ABSTRACT Time‐series forecasting is essential for predicting events in the future and for tracking objects. The conventional recurrent neural network model needs to pad the target with zeros when handling long inputs, resulting in a loss in accuracy. Recently, it was proposed to divide a time series input into patches and merge the learned weights. However, such a model is difficult to interpret. In this article, we consider a mixture of continuous and discrete Markov states to model long‐range time dependencies. For example, in a vehicle, each gear level can be a discrete state and the throttle input is continuously controlled to maximise the efficiency of the engine. Data collected from the sensor is prone to noise due to component faults or external disturbances. Hence, we apply a stability constraint to select samples for training. We validate our algorithm on three datasets: (1) Apple Watch, (2) Car engine and (3) Election tweets. On all datasets, we achieve an improvement in the range of 5%–20% in the F‐measure. Furthermore, the features learned are easy to explain in terms of real‐world scenarios.

Abstract The use of computer‐mediated communication has resulted in a new form of written text called Microtext, which is very different from well‐written text. Most previous approaches deal with microtext at the character level rather than just words resulting in increased processing time. In this paper, we propose to transform static word vectors to dynamic form by modelling the effect of neighbouring words and their sentiment strength in the AffectiveSpace. To evaluate the approach, we crawled Tweets from diverse topics and human annotation was used to label their sentiments. We also normalized the tweets to fix phonetic variations, spelling errors, and abbreviations manually. A total of 1432 out‐of‐vocabulary (OOV) texts and their IV texts made it to the final corpus with their corresponding polarity. To assess the quality of the corpus, we used several OOV classifiers such as linear regression and observed over 90% accuracy. Next, we inferred word vectors using a novel four‐gram model based on sentiment intensity and reported accuracy on both open domain and closed domain sentiment classifiers. We observed an improvement in the range of 4–20 on Twitter, Movie and Airline reviews over baselines.

Neural network-based methods, especially deep learning, have been a burgeoning area in AI research and have been successful in tackling the expanding data volume as we move into a digital age. Today, the neural network-based methods are not only used for low-level cognitive tasks, such as recognizing objects and spotting keywords, but they have also been deployed in various industrial information systems to assist high-level decision-making. In natural language processing, there have been two milestones for the past decade: one is word2vec [1], a group of neural models that learn word embeddings (vector representations of words) from large datasets; and one is the most recent GPT-based models [2], which combine reinforcement learning with a generative transformer in order to enable multi-round end-to-end conversations. While producing highly accurate predictions on datasets and generating human-like utterances, those neural network-based artifacts provide little understanding of the internal features and representations of the data. Many problems and concerns subsequently emerge from this black-box issue. Because some of the problems and concerns are also relevant in the context of sentiment analysis.

There is a large variety of products sold online and the websites are in several languages. Hence, it is desirable to train a model that can predict sentiments in different domains simultaneously. Previous authors have used deep learning to extract features from multiple domains. Here, each word is represented by a vector that is determined using co-occurrence data. Such a model requires that all sentences have the same length resulting in low accuracy. To overcome this challenge, we model the features in each sentence using a variable length tree called a Genetic Program. The polarity of clauses can be represented using mathematical operators such as '+' or '-' at internal nodes in the tree. The proposed model is evaluated on Amazon product reviews for different products and in different languages. We are able to outperform the accuracy of baseline multi-domain models in the range of 5-20%.

Automatic image manipulation can be used to make subtle changes at the pixel level resulting in morphism from one domain to another. This is desirable in tasks such as creating mock expressions for an individual or dynamic scene generation in autonomous driving. This type of morphism can be achieved using an adversarial model where the generator and the discriminator compete to produce fake images of the target domain. Due to high variance among the images, it is difficult to learn an optimal loss function. Previously, manifold matching of clusters in the source domain with labeled samples and the target domain that is generated was used to overcome this limitation. To generate videos it is common to use three-dimensional convolution however, such a model has very high complexity. Instead, in this paper we use manifold constrained model selection to do a constrained clustering of the combined manifold with fixed start and end images for the morphism. We show that each step in the principal path connecting the centroids is analogous to a single time delay in the video sequence. Hence, we can construct a cascade of models using samples from a pair of connected centroids such that one model is used to initialize the next. We apply the model to smile generation from neutral face expression and for predicting the next few frames while driving on real roads. We are able to outperform the baselines in the quality of images generated and the computational cost for training the model.

Building dialogue systems that can converse naturally with humans is a challenging yet intriguing problem of artificial intelligence. In open-domain human-computer conversation, where the conversational agent is expected to respond to human utterances in an interesting and engaging way, commonsense knowledge has to be integrated into the model effectively. In this paper, we investigate the impact of providing commonsense knowledge about the concepts covered in the dialogue. Our model represents the first attempt to integrating a large commonsense knowledge base into end-to-end conversational models. In the retrieval-based scenario, we propose a model to jointly take into account message content and related commonsense for selecting an appropriate response. Our experiments suggest that the knowledge-augmented models are superior to their knowledge-free counterparts.

Technology has enabled anyone with an Internet connection to easily create and share their ideas, opinions and content with millions of other people around the world. Much of the content being posted and consumed online is multimodal. With billions of phones, tablets and PCs shipping today with built-in cameras and a host of new video-equipped wearables like Google Glass on the horizon, the amount of video on the Internet will only continue to increase. It has become increasingly difficult for researchers to keep up with this deluge of multimodal content, let alone organize or make sense of it. Mining useful knowledge from video is a critical need that will grow exponentially, in pace with the global growth of content. This is particularly important in sentiment analysis, as both service and product reviews are gradually shifting from unimodal to multimodal. We present a novel method to extract features from visual and textual modalities using deep convolutional neural networks. By feeding such features to a multiple kernel learning classifier, we significantly outperform the state of the art of multimodal emotion recognition and sentiment analysis on different datasets.

Objective sentences lack sentiments and, hence, can reduce the accuracy of a sentiment classifier. Traditional methods prior to 2001 used hand-crafted templates to identify subjectivity and did not generalize well for resource-deficient languages such as Spanish. Later works published between 2002 and 2009 proposed the use of deep neural networks to automatically learn a dictionary of features (in the form of convolution kernels) that is portable to new languages. Recently, recurrent neural networks are being used to model alternating subjective and objective sentences within a single review. Such networks are difficult to train for a large vocabulary of words due to the problem of vanishing gradients. Hence, in this paper we consider use of a Lyapunov linear matrix inequality to classify Spanish text as subjective or objective by combining Spanish features and features obtained from the corresponding translated English text. The aligned features for each sentence are next evolved using multiple kernel learning. The proposed Lyapunov deep neural network outperforms baselines by over 10% and the features learned in the hidden layers improve our understanding subjective sentences in Spanish.

The advent of the Social Web has provided netizens with new tools for creating and sharing, in a time- and cost-efficient way, their contents, ideas, and opinions with virtually the millions of people connected to the World Wide Web. This huge amount of information, however, is mainly unstructured as specifically produced for human consumption and, hence, it is not directly machine-processable. In order to enable a more efficient passage from unstructured information to structured data, aspect-based opinion mining models the relations between opinion targets contained in a document and the polarity values associated with these. Because aspects are often implicit, however, spotting them and calculating their respective polarity is an extremely difficult task, which is closer to natural language understanding rather than natural language processing. To this end, Sentic LDA exploits common-sense reasoning to shift LDA clustering from a syntactic to a semantic level. Rather than looking at word co-occurrence frequencies, Sentic LDA leverages on the semantics associated with words and multi-word expressions to improve clustering and, hence, outperform state-of-the-art techniques for aspect extraction.

We introduce a probabilistic framework for building higher-order gene regulatory networks, which automatically finds the delays of regulatory interactions. A variable-order Markov chain Monte Carlo method with a new acceptance mechanism is proposed to find the optimal order and the structure of a dynamic Bayesian network (DBN). Experiments on cell cycle expression data indicate that the variable-order DBN (VDBN) better fits the data and gives biologically more plausible regulatory networks.