Correspondence-based analogies for choosing problem representations

Stockdill, Aaron; Raggi, Daniel; Jamnik, Mateja; Garcia, Grecia Garcia; Sutherland, Holly; Cheng, Peter; Sarkar, Advait

doi:10.1109/vl/hcc50065.2020.9127258

Cited by 7 publications

(8 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Full knowledge about the expressiveness of RSs and the problems solvable using their laws and tactics is rarely availableespecially when some systems have not been fully formalised, but are merely described. Thus, rep2rep exploits RS-and Q-descriptions for identifying relationships, using the novel notion of correspondences [19,25] between RSs, in the context of the given problem to be solved.…”

Section: A Informational Suitabilitymentioning

confidence: 99%

“…A correspondence between two RS-descriptions, r and r i , is a triple, α, β, s , where α and β are declarations stemming from RS-descriptions r and r i , respectively, and s ∈ [0, 1]. Generally, strengths may be informed by theoretical or empirical findings; for a probabilistic computation method, its statistical interpretation and some of its provable consequences, such as reversibility, composability, and extendability, see [25].…”

Section: A Informational Suitabilitymentioning

confidence: 99%

See 1 more Smart Citation

How to (Re)represent it?

Raggi

Stockdill

Jamnik

et al. 2020

2020 IEEE 32nd International Conference on Tools With Artificial Intelligence (ICTAI)

Self Cite

View full text Add to dashboard Cite

Choosing an effective representation is fundamental to the ability of the representation's user to exploit it for the intended purpose. The major contribution of this paper is to provide a novel, flexible framework, rep2rep, that can be used by AI systems to recommend effective representations. What makes an effective representation is determined by whether it expresses the necessary information, supports the execution of tasks, and reflects the user's cognitive abilities. In general, there is no single 'most effective' representation for every problem and every user, which makes it difficult to choose one from the plethora of possible representations. To address this, rep2rep includes: a domain-independent language for describing representations, algorithms that compute measures of informational suitability and overall cognitive cost, and uses these measures to recommend representations. We demonstrate the application of rep2rep in the probability domain. Importantly, our framework provides the foundations for personalised interaction with AI systems in the context of representation choice.

show abstract

Section: A Informational Suitabilitymentioning

confidence: 99%

Section: A Informational Suitabilitymentioning

confidence: 99%

How to (Re)represent it?

Raggi

Stockdill

Jamnik

et al. 2020

2020 IEEE 32nd International Conference on Tools With Artificial Intelligence (ICTAI)

Self Cite

View full text Add to dashboard Cite

show abstract

“…We can use RS and Q descriptions to compute important measures: informational suitability (presented in [9]), and cognitive cost. The Informational Suitability (IS) of an RS, r, given a problem q is the sum of the strengths of analogical correspondences [11] between components that match the source q and the target r, modulated by the importance of said components: The Cognitive Cost encodes the RS's processing cost to the user, and is calculated by computing a set of properties of the representation, all of which can be estimated by values computed from Q descriptions (out of the scope of this paper). These properties are based on established cognitive science concepts [1, 6-8, 12, 13], presented schematically in Figure 3.…”

Section: Evaluating Representationsmentioning

confidence: 99%

“…In previous work we introduced a language [9] for encoding the properties of representational systems, in addition to correspondences between them [11]. The purpose was to calculate an informational measure, which, given a problem, estimates the likelihood that the important information for this problem can be expressed in any given representational system.…”

Section: Introductionmentioning

confidence: 99%

Dissecting Representations

Raggi¹,

Stockdill²,

Jamnik³

et al. 2020

Diagrammatic Representation and Inference

Self Cite

View full text Add to dashboard Cite

Choosing effective representations for a problem and for the person solving it has benefits, including the ability or inability to solve it. We previously devised a novel framework consisting of a language to describe representations and computational methods to analyse them in terms of their formal and cognitive properties. In this paper we demonstrate the application of this framework to a variety of notations including natural languages, formal languages, and diagrams. We show how our framework, and the analysis of representations that it enables, gives us insight into how and why we can select representations which are appropriate for both the task and the user.

show abstract

“…For his part, Plappally [9] designed a model of analogies using concept maps for the first year of mechanical engineering courses. Additionally, Stockdill et al [10] implemented a correspondence-based analogy model to choose representations of mathematical problems, among others.…”

Section: Introductionmentioning

confidence: 99%

Discovery Model Based on Analogies for Teaching Computer Programming

2021

View full text Add to dashboard Cite

Teaching the fundamentals of computer programming in a first course (CS1) is a complex activity for the professor and is also a challenge for them. Nowadays, there are several teaching strategies for dealing with a CS1 at the university, one of which is the use of analogies to support the abstraction process that a student needs to carry for the appropriation of fundamental concepts. This article presents the results of applying a discovery model that allowed for the extraction of patterns, linguistic analysis, textual analytics, and linked data when using analogies for teaching the fundamental concepts of programming by professors in a CS1 in university programs that train software developers. For that reason, a discovery model based on machine learning and text mining was proposed using natural language processing techniques for semantic vector space modeling, distributional semantics, and the generation of synthetic data. The discovery process was carried out using nine supervised learning methods, three unsupervised learning methods, and one semi-supervised learning method involving linguistic analysis techniques, text analytics, and linked data. The main findings showed that professors include keywords, which are part of the technical computer terminology, in the form of verbs in the statement of the analogy and combine them in quantitative contexts with neutral or positive phrases, where numerical examples, cooking recipes, and games were the most used categories. Finally, a structure is proposed for the construction of analogies to teach programming concepts and this was validated by the professors and students.

show abstract

Correspondence-based analogies for choosing problem representations

Cited by 7 publications

References 19 publications

How to (Re)represent it?

How to (Re)represent it?

Dissecting Representations

Discovery Model Based on Analogies for Teaching Computer Programming

Contact Info

Product

Resources

About