach for the explicit specification of semantics for volume visualization. This explicit and graphical specification of semantics for volumetric features allows us to visually assign meaning to both input and output parameters of the visualization mapping. This is in contrast to the implicit way of specifying semantics using transfer functions. In particular, we demonstrate how to realize a dynamic specification of semantics which allows to flexibly explore a wide range of mappings. Our approach is based on three concepts. First, we use semantic shader augmentation to automatically add rule-based rendering functionality to static visualization mappings in a shader program, while preserving the visual abstraction that the initial shader encodes. With this technique we extend recent developments that define a mapping between data attributes and visual attributes with rules, which are evaluated using fuzzy logic. Second, we let users define the semantics by analogy through brushing on renderings of the data attributes of interest. Third, the rules are specified graphically in an interface that provides visual clues for potential modifications. Together, the presented methods offer a high degree of freedom in the specification and exploration of rule-based mappings and avoid the limitations of a linguistic rule formulation.
| S-EPMC3617880 | biostudies-literature
Project description:The use of analogy is an important component of human cognition. The type of analogy we produce and communicate depends heavily on a number of factors, such as the setting, the level of domain expertise present, and the speaker's goal or intent. In this observational study, we recorded economics experts during scientific discussion and examined the categorical distance and structural depth of the analogies they produced. We also sought to characterize the purpose of the analogies that were generated. Our results supported previous conclusions about the infrequency of superficial similarity in subject-generated analogs, but also showed that distance and depth characteristics were more evenly balanced than in previous observational studies. This finding was likely due to the nature of the goals of the participants, as well as the broader nature of their expertise. An analysis of analogical purpose indicated that the generation of concrete source examples of more general target concepts was most prevalent. We also noted frequent instances of analogies intended to form visual images of source concepts. Other common purposes for analogies were the addition of colorful speech, inclusion (i.e., subsumption) of a target into a source concept, or differentiation between source and target concepts. We found no association between depth and either of the other two characteristics, but our findings suggest a relationship between purpose and distance; i.e., that visual imagery typically entailed an outside-domain source whereas exemplification was most frequently accomplished using within-domain analogies. Overall, we observed a rich and diverse set of spontaneously produced analogical comparisons. The high degree of expertise within the observed group along with the richly comparative nature of the economics discipline likely contributed to this analogical abundance.
| S-EPMC4244599 | biostudies-literature
Project description:Traditionally, investigations in the area of stimulus equivalence have employed humans as experimental participants. Recently, however, artificial neural network models (often referred to as connectionist models [CMs]) have been developed to simulate performances seen among human participants when training various types of stimulus relations. Two types of neural network models have shown particular promise in recent years. RELNET has demonstrated its capacity to approximate human acquisition of stimulus relations using simulated matching-to-sample (MTS) procedures (e.g., Lyddy & Barnes-Holmes Journal of Speech and Language Pathology and Applied Behavior Analysis, 2, 14-24, 2007). Other newly developed connectionist algorithms train stimulus relations by way of compound stimuli (e.g., Tovar & Chavez The Psychological Record, 62, 747-762, 2012; Vernucio & Debert The Psychological Record, 66, 439-449, 2016). What makes all of these CMs interesting to many behavioral researchers is their apparent ability to simulate the acquisition of diversified stimulus relations as an analogue to human learning; that is, neural networks learn over a series of training epochs such that these models become capable of deriving novel or untrained stimulus relations. With the goal of explaining these quickly evolving approaches to practical and experimental endeavors in behavior analysis, we offer an overview of existing CMs as they apply to behavior-analytic theory and practice. We provide a brief overview of derived stimulus relations as applied to human academic remediation, and we argue that human and simulated human investigations have symbiotic experimental potential. Additionally, we provide a working example of a neural network referred to as emergent virtual analytics (EVA). This model demonstrates a process by which artificial neural networks can be employed by behavior-analytic researchers to understand, simulate, and predict derived stimulus relations made by human participants.
| S-EPMC6701726 | biostudies-literature
Project description:One aspect of consciousness phenomena, the temporal emergence of visual awareness, has been subject of a controversial debate. How can visual awareness, that is the experiential quality of visual stimuli, be characterized best? Is there a sharp discontinuous or dichotomous transition between unaware and fully aware states, or does awareness emerge gradually encompassing intermediate states? Previous studies yielded conflicting results and supported both dichotomous and gradual views. It is well conceivable that these conflicting results are more than noise, but reflect the dynamic nature of the temporal emergence of visual awareness. Using a psychophysical approach, the present research tested whether the emergence of visual awareness is context-dependent with a temporal two-alternative forced choice task. During backward masking of word targets, it was assessed whether the relative temporal sequence of stimulus thresholds is modulated by the task (stimulus presence, letter case, lexical decision, and semantic category) and by mask type. Four masks with different similarity to the target features were created. Psychophysical functions were then fitted to the accuracy data in the different task conditions as a function of the stimulus mask SOA in order to determine the inflection point (conscious threshold of each feature) and slope of the psychophysical function (transition from unaware to aware within each feature). Depending on feature-mask similarity, thresholds in the different tasks were highly dispersed suggesting a graded transition from unawareness to awareness or had less differentiated thresholds indicating that clusters of features probed by the tasks quite simultaneously contribute to the percept. The latter observation, although not compatible with the notion of a sharp all-or-none transition between unaware and aware states, suggests a less gradual or more discontinuous emergence of awareness. Analyses of slopes of the fitted psychophysical functions also indicated that the emergence of awareness of single features is variable and might be influenced by the continuity of the feature dimensions. The present work thus suggests that the emergence of awareness is neither purely gradual nor dichotomous, but highly dynamic depending on the task and mask type.
| S-EPMC5334328 | biostudies-literature