Cognitive Plausibility

Invited Keynote Delivered by Dr Fermín Moscoso del Prado Martín

In contrast with a broad corpus of research on the distribution of the frequencies of words, the distribution of the frequencies of phonemic contrasts is remarkably understudied. Using data from over one hundred language varieties, I will discuss that all of them can be modeled a priori by a virtually parameter-free distribution; in other words, just knowing the number of phonemic contrasts in a language variety, one can accurately guess their relative frequencies. Moreover, in situations with extreme undersampling —as would be the case when documenting new, unknown varieties— this knowledge can be used for guessing the expected number of existing contrasts. Finally, I discuss how informational and diachronic constraints give rise to these distributions. Incrementally complex models of phonological change can replicate these distributions with minimum assumptions.

While it is commonplace to retrieve photos showing a particular feature (e.g. through tools such as Google Pictures or Bing Images), spatial approaches for retrieving videos showing a particular feature (e.g. a building) have yet to be established. is article proposes ve ranking algorithms to query georeferenced videos for a speci c feature based on the videos' spatio-temporal metadata. 12 relevance criteria for feature-centric video ranking were compiled from a focus group discussion. From these, four criteria have been selected for implementation: "Feature Depiction", "Feature Illumination", "Feature Visibility Duration", and "Distance to Feature". ese criteria were implemented in ve algorithms and evaluated regarding e ciency and user perceived plausibility. e evaluation suggests that the "Feature Visibility Duration" of the video's viewshed with the queried feature geometry o ers a good trade-o between computationally performant and cognitive plausible ranking. e obtained results are relevant to user-centered approaches for interacting with georeferenced videos. CCS CONCEPTS •Information systems →Information retrieval; •Human-centered computing →Human computer interaction (HCI);

The informativity of a computational model of language acquisition is directly related to how closely it approximates the actual acquisition task, sometimes referred to as the model's cognitive plausibility. We suggest that though every computational model necessarily idealizes the modeled task, an informative language acquisition model can aim to be cognitively plausible in multiple ways. We discuss these cognitive plausibility checkpoints in general terms, and then apply them to a case study in word segmentation, investigating a promising Bayesian segmentation strategy. We create a more cognitively plausible model of this learning strategy which uses an age-appropriate unit of perceptual representation, evaluates the model output in terms of its utility, and incorporates cognitive constraints into the inference process. Our more cognitively plausible model of the Bayesian word segmentation strategy not only yields better performance than previous implementations but also shows more strongly the beneficial effect of cognitive constraints on segmentation. One interpretation of this effect is as a synergy between the naive theories of language structure that infants may have and the cognitive constraints that limit the fidelity of their inference processes, where less accurate inference approximations are better when the underlying assumptions about how words are generated are less accurate. More generally, these results highlight the utility of incorporating cognitive plausibility more fully into computational models of language acquisition.

In the field of cognitive science, the primary means of judging a model's viability is made on the basis of goodness-of-fit between model and human empirical data. Recent developments in model comparison reveal, however, that other criteria should be considered in evaluating the quality of a model. These criteria include model complexity, generalizability, predictive capability, and of course descriptive adequacy. The current investigation seeks to formally compare three variants of a mathematical model for performance prediction. The results raise the issue of how to go about selecting a model when formal comparison methods reveal equivalent values. A possibility briefly proposed at the end of the paper is that cognitive/neural plausibility is an appropriate tiebreaker among otherwise equivalent functional forms.

While it is commonplace to retrieve photos showing a particular feature (e.g. through tools such as Google Pictures or Bing Images), spatial approaches for retrieving videos showing a particular feature (e.g. a building) have yet to be established. This article proposes five ranking algorithms to query georeferenced videos for a specific feature based on the videos' spatio-temporal metadata. 12 relevance criteria for feature-centric video ranking were compiled from a focus group discussion. From these, four criteria have been selected for implementation: "Feature Depiction", "Feature Illumination", "Feature Visibility Duration", and "Distance to Feature". These criteria were implemented in five algorithms and evaluated regarding efficiency and user perceived plausibility. The evaluation suggests that the "Feature Visibility Duration" of the video's viewshed with the queried feature geometry offers a good trade-off between com...

Semantic relatedness (SR) measures form the algorithmic foundation of intelligent technologies in domains ranging from artificial intelligence to human-computer interaction. Although SR has been researched for decades, this work has focused on developing general SR measures rooted in graph and text mining algorithms that perform reasonably well for many different types of concepts. This paper introduces domain-specific SR, which augments general SR by identifying, capturing, and synthesizing domain-specific relationships between concepts. Using the domain of geography as a case study, we show that domain-specific SR -- and even geography-specific signals alone (e.g. distance, containment) without sophisticated graph or text mining algorithms -- significantly outperform the SR state-of-the-art for geographic concepts. In addition to substantially improving SR measures for geospatial technologies, an area that is rapidly increasing in importance, this work also unlocks an important ne...

Semantic relatedness (SR) measures form the algorithmic foundation of intelligent technologies in domains ranging from artificial intelligence to human-computer interaction. Although SR has been researched for decades, this work has focused on developing general SR measures rooted in graph and text mining algorithms that perform reasonably well for many different types of concepts. This paper introduces domain-specific SR, which augments general SR by identifying, capturing, and synthesizing domain-specific relationships between concepts. Using the domain of geography as a case study, we show that domain-specific SR -- and even geography-specific signals alone (e.g. distance, containment) without sophisticated graph or text mining algorithms -- significantly outperform the SR state-of-the-art for geographic concepts. In addition to substantially improving SR measures for geospatial technologies, an area that is rapidly increasing in importance, this work also unlocks an important ne...

Categorization is a central concept for spatial representations in human cognition and artificial intelligence. With the present research, we aimed at building a bridge between those two research fields by asking whether motion categorizations designed in artificial intelligence can inform the psychological understanding of human perception and memory of motion scenes. We report the results of four experiments investigating the influence of Motion-RCC and Motion-OPRA1 categorizations on human perception and memory. Participants viewed simple motion scenes and judged the similarity of transformed scenes with this reference scene. Those transformed scenes differed in none, one or both Motion-RCC and Motion-OPRA1 categories. Importantly, we applied an equal absolute metric change to those transformed scenes. When the reference stimulus and transformed stimuli were visible at the same time (Experiments 1a and 1b: perception), both Motion-OPRA1 and Motion-RCC influenced the similarity judgements, with a stronger influence of Motion-OPRA1 than Motion-RCC. When the participants first memorized the reference stimulus and viewed the transformed stimuli after a short blank (Experiments 2a and 2b: memory), only Motion-OPRA1 had marked influences on the similarity judgements. We conclude that human perception and short-term memory utilize some of the properties underlying those categorizations and given this link between human cognition and qualitative reasoning, we argue for a continued and close multidisciplinary approach to investigating the spatial representation of motion scenes.

In this article, I argue that the artificial components of hybrid bionic systems do not play a direct explanatory role, i.e., in simulative terms, in the overall context of the systems in which they are embedded in. More precisely, I claim that the internal procedures determining the output of such artificial devices, replacing biological tissues and connected to other biological tissues, cannot be used to directly explain the corresponding mechanisms of the biological component(s) they substitute (and therefore cannot be used to explain the local mechanisms determining an overall biological or cognitive function replicated by such bionic models). I ground this analysis on the use of the Minimal Cognitive Grid (MCG), a novel framework proposed in Lieto (Cognitive design for artificial minds, 2021) to rank the epistemological and explanatory status of biologically and cognitively inspred artificial systems. Despite the lack of such a direct mechanistic explanation from the artificial component, however, I also argue that the hybrid bionic systems can have an indirect explanatory role similar to the one played by some AI systems built by using an overall structural design approach (but including the partial adoption of functional components). In particular, the artificial replacement of part(s) of a biological system can provide i) a local functional account of that part(s) in the context of the overall functioning of the hybrid biological-artificial system and ii) global insights about the structural mechanisms of the biological elements connected to such artificial devices.

In geographic information science and semantics, the computation of semantic similarity is widely recognised as key to supporting a vast number of tasks in information integration and retrieval. By contrast, the role of geosemantic relatedness has been largely ignored. In natural language processing, semantic relatedness is often confused with the more specific semantic similarity. In this article, we discuss a notion of geo-semantic relatedness based on Lehrer's semantic fields, and we compare it with geo-semantic similarity. We then describe and validate the Geo Relatedness and Similarity Dataset (GeReSiD), a new open dataset designed to evaluate computational measures of geo-semantic relatedness and similarity. This dataset is larger than existing datasets of this kind, and includes 97 geographic terms combined into 50 term pairs rated by 203 human subjects. GeReSiD is available online and can be used as an evaluation baseline to determine empirically to what degree a given computational model approximates geo-semantic relatedness and similarity.

While it is commonplace to retrieve photos showing a particular feature (e.g. through tools such as Google Pictures or Bing Images), spatial approaches for retrieving videos showing a particular feature (e.g. a building) have yet to be established. This article proposes five ranking algorithms to query georeferenced videos for a specific feature based on the videos' spatio-temporal metadata. 12 relevance criteria for feature-centric video ranking were compiled from a focus group discussion. From these, four criteria have been selected for implementation: "Feature Depiction", "Feature Illumination", "Feature Visibility Duration", and "Distance to Feature". These criteria were implemented in five algorithms and evaluated regarding efficiency and user perceived plausibility. The evaluation suggests that the "Feature Visibility Duration" of the video's viewshed with the queried feature geometry offers a good trade-off between computationally performant and cognitive plausible ranking. The obtained results are relevant to user-centered approaches for interacting with georeferenced videos.

Computational measures of semantic similarity between geographic terms provide valuable support across geographic information retrieval, data mining, and information integration. To date, a wide variety of approaches to geo-semantic similarity have been devised. A judgement of similarity is not intrinsically right or wrong, but obtains a certain degree of cognitive plausibility, depending on how closely it mimics human behaviour. Thus selecting the most appropriate measure for a specific task is a significant challenge. To address this issue, we make an analogy between computational similarity measures and soliciting domain expert opinions, which incorporate a subjective set of beliefs, perceptions, hypotheses, and epistemic biases. Following this analogy, we define the semantic similarity ensemble (SSE) as a composition of different similarity measures, acting as a panel of experts having to reach a decision on the semantic similarity of a set of geographic terms. The approach is evaluated in comparison to human judgements, and results indicate that an SSE performs better than the average of its parts. Although the best member tends to outperform the ensemble, all ensembles outperform the average performance of each ensemble's member. Hence, in contexts where the best measure is unknown, the ensemble provides a more cognitively plausible approach.

Volunteered geographic information (VGI) is generated by heterogenous 'information communities' that co-operate to produce reusable units of geographic knowledge. A consensual lexicon is a key factor to enable this open production model. Lexical definitions help demarcate the boundaries of terms, forming a thin semantic ground on which knowledge can travel. In VGI, lexical definitions often appear to be inconsistent, circular, noisy and highly idiosyncratic. Computing the semantic similarity of these 'volunteered lexical definitions' has a wide range of applications in GIScience, including information retrieval, data mining and information integration. This article describes a knowledge-based approach to quantify the semantic similarity of lexical definitions. Grounded in the recursive intuition that similar terms are described using similar terms, the approach relies on paraphrase-detection techniques and the lexical database WordNet. The cognitive plausibility of the approach is evaluated in the context of the OpenStreetMap (OSM) Semantic Network, obtaining high correlation with human judgements. Guidelines are provided for the practical usage of the approach.

In geographic information science and semantics, the computation of semantic similarity is widely recognised as key to supporting a vast number of tasks in information integration and retrieval. By contrast, the role of geosemantic relatedness has been largely ignored. In natural language processing, semantic relatedness is often confused with the more specific semantic similarity. In this article, we discuss a notion of geo-semantic relatedness based on Lehrer's semantic fields, and we compare it with geo-semantic similarity. We then describe and validate the Geo Relatedness and Similarity Dataset (GeReSiD), a new open dataset designed to evaluate computational measures of geo-semantic relatedness and similarity. This dataset is larger than existing datasets of this kind, and includes 97 geographic terms combined into 50 term pairs rated by 203 human subjects. GeReSiD is available online and can be used as an evaluation baseline to determine empirically to what degree a given computational model approximates geo-semantic relatedness and similarity.

Over the past decade, rapid advances in web technologies, coupled with innovative models of spatial data collection and consumption, have generated a robust growth in geo-referenced information, resulting in spatial information overload. Increasing 'geographic intelligence' in traditional text-based information retrieval has become a prominent approach to respond to this issue and to fulfill users' spatial information needs. Numerous efforts in the Semantic Geospatial Web, Volunteered Geographic Information (VGI), and the Linking Open Data initiative have converged in a constellation of open knowledge bases, freely available online. In this article, we survey these open knowledge bases, focusing on their geospatial dimension. Particular attention is devoted to the crucial issue of the quality of geoknowledge bases, as well as of crowdsourced data. A new knowledge base, the OpenStreetMap Semantic Network, is outlined as our contribution to this area. Research directions in information integration and Geographic Information Retrieval (GIR) are then reviewed, with a critical discussion of their current limitations and future prospects.

In recent years, geographic information has entered the mainstream, deeply altering the pre-existing patterns of its production, distribution, and consumption. Through web mapping, millions of online users utilise spatial data in interactive digital maps. The typical unit of visualisation of geo-data is a viewport, defined as a bi-dimensional image of a map, fixed at a given scale, in a rectangular frame. In a viewport, the user performs analytical tasks, observing individual map features, or drawing high-level judgements about the objects in ...

Graphs have become ubiquitous structures to encode geographic knowledge online. The Semantic Web’s linked open data, folksonomies, wiki websites and open gazetteers can be seen as geo-knowledge graphs, that is labeled graphs whose vertices represent geographic concepts and whose edges encode the relations between concepts. To compute the semantic similarity of concepts in such structures, this article defines the network-lexical similarity measure (NLS). This measure estimates similarity by combining two complementary sources of information: the network similarity of vertices and the semantic similarity of the lexical definitions. NLS is evaluated on the OpenStreetMap Semantic Network, a crowdsourced geo-knowledge graph that describes geographic concepts. The hybrid approach outperforms both network and lexical measures, obtaining very strong correlation with the similarity judgments of human subjects.

Computational measures of semantic similarity between geographic terms pro- vide valuable support across geographic information retrieval, data mining, and information integration. To date, a wide variety of approaches to geo-semantic similarity have been devised. A judgment of similarity is not intrinsically right or wrong, but obtains a certain degree of cognitive plausibility, depending on how closely it mimics human behavior. Thus selecting the most appropriate measure for a specific task is a significant challenge. To address this issue, we make an analogy between computational similarity measures and soliciting domain expert opinions, which incorporate a subjective set of beliefs, perceptions, hypotheses, and epistemic biases. Following this analogy, we define the semantic similarity ensemble (SSE) as a composition of different similarity measures, acting as a panel of experts having to reach a decision on the semantic similarity of a set of geographic terms. The ap- proach is evaluated in comparison to human judgments, and results indicate that an SSE performs better than the average of its parts. Although the best member tends to outperform the ensemble, all ensembles outperform the average performance of each ensemble’s member. Hence, in contexts where the best measure is unknown, the ensemble provides a more cognitively plausible approach.

In geographic information science and semantics, the computation of semantic similarity is widely recognised as key to supporting a vast number of tasks in information integration and retrieval. By contrast, the role of geo-semantic relatedness has been largely ignored. In natural language processing, semantic relatedness is often confused with the more specific semantic similarity. In this article, we discuss a notion of geo-semantic relatedness based on Lehrer’s semantic fields, and we compare it with geo-semantic similarity. We then describe and validate the Geo Relatedness and Similarity Dataset (GeReSiD), a new open dataset designed to evaluate computational measures of geo-semantic relatedness and similarity. This dataset is larger than existing datasets of this kind, and includes 97 geographic terms combined into 50 term pairs rated by 203 human subjects. GeReSiD is available online and can be used as an evaluation baseline to determine empirically to what degree a given computational model approximates geo-semantic relatedness and similarity.

Volunteered geographic information (VGI) is generated by heterogenous ‘information communities’ that co-operate to produce reusable units of geographic knowledge. A consensual lexicon is a key factor to enable this open production model. Lexical definitions help demarcate the boundaries of terms, forming a thin semantic ground on which knowledge can travel. In VGI, lexical definitions often appear to be inconsistent, circular, noisy and highly idiosyncratic. Computing the semantic similarity of these ‘volunteered lexical definitions’ has a wide range of applications in GIScience, including information retrieval, data mining and information integration. This article describes a knowledge-based approach to quantify the semantic similarity of lexical definitions. Grounded in the recursive intuition that similar terms are described using similar terms, the approach relies on paraphrase-detection techniques and the lexical database WordNet. The cognitive plausibility of the approach is evaluated in the context of the OpenStreetMap (OSM) Semantic Network, obtaining high correlation with human judgements. Guidelines are provided for the practical usage of the approach.

A cognitively plausible measure of semantic similarity between geographic concepts is valuable across several areas, including geographic information retrieval, data mining, and ontology alignment. Semantic similarity measures are not intrinsically right or wrong, but obtain a certain degree of cognitive plausibility in the context of a given application. A similarity measure can therefore be seen as a domain expert summoned to judge the similarity of a pair of concepts according to her subjective set of beliefs, perceptions, hypotheses, and epistemic biases. Following this analogy, we first define the similarity jury as a panel of experts having to reach a decision on the semantic similarity of a set of geographic concepts. Second, we have conducted an evaluation of 8 WordNet-based semantic similarity measures on a subset of OpenStreetMap geographic concepts. This empirical evidence indicates that a jury tends to perform better than individual experts, but the best expert often outperforms the jury. In some cases, the jury obtains higher cognitive plausibility than its best expert.

In recent years, a web phenomenon known as Volunteered Geographic Information (VGI) has produced large crowdsourced geographic data sets. OpenStreetMap (OSM), the leading VGI project, aims at building an open-content world map through user contributions. OSM semantics consists of a set of properties (called ‘tags’) describing geographic classes, whose usage is defined by project contributors on a dedicated Wiki website. Because of its simple and open semantic structure, the OSM approach often results in noisy and ambiguous data, limiting its usability for analysis in information retrieval, recommender systems and data mining. Devising a mechanism for computing the semantic similarity of the OSM geographic classes can help alleviate this semantic gap. The contribution of this paper is twofold. It consists of (1) the development of the OSM Semantic Network by means of a web crawler tailored to the OSM Wiki website; this semantic network can be used to compute semantic similarity through co-citation measures, providing a novel semantic tool for OSM and GIS communities; (2) a study of the cognitive plausibility (i.e. the ability to replicate human judgement) of co-citation algorithms when applied to the computation of semantic similarity of geographic concepts. Empirical evidence supports the usage of co-citation algorithms—SimRank showing the highest plausibility—to compute concept similarity in a crowdsourced semantic network.

In recent years, geographic information has entered the mainstream, deeply altering the pre-existing patterns of its production, distribution, and consumption. Through web mapping, millions of online users utilise spatial data in interactive digital maps. The typical unit of visualisation of geo-data is a viewport, defined as a bi-dimensional image of a map, fixed at a given scale, in a rectangular frame. In a viewport, the user performs analytical tasks, observing individual map features, or drawing high-level judgements about the objects in the viewport as a whole. Current geographic information retrieval (GIR) systems aim at facilitating analytical tasks, and little emphasis is put on the retrieval and indexing of visualised units, i.e. viewports. In this paper we outline a holistic, viewport-based GIR system, offering an alternative approach to feature-based GIR. Such a system indexes viewports, rather than individual map features, extracting descriptors of their high-level, overall semantics in a vector space model. This approach allows for efficient comparison, classification, clustering, and indexing of viewports. A case study describes in detail how our GIR system models viewports representing geographical locations in Ireland. The results indicate advantages and limitations of the viewport-based approach, which allows for a novel exploration of geographic data, using holistic semantics.