No abstract

In helping students learn engineering design, it is very important that they explore complex scenarios that are realistic, and fall outside the domain of standard and over-simplified textbook problems that typically have an answer. A majority of the current educational methods and computer-based tools do not bridge this gap and lack affordances for design exploration. Although computational methods such as Finite Element Analysis have this potential, they are hard to use requiring the users to spend a significant effort. Also, several instructors have identified significant knowledge gaps in concepts related to structural design and strength of materials when the students reach their senior year. To this end, we have developed a problem-based framework to allow for rapid design exploration within engineering design curricula using an easy-to-use, simplified and constrained version of finite elements for stress analysis and exploration. Our framework makes it possible for users to rapidly explore various design options by incorporating a Finite Element Analysis (FEA) backend for design exploration. Our approach uses a constrained design problem for weight minimization that incorporates elements of structural topology optimization but does not automate it. Instead we provide the user the control on decision making for changing the shape through material removal. Using this framework, we explore the decision making of users, and their methodology in the course of the activities that provide a context of control, challenge and reflection. Using video and verbal protocol analysis we integrate assessment in ways that are important and interesting for learning. Our framework demonstrates that the ability of computational tools that are transformed for learning purposes can scaffold and augment learning processes in new ways.

No abstract

Gesture typing is a method of typing words on a touch-based keyboard by creating a continuous trace passing through the relevant keys. This work is aimed at developing a keyboard that supports gesture typing in Indic languages. We begin by noting that when dealing with Indic languages, one needs to cater to two different sets of users: (i) users who prefer to type in the native Indic script (Devanagari, Bengali, etc.) and (ii) users who prefer to type in the English script but want the transliterated output in the native script. In both cases, we need a model that takes a trace as input and maps it to the intended word. To enable the development of these models, we create and release two datasets. First, we create a dataset containing keyboard traces for 193,658 words from 7 Indic languages. Second, we curate 104,412 English-Indic transliteration pairs from Wikidata across these languages. Using these datasets we build a model that performs path decoding, transliteration and transliteration correction. Unlike prior approaches, our proposed model does not make co-character independence assumptions during decoding. The overall accuracy of our model across the 7 languages varies from 70-95%.

Gesture typing is a method of typing words on a touch-based keyboard by creating a continuous trace passing through the relevant keys. This work is aimed at developing a keyboard that supports gesture typing in Indic languages. We begin by noting that when dealing with Indic languages, one needs to cater to two different sets of users: (i) users who prefer to type in the native Indic script (Devanagari, Bengali, etc.) and (ii) users who prefer to type in the English script but want the transliterated output in the native script. In both cases, we need a model that takes a trace as input and maps it to the intended word. To enable the development of these models, we create and release two datasets. First, we create a dataset containing keyboard traces for 193,658 words from 7 Indic languages. Second, we curate 104,412 English-Indic transliteration pairs from Wikidata across these languages. Using these datasets we build a model that performs path decoding, transliteration and transliteration correction. Unlike prior approaches, our proposed model does not make co-character independence assumptions during decoding. The overall accuracy of our model across the 7 languages varies from 70-95%.