The deep learning algorithms of artificial intelligence can identify patterns that help inventors think laterally, make connections between nonobvious ideas, pinpoint hidden invention features, and exploit new science and technology-based opportunities. "To invent, you need a good imagination and a pile of junk." So said Thomas Edison, America's most prolific inventor. Yet the march of technology is now changing the great man's inventive equation: powerful algorithmic advisory systems are now giving inventors far more fertile imaginations, even if they don't have very much of one themselves. After being fed vast datasets of information on a field of inventive endeavor, deep learning algorithms identify patterns that help inventors think laterally, make connections between nonobvious ideas, pinpoint hidden invention features that rivals have missed, and exploit new science and technology-based opportunities from, say, patents and journals.

Argument mining is a rising area of Natural Language Pro- cessing (NLP) concerned with the automatic recognition and interpretation of argument components and their relations. Neural models are by now mature technologies to be ex- ploited for automating the argument mining tasks, despite the issue of data sparseness. This could ease much of the man- ual effort involved in these tasks, taking into account hetero- geneous types of texts and topics. In this work, we evaluate different attention mechanisms applied over a state-of-the-art architecture for sequence labeling. We assess the impact of different flavors of attention in the task of argument compo- nent detection over two datasets: essays and legal domain. We show that attention not models the problem better but also supports interpretability.

Today, significant impacts of comments on the web affect people decisions while they are about to choose a product. Unfavorable effect of spam attacks in these reviews follows heavy damages for customers and organizations. The majority of methods so far classify reviews to spam and non-spam groups. Therefore, most researches are done on feature learning techniques to enhance the classification performance. From another point of view, presence of huge amount of features makes text classification overwhelming. Attention mechanism has lately been used to improve neural networks performance on sequence modeling. Instead of mining all existing features, attention can enables the model to concentrate on most important parts of the data. To these ends, we applied an attention based deep structure for detecting deceptive reviews. This model contributes distinguishing between truthful and fake reviews and benefits an attentional part to engineering better features. Our proposed model accuracy and scalability is comparable regard to the other common models.

Wireless traffic prediction is a fundamental enabler to proactive network optimisation in 5G and beyond. Forecasting extreme demand spikes and troughs is essential to avoiding outages and improving energy efficiency. However, current forecasting methods predominantly focus on overall forecast performance and/or do not offer probabilistic uncertainty quantification. Here, we design a feature embedding (FE) kernel for a Gaussian Process (GP) model to forecast traffic demand. The FE kernel enables us to trade-off overall forecast accuracy against peak-trough accuracy. Using real 4G base station data, we compare its performance against both conventional GPs, ARIMA models, as well as demonstrate the uncertainty quantification output. The advantage over neural network (e.g. CNN, LSTM) models is that the probabilistic forecast uncertainty can directly feed into decision processes in self-organizing-network (SON) modules.

With the rapid proliferation and increased sophistication of malicious software (malware), detection methods no longer rely only on manually generated signatures but have also incorporated more general approaches like Machine Learning (ML) detection. Although powerful for conviction of malicious artifacts, these methods do not produce any further information about the type of malware that has been detected. In this work, we address the information gap between ML and signature-based detection methods by introducing an ML-based tagging model that generates human interpretable semantic descriptions of malicious software (e.g. file-infector, coin-miner), and argue that for less prevalent malware campaigns these provide potentially more useful and flexible information than malware family names. For this, we first introduce a method for deriving high-level descriptions of malware files from an ensemble of vendor family names. Then we formalize the problem of malware description as a tagging problem and propose a joint embedding deep neural network architecture that can learn to characterize portable executable (PE) files based on static analysis, thus not requiring a dynamic trace to identify behaviors at deployment time. We empirically demonstrate that when evaluated against tags extracted from an ensemble of anti-virus detection names, the proposed tagging model correctly identifies more than 93.7% of eleven possible tag descriptions for a given sample, at a deployable false positive rate (FPR) of 1% per tag. Furthermore, we show that when evaluating this model against ground truth tags derived from the results of dynamic analysis, it correctly predicts 93.5% of the labels for a given sample. These results suggest that an ML tagging model can be effectively deployed alongside a detection model for malware description.

Recent technological advances have proliferated the available computing power, memory, and speed of modern Central Processing Units (CPUs), Graphics Processing Units (GPUs), and Field Programmable Gate Arrays (FPGAs). Consequently, the performance and complexity of Artificial Neural Networks (ANNs) is burgeoning. While GPU accelerated Deep Neural Networks (DNNs) currently offer state-of-the-art performance, they consume large amounts of power. Training such networks on CPUs is inefficient, as data throughput and parallel computation is limited. FPGAs are considered a suitable candidate for performance critical, low power systems, e.g. the Internet of Things (IOT) edge devices. Using the Xilinx SDAccel or Intel FPGA SDK for OpenCL development environment, networks described using the high-level OpenCL framework can be accelerated on heterogeneous platforms. Moreover, the resource utilization and power consumption of DNNs can be further enhanced by utilizing regularization techniques that binarize network weights. In this paper, we introduce, to the best of our knowledge, the first FPGA-accelerated stochastically binarized DNN implementations, and compare them to implementations accelerated using both GPUs and FPGAs. Our developed networks are trained and benchmarked using the popular MNIST and CIFAR-10 datasets, and achieve near state-of-the-art performance, while offering a >16-fold improvement in power consumption, compared to conventional GPU-accelerated networks. Both our FPGA-accelerated determinsitic and stochastic BNNs reduce inference times on MNIST and CIFAR-10 by >9.89x and >9.91x, respectively.

A simple, flexible approach to creating expressive priors in Gaussian process (GP) models makes new kernels from a combination of basic kernels, e.g. summing a periodic and linear kernel can capture seasonal variation with a long term trend. Despite a well-studied link between GPs and Bayesian neural networks (BNNs), the BNN analogue of this has not yet been explored. This paper derives BNN architectures mirroring such kernel combinations. Furthermore, it shows how BNNs can produce periodic kernels, which are often useful in this context. These ideas provide a principled approach to designing BNNs that incorporate prior knowledge about a function. We showcase the practical value of these ideas with illustrative experiments in supervised and reinforcement learning settings.

Many real world data such as social networks, collections of documents and chemical structures are naturally represented as graphs. Consequently there exists great potential for the application of machine learning to graphs. Given the great successes of neural networks or deep learning to the analysis of images, there has recently been much research considering the application or generalization of neural networks to graphs. In many cases this has resulted in state of the art performance in many tasks (Wu et al., 2019). Graph convolutional is a neural network architecture commonly applied to graphs. This architecture consists of a sequence of convolutional layers where each layer iteratively updates a representation or embedding of each vertex. This update is achieved through the application of an operation which considers the current representation of each vertex plus the current representation of its adjacent neighbours (Gilmer et al., 2017). The output of a sequence of convolutional layers is a representation of each vertex which encodes properties of the vertex in question and vertices in its neighbourhood. If one wishes to perform a vertex centric task such as vertex classification, then one may operate directly on the set of vertex representations output from a sequence of convolutional layers.

The FLAIRS poster track is designed to promote discussion of emerging ideas and work in order to encourage and help guide researchers — especially new researchers — who are able to present a full poster in the conference poster session and receive that critical work-shaping feedback that helps guide good work into great work. Abstracts of those posters appear here, which we hope to see fully developed into future FLAIRS papers..

This paper presents the first results of the research into AI-based support of the room configuration process during the early design phases in architecture. Room configuration (also: room layout or space layout) is an essential stage of the initial design phase: its results are crucial for user-friendliness and success of the planned utilization of the architectural object. Our approach takes into account different possible actions of the configuration process, such as adding, removing, or (re)assigning of the room type. Its mode of operation is based on specific process chain clusters, where each cluster represents a contextual subset of previous configuration steps and provides a recurrent neural network trained on this cluster data only to suggest the next step, and a case base that is used to determine if the current process chain belongs to this cluster. The most similar cluster then tries to suggest the next step of the process. The approach is implemented in a distributed CBR framework for support of early conceptual design in architecture and was evaluated with a high number of process chain queries to prove its general suitability.