The field of learning analytics needs to adopt a more rigorous approach for predictive model evaluation that matches the complex practice of model-building. In this work, we present a procedure to statistically test hypotheses about model performance which goes beyond the state-of-the-practice in the community to analyze both algorithms and feature extraction methods from raw data. We apply this method to a series of algorithms and feature sets derived from a large sample of Massive Open Online Courses (MOOCs). While a complete comparison of all potential modeling approaches is beyond the scope of this paper, we show that this approach reveals a large gap in dropout prediction performance between forum-, assignment-, and clickstream-based feature extraction methods, where the latter is significantly better than the former two, which are in turn indistinguishable from one another. This work has methodological implications for evaluating predictive or AI-based models of student success, and practical implications for the design and targeting of at-risk student models and interventions.

Interactive training environments typically include feedback mechanisms designed to help trainees improve their performance through either guided or self-reflection. In this context, trainees are candidate teachers who need to hone their social skills as well as other pedagogical skills for their future classroom. We chose an avatar-mediated interactive virtual training system–TeachLivE–as the basic research environment to investigate the motions and embodiment of the trainees. Using tracking sensors, and customized improvements for existing gesture recognition utilities, we created a gesture database and employed it for the implementation of our real-time gesture recognition and feedback application. We also investigated multiple methods of feedback provision, including visual and haptics. The results from the conducted user studies and user evaluation surveys indicate the positive impact of the proposed feedback applications and informed body language. In this paper, we describe the context in which the utilities have been developed, the importance of recognizing nonverbal communication in the teaching context, the means of providing automated feedback associated with nonverbal messaging, and the preliminary studies developed to inform the research.

Data analysis education plays an important role in accelerating the efficient use of data analysis technologies in various domains. Not only the knowledge of statistics and machine learning, but also practical skills of deploying machine learning and data analysis techniques, are required for conducting data analysis projects in the real world. Data analysis competitions, such as Kaggle, have been considered as an efficient system for learning such skills by addressing real data analysis problems. However, current data analysis competitions are not designed for educational purposes and it is not well studied how data analysis competition platforms should be designed for enhancing educational effectiveness. To answer this research question, we built, and subsequently operated an educational data analysis competition platform called University of Big Data for several years. In this paper, we present our approaches for supporting and motivating learners and the results of our case studies. We found that providing a tutorial article is beneficial for encouraging active participation of learners, and a leaderboard system allowing an unlimited number of submissions can motivate the efforts of learners. We further discuss future directions of educational data analysis competitions.

Due to the limited natural water resources and the increase in population, managing water consumption is becoming an increasingly important subject worldwide. In this paper, we present and compare different machine learning models that are able to predict water demand for Central Indiana. The models are developed for two different time scales: daily and monthly. The input features for the proposed model include weather conditions (temperature, rainfall, snow), social features (holiday, median income), date (day of the year, month), and operational features (number of customers, previous water demand levels). The importance of these input features as accurate predictors is investigated. The results show that daily and monthly models based on recurrent neural networks produced the best results with an average error in prediction of 1.69% and 2.29%, respectively for 2016. These models achieve a high accuracy with a limited set of input features.

Sketching can be a valuable tool for science education, but it is currently underutilized. Sketch worksheets were developed to help change this, by using AI technology to give students immediate feedback and to give instructors assistance in grading. Sketch worksheets use visual representations automatically computed by CogSketch, which are combined with conceptual information from the OpenCyc ontology. Feedback is provided to students by comparing an instructor’s sketch to a student’s sketch, using the Structure-Mapping Engine. This paper describes our experiences in deploying sketch worksheets in two types of classes: Geoscience and AI. Sketch worksheets for introductory geoscience classes were developed by geoscientists at University of Wisconsin-Madison, authored using CogSketch and used in classes at both Wisconsin and Northwestern University. Sketch worksheets were also developed and deployed for a knowledge representation and reasoning course at Northwestern. Our experience indicates that sketch worksheets can provide helpful on-the-spot feedback to students, and significantly improve grading efficiency, to the point where sketching assignments can be more practical to use broadly in STEM education.

Driven by successes in deep learning, computer vision research has begun to move beyond object detection and image classification to more sophisticated tasks like image captioning or visual question answering. Motivating such endeavors is the desire for models to capture not only objects present in an image, but more fine-grained aspects of a scene such as relationships between objects and their attributes. Scene graphs provide a formal construct for capturing these aspects of an image. Despite this, there have been only a few recent efforts to generate scene graphs from imagery. Previous works limit themselves to settings where bounding box information is available at train time and do not attempt to generate scene graphs with attributes. In this paper we propose a method, based on recent advancements in Generative Adversarial Networks, to overcome these deficiencies. We take the approach of first generating small subgraphs, each describing a single statement about a scene from a specific region of the input image chosen using an attention mechanism. By doing so, our method is able to produce portions of the scene graphs with attribute information without the need for bounding box labels. Then, the complete scene graph is constructed from these subgraphs. We show that our model improves upon prior work in scene graph generation on state-of-the-art data sets and accepted metrics. Further, we demonstrate that our model is capable of handling a larger vocabulary size than prior work has attempted.

Continuous Sign Language Translation (SLT) is a challenging task due to its specific linguistics under sequential gesture variation without word alignment. Current hybrid HMM and CTC (Connectionist temporal classification) based models are proposed to solve frame or word level alignment. They may fail to tackle the cases with messing word order corresponding to visual content in sentences. To solve the issue, this paper proposes a hierarchical-LSTM (HLSTM) encoder-decoder model with visual content and word embedding for SLT. It tackles different granularities by conveying spatio-temporal transitions among frames, clips and viseme units. It firstly explores spatio-temporal cues of video clips by 3D CNN and packs appropriate visemes by online key clip mining with adaptive variable-length. After pooling on recurrent outputs of the top layer of HLSTM, a temporal attention-aware weighting mechanism is proposed to balance the intrinsic relationship among viseme source positions. At last, another two LSTM layers are used to separately recurse viseme vectors and translate semantic. After preserving original visual content by 3D CNN and the top layer of HLSTM, it shortens the encoding time step of the bottom two LSTM layers with less computational complexity while attaining more nonlinearity. Our proposed model exhibits promising performance on singer-independent test with seen sentences and also outperforms the comparison algorithms on unseen sentences.

There has been substantial work in recent years on grounded language acquisition, in which language and sensor data are used to create a model relating linguistic constructs to the perceivable world. While powerful, this approach is frequently hindered by ambiguities, redundancies, and omissions found in natural language. We describe an unsupervised system that learns language by training visual classifiers, first selecting important terms from object descriptions, then automatically choosing negative examples from a paired corpus of perceptual and linguistic data. We evaluate the effectiveness of each stage as well as the system's performance on the overall learning task.

In this paper, we focus on multiple-choice reading comprehension which aims to answer a question given a passage and multiple candidate options. We present the hierarchical attention flow to adequately leverage candidate options to model the interactions among passages, questions and candidate options. We observe that leveraging candidate options to boost evidence gathering from the passages play a vital role in this task, which is ignored in previous works. In addition, we explicitly model the option correlations with attention mechanism to obtain better option representations, which are further fed into a bilinear layer to obtain the ranking score for each option. On a large-scale multiple-choice reading comprehension dataset (i.e. the RACE dataset), the proposed model outperforms two previous neural network baselines on both RACE-M and RACE-H subsets and yields the state-of-the-art overall results.

Deep learning has demonstrated tremendous potential for Automatic Text Scoring (ATS) tasks. In this paper, we describe a new neural architecture that enhances vanilla neural network models with auxiliary neural coherence features. Our new method proposes a new SkipFlow mechanism that models relationships between snapshots of the hidden representations of a long short-term memory (LSTM) network as it reads. Subsequently, the semantic relationships between multiple snapshots are used as auxiliary features for prediction. This has two main benefits. Firstly, essays are typically long sequences and therefore the memorization capability of the LSTM network may be insufficient. Implicit access to multiple snapshots can alleviate this problem by acting as a protection against vanishing gradients. The parameters of the SkipFlow mechanism also acts as an auxiliary memory. Secondly, modeling relationships between multiple positions allows our model to learn features that represent and approximate textual coherence. In our model, we call this neural coherence features. Overall, we present a unified deep learning architecture that generates neural coherence features as it reads in an end-to-end fashion. Our approach demonstrates state-of-the-art performance on the benchmark ASAP dataset, outperforming not only feature engineering baselines but also other deep learning models.