-
Creating A Neural Pedagogical Agent by Jointly Learning to Review and Assess
Authors:
Youngnam Lee,
Youngduck Choi,
Junghyun Cho,
Alexander R. Fabbri,
Hyunbin Loh,
Chanyou Hwang,
Yongku Lee,
Sang-Wook Kim,
Dragomir Radev
Abstract:
Machine learning plays an increasing role in intelligent tutoring systems as both the amount of data available and specialization among students grow. Nowadays, these systems are frequently deployed on mobile applications. Users on such mobile education platforms are dynamic, frequently being added, accessing the application with varying levels of focus, and changing while using the service. The e…
▽ More
Machine learning plays an increasing role in intelligent tutoring systems as both the amount of data available and specialization among students grow. Nowadays, these systems are frequently deployed on mobile applications. Users on such mobile education platforms are dynamic, frequently being added, accessing the application with varying levels of focus, and changing while using the service. The education material itself, on the other hand, is often static and is an exhaustible resource whose use in tasks such as problem recommendation must be optimized. The ability to update user models with respect to educational material in real-time is thus essential; however, existing approaches require time-consuming re-training of user features whenever new data is added. In this paper, we introduce a neural pedagogical agent for real-time user modeling in the task of predicting user response correctness, a central task for mobile education applications. Our model, inspired by work in natural language processing on sequence modeling and machine translation, updates user features in real-time via bidirectional recurrent neural networks with an attention mechanism over embedded question-response pairs. We experiment on the mobile education application SantaTOEIC, which has 559k users, 66M response data points as well as a set of 10k study problems each expert-annotated with topic tags and gathered since 2016. Our model outperforms existing approaches over several metrics in predicting user response correctness, notably out-performing other methods on new users without large question-response histories. Additionally, our attention mechanism and annotated tag set allow us to create an interpretable education platform, with a smart review system that addresses the aforementioned issue of varied user attention and problem exhaustion.
△ Less
Submitted 1 July, 2019; v1 submitted 26 June, 2019;
originally announced June 2019.
-
What Should I Learn First: Introducing LectureBank for NLP Education and Prerequisite Chain Learning
Authors:
Irene Li,
Alexander R. Fabbri,
Robert R. Tung,
Dragomir R. Radev
Abstract:
Recent years have witnessed the rising popularity of Natural Language Processing (NLP) and related fields such as Artificial Intelligence (AI) and Machine Learning (ML). Many online courses and resources are available even for those without a strong background in the field. Often the student is curious about a specific topic but does not quite know where to begin studying. To answer the question o…
▽ More
Recent years have witnessed the rising popularity of Natural Language Processing (NLP) and related fields such as Artificial Intelligence (AI) and Machine Learning (ML). Many online courses and resources are available even for those without a strong background in the field. Often the student is curious about a specific topic but does not quite know where to begin studying. To answer the question of "what should one learn first," we apply an embedding-based method to learn prerequisite relations for course concepts in the domain of NLP. We introduce LectureBank, a dataset containing 1,352 English lecture files collected from university courses which are each classified according to an existing taxonomy as well as 208 manually-labeled prerequisite relation topics, which is publicly available. The dataset will be useful for educational purposes such as lecture preparation and organization as well as applications such as reading list generation. Additionally, we experiment with neural graph-based networks and non-neural classifiers to learn these prerequisite relations from our dataset.
△ Less
Submitted 26 November, 2018;
originally announced November 2018.
-
A quantitative approach to evolution of music and philosophy
Authors:
Vilson Vieira,
Renato Fabbri,
Gonzalo Travieso,
Osvaldo N. Oliveira Jr.,
Luciano da Fontoura Costa
Abstract:
The development of new statistical and computational methods is increasingly making it possible to bridge the gap between hard sciences and humanities. In this study, we propose an approach based on a quantitative evaluation of attributes of objects in fields of humanities, from which concepts such as dialectics and opposition are formally defined mathematically. As case studies, we analyzed the t…
▽ More
The development of new statistical and computational methods is increasingly making it possible to bridge the gap between hard sciences and humanities. In this study, we propose an approach based on a quantitative evaluation of attributes of objects in fields of humanities, from which concepts such as dialectics and opposition are formally defined mathematically. As case studies, we analyzed the temporal evolution of classical music and philosophy by obtaining data for 8 features characterizing the corresponding fields for 7 well-known composers and philosophers, which were treated with multivariate statistics and pattern recognition methods. A bootstrap method was applied to avoid statistical bias caused by the small sample data set, with which hundreds of artificial composers and philosophers were generated, influenced by the 7 names originally chosen. Upon defining indices for opposition, skewness and counter-dialectics, we confirmed the intuitive analysis of historians in that classical music evolved according to a master-apprentice tradition, while in philosophy changes were driven by opposition. Though these case studies were meant only to show the possibility of treating phenomena in humanities quantitatively, including a quantitative measure of concepts such as dialectics and opposition the results are encouraging for further application of the approach presented here to many other areas, since it is entirely generic.
△ Less
Submitted 13 November, 2013;
originally announced March 2014.
-
A Quantitative Approach to Painting Styles
Authors:
Vilson Vieira,
Renato Fabbri,
David Sbrissa,
Luciano da Fontoura Costa,
Gonzalo Travieso
Abstract:
This research extends a method previously applied to music and philosophy,representing the evolution of art as a time-series where relations like dialectics are measured quantitatively. For that, a corpus of paintings of 12 well-known artists from baroque and modern art is analyzed. A set of 93 features is extracted and the features which most contributed to the classification of painters are sele…
▽ More
This research extends a method previously applied to music and philosophy,representing the evolution of art as a time-series where relations like dialectics are measured quantitatively. For that, a corpus of paintings of 12 well-known artists from baroque and modern art is analyzed. A set of 93 features is extracted and the features which most contributed to the classification of painters are selected. The projection space obtained provides the basis to the analysis of measurements. This quantitative measures underlie revealing observations about the evolution of painting styles, specially when compared with other humanity fields already analyzed: while music evolved along a master-apprentice tradition (high dialectics) and philosophy by opposition, painting presents another pattern: constant increasing skewness, low opposition between members of the same movement and opposition peaks in the transition between movements. Differences between baroque and modern movements are also observed in the projected "painting space": while baroque paintings are presented as an overlapped cluster, the modern paintings present minor overlap** and are disposed more widely in the projection than the baroque counterparts. This finding suggests that baroque painters shared aesthetics while modern painters tend to "break rules" and develop their own style.
△ Less
Submitted 13 November, 2013;
originally announced March 2014.
-
Can the evolution of music be analyzed in a quantitative manner?
Authors:
Vilson Vieira,
Renato Fabbri,
Gonzalo Travieso,
Luciano da Fontoura Costa
Abstract:
We propose a methodology to study music development by applying multivariate statistics on composers characteristics. Seven representative composers were considered in terms of eight main musical features. Grades were assigned to each characteristic and their correlations were analyzed. A bootstrap method was applied to simulate hundreds of artificial composers influenced by the seven representati…
▽ More
We propose a methodology to study music development by applying multivariate statistics on composers characteristics. Seven representative composers were considered in terms of eight main musical features. Grades were assigned to each characteristic and their correlations were analyzed. A bootstrap method was applied to simulate hundreds of artificial composers influenced by the seven representatives chosen. Afterwards we quantify non-numeric relations like dialectics, opposition and innovation. Composers differences on style and technique were represented as geometrical distances in the feature space, making it possible to quantify, for example, how much Bach and Stockhausen differ from other composers or how much Beethoven influenced Brahms. In addition, we compared the results with a prior investigation on philosophy. Opposition, strong on philosophy, was not remarkable on music. Supporting an observation already considered by music theorists, strong influences were identified between composers by the quantification of dialectics, implying inheritance and suggesting a stronger master-disciple evolution when compared to the philosophy analysis.
△ Less
Submitted 4 March, 2012; v1 submitted 21 September, 2011;
originally announced September 2011.
