Papers.

A Contextual Discretization framework for compressing Recurrent Neural Networks

OSoMe: the IUNI observatory on social media

Mango: data exploration on large genomic datasets

Distributed visualization for genomic analysis

Reducing card-not-present fraud using pre-approved transactions

CASSAurus: A Resource of Simpler Spanish Synonyms

Building Language Resources for Exploring Autism Spectrum Disorders

Deeplearning4j: Distributed, open-source deep learning for Java and Scala on Hadoop and Spark

Deeplearning4j implements deep artificial neural networks on the JVM, with a Java API and C++ for computation. Deeplearning4j includes recurrent neural networks such as long short-term memory networks, as well as convolutional networks, restricted boltzmann machines, deep-belief networks, deep autoencoders and shallow neural networks such as word2vec. Deeplearning4j integrates with Hadoop and Spark, and is built to train neural networks in parallel on CPUs and GPUs.

Session details: Session 22

Refinement Driven Processing of Aggregation Constrained Queries

metricsgraphics: metricsgraphics htmlwidget package version 0.9.1

An Architecture-Centric Process for MILS Development

Safety-critical embedded systems are now software-reliant and evolving at an incredible pace. With the emerging Internet of Things (IoT) ecosystem, these systems are now interconnected to several networks and exposed to potential attackers. This increases the potential surface of attack and, ultimately, the likelihood of a successful attack that would penetrate the system. Until recently, many security efforts were focused on code analysis, but studies have shown that security is also a matter of good software architecture design and practices. For example, MILS requires isolating security domains in partitions using appropriate security components. However, because embedded systems are evolving quickly, new design methods are now required to overcome the challenges of developing them. In this paper, we introduce a research agenda for a new architecturecentric development approach forMILS systems. This would leverage architecture models and augment them with security information in order to perform the different activities of the development process, including security policy validation, implementation, and testing. Using the same model throughout development improves the consistency of the development process by avoiding any translation between different—and potentially inconsistent—representations. In addition, automating the generation of implementation and tests avoids the traditional mistakes of manual code production, such as bugs and developers’ assumptions about ambiguous requirements.

Availability Considerations for Mission Critical Applications in the Cloud

Cloud environments offer flexibility, elasticity, and low cost compute infrastructure. Enterprise-level workloads â?? such as SAP and Oracle workloads - require infrastructure with high availability, clustering, or physical server appliances. These features are often not part of a typical cloud offering, and as a result, businesses are forced to run enterprise workloads in their legacy environments. To enable enterprise customers to use these workloads in a cloud, we enabled a large number of SAP and Oracle workloads in the IBM Cloud Managed Services (CMS) for both virtualized and non-virtualized cloud environments. In this paper, we discuss the challenges in enabling enterprise class applications in the cloud based on our experience on providing a diverse set of platforms implemented in the IBM CMS offering.

A real-time dashboard for managing pathology processes

Context: The Eastern Ontario Regional Laboratory Association (EORLA) is a newly established association of all the laboratory and pathology departments of Eastern Ontario that currently includes facilities from eight hospitals. All surgical specimens for EORLA are processed in one central location, the Department of Pathology and Laboratory Medicine (DPLM) at The Ottawa Hospital (TOH), where the rapid growth and influx of surgical and cytology specimens has created many challenges in ensuring the timely processing of cases and reports. Although the entire process is maintained and tracked in a clinical information system, this system lacks pre-emptive warnings that can help management address issues as they arise. Aims: Dashboard technology provides automated, real-time visual clues that could be used to alert management when a case or specimen is not being processed within predefined time frames. We describe the development of a dashboard helping pathology clinical management to make informed decisions on specimen allocation and tracking. Methods: The dashboard was designed and developed in two phases, following a prototyping approach. The first prototype of the dashboard helped monitor and manage pathology processes at the DPLM. Results: The use of this dashboard helped to uncover operational inefficiencies and contributed to an improvement of turn-around time within The Ottawa Hospital′s DPML. It also allowed the discovery of additional requirements, leading to a second prototype that provides finer-grained, real-time information about individual cases and specimens. Conclusion: We successfully developed a dashboard that enables managers to address delays and bottlenecks in specimen allocation and tracking. This support ensures that pathology reports are provided within time frame standards required for high-quality patient care. Given the importance of rapid diagnostics for a number of diseases, the use of real-time dashboards within pathology departments could

Production and Perception of Tone 3 Focus in Mandarin Chinese

The Off-Switch Game

It is clear that one of the primary tools we can use to mitigate the potential risk from a misbehaving AI system is the ability to turn the system off. As the capabilities of AI systems improve, it is important to ensure that such systems do not adopt subgoals that prevent a human from switching them off. This is a challenge because many formulations of rational agents create strong incentives for self-preservation. This is not caused by a built-in instinct, but because a rational agent will maximize expected utility and cannot achieve whatever objective it has been given if it is dead. Our goal is to study the incentives an agent has to allow itself to be switched off. We analyze a simple game between a human H and a robot R, where H can press R's off switch but R can disable the off switch. A traditional agent takes its reward function for granted: we show that such agents have an incentive to disable the off switch, except in the special case where H is perfectly rational. Our key insight is that for R to want to preserve its off switch, it needs to be uncertain about the utility associated with the outcome, and to treat H's actions as important observations about that utility. (R also has no incentive to switch itself off in this setting.) We conclude that giving machines an appropriate level of uncertainty about their objectives leads to safer designs, and we argue that this setting is a useful generalization of the classical AI paradigm of rational agents.

Psychophysical Study of Color Verbalization Using Fuzzy Logic

We develop and implement a fuzzy logic algorithm able to determine which areas of the CIE-Lab color space correspond, with different weights, to which conventional color names (red, green, blue, yellow, magenta, purple, brown, etc.)Parameters of the underlying model are estimated using data derived from a psychophysical experiment. The resulting color naming application will benefit developers of computer vision and image classification applications, graphic designers, imaging hardware and software manufacturers, and other companies and individuals, including those involved in publishing, advertisement, and commercial photography.

Approximate Subgraph Isomorphism for Image Localization

Pauses and pause fillers in Mandarin monologue speech: The effects of sex and proficiency

Voice quality as a pitch-range indicator

The Rhythmic Constraint on Prosodic Boundaries in Mandarin Chinese Based on Corpora of Silent Reading and Speech Perception

Phoneme, Phone Boundary, and Tone in Automatic Scoring of Mandarin Proficiency

Pitch-Range Perception: The Dynamic Interaction Between Voice Quality and Fundamental Frequency

Automatic Analysis of Phonetic Speech Style Dimensions

Data Privacy Issues in Cloud Computing

Towards a continuous modeling of natural language domains

Exploring Autism Spectrum Disorders Using HLT

INSIGHT-1 at SemEval-2016 Task 5: Deep Learning for Multilingual Aspect-based Sentiment Analysis

INSIGHT-1 at SemEval-2016 Task 4: Convolutional Neural Networks for Sentiment Classification and Quantification

Probabilistic Models for Learning a Semantic Parser Lexicon

A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks

Transfer and multi-task learning have traditionally focused on either a single source-target pair or very few, similar tasks. Ideally, the linguistic levels of morphology, syntax and semantics would benefit each other by being trained in a single model. We introduce a joint many-task model together with a strategy for successively growing its depth to solve increasingly complex tasks. Higher layers include shortcut connections to lower-level task predictions to reflect linguistic hierarchies. We use a simple regularization term to allow for optimizing all model weights to improve one task’s loss without exhibiting catastrophic interference of the other tasks. Our single end-to-end model obtains state-of-the-art or competitive results on five different tasks from tagging, parsing, relatedness, and entailment tasks.

A Hierarchical Model of Reviews for Aspect-based Sentiment Analysis

Semantic Parsing to Probabilistic Programs for Situated Question Answering

Distinguishing Past, On-going, and Future Events: The EventStatus Corpus

Determining whether a major societal event has already happened, is still on-going, or may occur in the future is crucial for event prediction, timeline generation, and news summarization. We introduce a new task and a new corpus, EventStatus , which has 4500 English and Spanish articles about civil unrest events labeled as P AST , O N -G OING , or F U - TURE . We show that the temporal status of these events is difficult to classify because lo-cal tense and aspect cues are often lacking, time expressions are insufficient, and the linguistic contexts have rich semantic compositionality. We explore two approaches for event status classification: (1) a feature-based SVM classifier augmented with a novel induced lexicon of future-oriented verbs, such as “threat-ened” and “planned”, and (2) a convolutional neural net. Both types of classifiers improve event status recognition over a state-of-the-art TempEval model, and our analysis offers linguistic insights into the semantic composition-ality challenges for this new task.

Medulloblastoma-associated DDX3 variant selectively alters the translational response to stress

A Survey of Current Practice and Teaching of AI

The field of AI has changed significantly in the past couple of years and will likely continue to do so. Driven by a desire to expose our students to relevant and modern materials, we conducted two surveys, one of AI instructors and one of AI practitioners. The surveys were aimed at gathering infor-mation about the current state of the art of introducing AI as well as gathering input from practitioners in the field on techniques used in practice. In this paper, we present and briefly discuss the responses to those two surveys.

Instructable Intelligent Personal Agent

SystemML: Declarative Machine Learning on Spark

Yelp Reviews Of Hospital Care Can Supplement And Inform Traditional Surveys Of The Patient Experience Of Care

Little is known about how real-time online rating platforms such as Yelp may complement the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey, which is the US standard for evaluating patients' experiences after hospitalization. We compared the content of Yelp narrative reviews of hospitals to the topics in the HCAHPS survey, called domains in HCAHPS terminology. While the domains included in Yelp reviews covered the majority of HCAHPS domains, Yelp reviews covered an additional twelve domains not found in HCAHPS. The majority of Yelp topics that most strongly correlate with positive or negative reviews are not measured or reported by HCAHPS. The large collection of patient- and caregiver-centered experiences found on Yelp can be analyzed with natural language processing methods, identifying for policy makers the measures of hospital quality that matter most to patients and caregivers. The Yelp measures and analysis can also provide actionable feedback for hospitals.

Direct Measurements of Smartphone Screen-Time: Relationships with Demographics and Sleep

Growing Homophilic Networks Are Natural Navigable Small Worlds

Navigability, an ability to find a logarithmically short path between elements using only local information, is one of the most fascinating properties of real-life networks. However, the exact mechanism responsible for the formation of navigation properties remained unknown. We show that navigability can be achieved by using only two ingredients present in the majority of networks: network growth and local homophily, giving a persuasive answer how the navigation appears in real-life networks. A very simple algorithm produces hierarchical self-similar optimally wired navigable small world networks with exponential degree distribution by using only local information. Adding preferential attachment produces a scale-free network which has shorter greedy paths, but worse (power law) scaling of the information extraction locality (algorithmic complexity of a search). Introducing saturation of the preferential attachment leads to truncated scale-free degree distribution that offers a good tradeoff between these parameters and can be useful for practical applications. Several features of the model are observed in real-life networks, in particular in the brain neural networks.

Short Time-Scale Sensory Coding in S1 during Discrimination of Whisker Vibrotactile Sequences

Measures how rat primary somatosensory cortex encodes whisker vibrotactile sequences on short time scales during a discrimination task, showing that spike-count coding within brief windows carries the information used for behavioural discrimination.

Distilling the Wisdom of Crowds: Prediction Markets vs. Prediction Polls

We report the results of the first large-scale, long-term, experimental test between two crowdsourcing methods: prediction markets and prediction polls. More than 2,400 participants made forecasts on 261 events over two seasons of a geopolitical prediction tournament. Forecasters were randomly assigned to either prediction markets (continuous double auction markets) in which they were ranked based on earnings, or prediction polls in which they submitted probability judgments, independently or in teams, and were ranked based on Brier scores. In both seasons of the tournament, prices from the prediction market were more accurate than the simple mean of forecasts from prediction polls. However, team prediction polls outperformed prediction markets when forecasts were statistically aggregated using temporal decay, differential weighting based on past performance, and recalibration. The biggest advantage of prediction polls was at the beginning of long-duration questions. Results suggest that prediction polls with proper scoring feedback, collaboration features, and statistical aggregation are an attractive alternative to prediction markets for distilling the wisdom of crowds. This paper was accepted by Uri Gneezy, behavioral economics.

Variation and Change in the Use of Hesitation Markers in Germanic Languages

A Latent Variable Model Approach to PMI-based Word Embeddings.

Quantization for Robust Distributed Coding

Enabling enterprise-level workloads in the enterprise-class cloud

Session details: Keynote I

Towards Visualization Recommendation Systems

Processing uncertain spatial data resulting from differentially-private sanitization