Northeastern University


A Monte Carlo Approach to Skill-Based Automated Playtesting

AAAI Conferences

In order to create well-crafted learning progressions, designers guide players as they present game skills and give ample time for the player to master those skills. However, analyzing the quality of learning progressions is challenging, especially during the design phase, as content is ever-changing. This research presents the application of Stratabots — automated player simulations based on models of players with varying sets of skills — to the human computation game Foldit. Stratabot performance analysis coupled with player data reveals a relatively smooth learning progression within tutorial levels, yet still shows evidence for improvement. Leveraging existing general gameplaying algorithms such as Monte Carlo Evaluation can reduce the development time of this approach to automated playtesting without losing predicitive power of the player model.


Exploratory Automated Analysis of Structural Features of Interactive Narrative

AAAI Conferences

Analysis of interactive narrative is a complex undertaking, requiring understanding of the narrative's design, its affordances, and its impact on players. Analysis is often performed by an expert, but this is expensive and difficult for complex interactive narratives. Automated analysis of structure, the organization of interaction elements, could help augment an expert's analysis. For this purpose we developed a model consisting of a set of metrics to analyze interactive narrative structure, enabled by a novel multi-graph representation. We implemented this model for an interactive scenario authoring tool called StudyCrafter and analyzed 20 student-designed scenarios. We show that the model illuminates the structures and groupings of the scenarios. This work provides insight for manual analysis of attributes of interactive narratives and a starting point for automated design assistance.


Predicting News Coverage of Scientific Articles

AAAI Conferences

Journalists act as gatekeepers to the scientific world, controlling what information reaches the public eye and how it is presented. Analyzing the kinds of research that typically receive more media attention is vital to understanding issues such as the “science of science communication” (National Academies of Sciences, Engineering, and Medicine 2017), patterns of misinformation, and the “cycle of hype.” We track the coverage of 91,997 scientific articles published in 2016 across various disciplines, publishers, and news outlets using metadata and text data from a leading tracker of scientific coverage in social and traditional media, Altmetric. We approach the problem as one of ranking each day’s, or week’s, papers by their likely level of media attention, using the learning-to-rank model lambdaMART (Burges 2010). We find that ngram features from the title, abstract and press release significantly improve performance over the metadata features journal, publisher, and subjects.


High Rank Matrix Completion With Side Information

AAAI Conferences

We address the problem of high-rank matrix completion with side information. In contrast to existing work dealing with side information, which assume that the data matrix is low-rank, we consider the more general scenario where the columns of the data matrix are drawn from a union of low-dimensional subspaces, which can lead to a high rank matrix. Our goal is to complete the matrix while taking advantage of the side information. To do so, we use the self-expressive property of the data, searching for a sparse representation of each column of matrix as a combination of a few other columns. More specifically, we propose a factorization of the data matrix as the product of side information matrices with an unknown interaction matrix, under which each column of the data matrix can be reconstructed using a sparse combination of other columns. As our proposed optimization, searching for missing entries and sparse coefficients, is non-convex and NP-hard, we propose a lifting framework, where we couple sparse coefficients and missing values and define an equivalent optimization that is amenable to convex relaxation. We also propose a fast implementation of our convex framework using a Linearized Alternating Direction Method. By extensive experiments on both synthetic and real data, and, in particular, by studying the problem of multi-label learning, we demonstrate that our method outperforms existing techniques in both low-rank and high-rank data regimes.


Towards Ultra-High Performance and Energy Efficiency of Deep Learning Systems: An Algorithm-Hardware Co-Optimization Framework

AAAI Conferences

Hardware accelerations of deep learning systems have been extensively investigated in industry and academia. The aim of this paper is to achieve ultra-high energy efficiency and performance for hardware implementations of deep neural networks (DNNs). An algorithm-hardware co-optimization framework is developed, which is applicable to different DNN types, sizes, and application scenarios. The algorithm part adopts the general block-circulant matrices to achieve a fine-grained tradeoff of accuracy and compression ratio. It applies to both fully-connected and convolutional layers and contains a mathematically rigorous proof of the effectiveness of the method. The proposed algorithm reduces computational complexity per layer from O(n 2 ) to O(n log n) and storage complexity from O(n 2 ) to O(n), both for training and inference. The hardware part consists of highly efficient Field Programmable Gate Array (FPGA)-based implementations using effective reconfiguration, batch processing, deep pipelining, resource re-using, and hierarchical control. Experimental results demonstrate that the proposed framework achieves at least 152X speedup and 71X energy efficiency gain compared with IBM TrueNorth processor under the same test accuracy. It achieves at least 31X energy efficiency gain compared with the reference FPGA-based work.


Action Prediction From Videos via Memorizing Hard-to-Predict Samples

AAAI Conferences

Action prediction based on video is an important problem in computer vision field with many applications, such as preventing accidents and criminal activities. It's challenging to predict actions at the early stage because of the large variations between early observed videos and complete ones. Besides, intra-class variations cause confusions to the predictors as well. In this paper, we propose a mem-LSTM model to predict actions in the early stage, in which a memory module is introduced to record several "hard-to-predict" samples and a variety of early observations. Our method uses Convolution Neural Network (CNN) and Long Short-Term Memory (LSTM) to model partial observed video input. We augment LSTM with a memory module to remember challenging video instances. With the memory module, our mem-LSTM model not only achieves impressive performance in the early stage but also makes predictions without the prior knowledge of observation ratio. Information in future frames is also utilized using a bi-directional layer of LSTM. Experiments on UCF-101 and Sports-1M datasets show that our method outperforms state-of-the-art methods.


An Interpretable Joint Graphical Model for Fact-Checking From Crowds

AAAI Conferences

Assessing the veracity of claims made on the Internet is an important, challenging, and timely problem. While automated fact-checking models have potential to help people better assess what they read, we argue such models must be explainable, accurate, and fast to be useful in practice; while prediction accuracy is clearly important, model transparency is critical in order for users to trust the system and integrate their own knowledge with model predictions. To achieve this, we propose a novel probabilistic graphical model (PGM) which combines machine learning with crowd annotations. Nodes in our model correspond to claim veracity, article stance regarding claims, reputation of news sources, and annotator reliabilities. We introduce a fast variational method for parameter estimation. Evaluation across two real-world datasets and three scenarios shows that: (1) joint modeling of sources, claims and crowd annotators in a PGM improves the predictive performance and interpretability for predicting claim veracity; and (2) our variational inference method achieves scalably fast parameter estimation, with only modest degradation in performance compared to Gibbs sampling. Regarding model transparency, we designed and deployed a prototype fact-checker Web tool, including a visual interface for explaining model predictions. Results of a small user study indicate that model explanations improve user satisfaction and trust in model predictions. We share our web demo, model source code, and the 13K crowd labels we collected.


Model AI Assignments 2018

AAAI Conferences

The Model AI Assignments session seeks to gather and disseminate the best assignment designs of the Artificial Intelligence (AI) Education community. Recognizing that assignments form the core of student learning ex- perience, we here present abstracts of seven AI assign- ments from the 2018 session that are easily adoptable, playfully engaging, and flexible for a variety of instruc- tor needs.


VSE-ens: Visual-Semantic Embeddings with Efficient Negative Sampling

AAAI Conferences

Jointing visual-semantic embeddings (VSE) have become a research hotpot for the task of image annotation, which suffers from the issue of semantic gap, i.e., the gap between images' visual features (low-level) and labels' semantic features (high-level). This issue will be even more challenging if visual features cannot be retrieved from images, that is, when images are only denoted by numerical IDs as given in some real datasets. The typical way of existing VSE methods is to perform a uniform sampling method for negative examples that violate the ranking order against positive examples, which requires a time-consuming search in the whole label space. In this paper, we propose a fast adaptive negative sampler that can work well in the settings of no figure pixels available. Our sampling strategy is to choose the negative examples that are most likely to meet the requirements of violation according to the latent factors of images. In this way, our approach can linearly scale up to large datasets. The experiments demonstrate that our approach converges 5.02x faster than the state-of-the-art approaches on OpenImages, 2.5x on IAPR-TCI2 and 2.06x on NUS-WIDE datasets, as well as better ranking accuracy across datasets.


Discriminative Semi-Coupled Projective Dictionary Learning for Low-Resolution Person Re-Identification

AAAI Conferences

Person re-identification (re-ID) is a fundamental task in automated video surveillance. In real-world visual surveillance systems, a person is often captured in quite low resolutions. So we often need to perform low-resolution person re-ID, where images captured by different cameras have great resolution divergences. Existing methods cope problem via some complicated and time-consuming strategies, making them less favorable in practice, and their performances are far from satisfactory. In this paper, we design a novel Discriminative Semi-coupled Projective Dictionary Learning (DSPDL) model to effectively and efficiently solve this problem. Specifically, we propose to jointly learn a pair of dictionaries and a mapping to bridge the gap across low(er) and high(er) resolution person images. Besides, we develop a novel graph regularizer to incorporate positive and negative image pair information in a parameterless fashion. Meanwhile, we adopt the efficient and powerful projective dictionary learning technique to boost the our efficiency. Experiments on three public datasets show the superiority of the proposed method to the state-of-the-art ones.