Commonsense reasoning about containers using radically incomplete information

Davis, Ernest; Marcus, Gary; Frazier-Logue, Noah

doi:10.1016/j.artint.2017.03.004

Cited by 25 publications

(24 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Region-based approach is also closer to the everyday common-sense reasoning about physical world (the naïve worldview; cp. (Hayes, 1985), (Aurnague and Vieu, 1993), (Vieu, 1997), (Davis et al, 2017)) and therefore is a feasible framework for AI systems.…”

Section: Rcc+f Size Distance Occlusion As the Framework Of Modellimentioning

confidence: 99%

Topologic and Geometric Structure of Spatial Relations in Latvian: an Experimental Analysis of RCC

Šķilters¹,

Zariņa²,

Žilinskaitė-Šinkūnienė³

et al. 2020

BJMC

View full text Add to dashboard Cite

In this study we experimentally test topological and geometric relations as encoded in Latvian. The task of the subject is to describe different combinations of two geometric objects presented in a randomized order. For the in-group experiment two circlesdark and light-were used according to the topological principles of Region Connection Calculus further extended with simple relational variables representing proximity, orientation, object size, and partial occlusion. The results show that both topological and geometric features determine the number of words used in the description of the respective relations and accuracy of the description. Further, we explored the most common words used for the description of general spatial relations and tested the differences associated with the experimental variables. It was concluded that topological and geometric relations matter in the linguistic representation of space but to a differing degree. Our results also indicate that spatial relations at the linguistic level are represented categorically and rarely encode fine-grained information.

show abstract

Section: Rcc+f Size Distance Occlusion As the Framework Of Modellimentioning

confidence: 99%

Topologic and Geometric Structure of Spatial Relations in Latvian: an Experimental Analysis of RCC

Šķilters¹,

Zariņa²,

Žilinskaitė-Šinkūnienė³

et al. 2020

BJMC

View full text Add to dashboard Cite

show abstract

“…Closed containment: A Figure object is inside a solid closed object (container; Ground) such that it cannot come out of it even if the container is moved with it (cp. Davis et al, 2017), e.g. a jam is in a closed jar.…”

Section: Containment and Support In Rcc+fmentioning

confidence: 99%

Containment and support in Baltic Languages: Overview, experimental evidence, and an extended RCC as applied to Latvian and Lithuania

Žilinskaitė-Šinkūnienė¹,

Šķilters²,

Zariņa³

2019

BJMC

View full text Add to dashboard Cite

The paper deals with the spatial relations of containment and support in the Baltic languages in a geometric framework. The geometric framework in our research is based on Region Connection Calculus representing different combinations of two circles depending on primitive topological relation of connectedness and extended with the distance, size, orientation and partial occlusion primitives. The research summarises results of two experiments conducted with 105 Latvian and 106 Lithuanian native speakers, provides the interpretation of 8 support and 5 containment stimuli, and gives a preliminary overview of correlation between different independent variables and the granularity and length of the spatial descriptions. Our results indicate a great variation in interpreting the given stimuli in each language, but the contrastive empirical analysis reveals that heterogeneity of interpretation in both languages is similar. For perception of support and containment relations, axial information and connectedness are the determining primitives.

show abstract

“…In support of this theory, Caramazza et al (1981) found that biases are attenuated by formal instruction which suggests that this cognitive process can be easily changed by verbal information. To summarize, the "intuitive physics engine" requires rich information about the world but can provide more precise, quantitative information, whereas rule-based systems can handle less specified world models, but cannot provide as precise outcomes (Davis, Marcus, & Frazier-Logue, 2017).…”

Section: Cognitive Systems For Physical Reasoningmentioning

confidence: 99%

Different physical intuitions exist between tasks, not domains

Battaglia¹,

Vul²

2018

Preprint

View full text Add to dashboard Cite

Does human behavior exploit deep and accurate knowledge about how the world works, or does it rely on shallow and often inaccurate heuristics? This fundamental question is rooted in a classic dichotomy in psychology: human intuitions about even simple scenarios can be poor, yet their behaviors can exceed the capabilities of even the most advanced machines. One domain where such a dichotomy has classically been demonstrated is intuitive physics. Here we demonstrate that this dichotomy is rooted in how physical knowledge is measured: extrapolation of ballistic motion is idiosyncratic and erroneous when people draw the trajectories, but consistent with accurate physical inferences under uncertainty when people use the same trajectories to catch a ball or release it to hit a target. Our results suggest that the contrast between rich and calibrated, versus poor and inaccurate patterns of physical reasoning exist as a result of using different systems of knowledge across tasks, rather than as a universal system of knowledge that is inconsistent across physical principles.

show abstract

Commonsense reasoning about containers using radically incomplete information

Cited by 25 publications

References 10 publications

Topologic and Geometric Structure of Spatial Relations in Latvian: an Experimental Analysis of RCC

Topologic and Geometric Structure of Spatial Relations in Latvian: an Experimental Analysis of RCC

Containment and support in Baltic Languages: Overview, experimental evidence, and an extended RCC as applied to Latvian and Lithuania

Different physical intuitions exist between tasks, not domains

Contact Info

Product

Resources

About