Although stance detection has made great progress in the past few years, it is still facing the problem of unseen targets. In this study, we investigate the domain difference between targets and thus incorporate attention-based conditional encoding with adversarial domain generalization to perform unseen target stance detection. Experimental results show that our approach achieves new state-of-the-art performance on the SemEval-2016 dataset, demonstrating the importance of domain difference between targets in unseen target stance detection.

Sign language translation (SLT) aims to interpret sign video sequences into textbased natural language sentences. Sign videos consist of continuous sequences of sign gestures with no clear boundaries in between. Existing SLT models usually represent sign visual features in a frame-wise manner so as to avoid needing to explicitly segmenting the videos into isolated signs. However, these methods neglect the temporal information of signs and lead to substantial ambiguity in translation. In this paper, we explore the temporal semantic structures of sign videos to learn more discriminative features. To this end, we first present a novel sign video segment representation which takes into account multiple temporal granularities, thus alleviating the need for accurate video segmentation. Taking advantage of the proposed segment representation, we develop a novel hierarchical sign video feature learning method via a temporal semantic pyramid network, called TSPNet. Specifically, TSPNet introduces an inter-scale attention to evaluate and enhance local semantic consistency of sign segments and an intra-scale attention to resolve semantic ambiguity by using non-local video context. Experiments show that our TSPNet outperforms the state-of-the-art with significant improvements on the BLEU score (from 9.58 to 13.41) and ROUGE score (from 31.80 to 34.96) on the largest commonly-used SLT dataset.

While reinforcement learning has achieved considerable successes in recent years, state-of-the-art models are often still limited by the size of state and action spaces. Model-free reinforcement learning approaches use some form of state representations and the latest work has explored embedding techniques for actions, both with the aim of achieving better generalization and applicability. However, these approaches consider only states or actions, ignoring the interaction between them when generating embedded representations. In this work, we propose a new approach for jointly embedding states and actions that combines aspects of model-free and model-based reinforcement learning, which can be applied in both discrete and continuous domains. Specifically, we use a model of the environment to obtain embeddings for states and actions and present a generic architecture that uses these to learn a policy. In this way, the embedded representations obtained via our approach enable better generalization over both states and actions by capturing similarities in the embedding spaces. Evaluations of our approach on several gaming and recommender system environments show it significantly outperforms state-of-the-art models in discrete domains with large state/action space, thus confirming the efficacy of joint embedding and its overall superior performance.

Text autoencoders are commonly used for conditional generation tasks such as style transfer. We propose methods which are plug and play, where any pretrained autoencoder can be used, and only require learning a mapping within the autoencoder's embedding space, training embedding-to-embedding (Emb2Emb). This reduces the need for labeled training data for the task and makes the training procedure more efficient. Crucial to the success of this method is a loss term for keeping the mapped embedding on the manifold of the autoencoder and a mapping which is trained to navigate the manifold by learning offset vectors. Evaluations on style transfer tasks both with and without sequence-to-sequence supervision show that our method performs better than or comparable to strong baselines while being up to four times faster.

To drive progress in the field of data science, we propose 10 challenge areas for the research community to pursue. Since data science is broad, with methods drawing from computer science, statistics, and other disciplines, and with applications appearing in all sectors, these challenge areas speak to the breadth of issues spanning science, technology, and society. We preface our enumeration with meta-questions about whether data science is a discipline. We then describe each of the 10 challenge areas. The goal of this article is to start a discussion on what could constitute a basis for a research agenda in data science, while recognizing that the field of data science is still evolving. Although data science builds on knowledge from computer science, engineering, mathematics, statistics, and other disciplines, data science is a unique field with many mysteries to unlock: fundamental scientific questions and pressing problems of societal importance.

LoginRadius Engineering Blog - Posts related to Deep Learning

Deep Learning is changing the way we look at technologies. There is a lot of excitement around Artificial Intelligence (AI) along with its branches namely Machine Learning (ML) and Deep Learning at the moment. With massive amounts of computational power, machines can now recognize objects and translate speech in real time. Artificial intelligence is finally getting smart. It's predicted that many deep learning applications will affect your life in the near future.

DNA and RNA have been compared to "instruction manuals" containing the information needed for living "machines" to operate. But while electronic machines like computers and robots are designed from the ground up to serve a specific purpose, biological organisms are governed by a much messier, more complex set of functions that lack the predictability of binary code. Inventing new solutions to biological problems requires teasing apart seemingly intractable variables--a task that is daunting to even the most intrepid human brains. Two teams of scientists from the Wyss Institute at Harvard University and the Massachusetts Institute of Technology have devised pathways around this roadblock by going beyond human brains; they developed a set of machine learning algorithms that can analyze reams of RNA-based "toehold" sequences and predict which ones will be most effective at sensing and responding to a desired target sequence. As reported in two papers published concurrently today in Nature Communications, the algorithms could be generalizable to other problems in synthetic biology as well, and could accelerate the development of biotechnology tools to improve science and medicine and help save lives.

Unlike any other time, the past decade has seen unprecedented development in the field of Artificial Intelligence (AI). There are a lot of talks on machine learning doing things humans currently do in our workplace. Deep learning is leading in some of the fronts of machine learning making practical changes. Deep learning is an artificial intelligence function that imitates the working of the human brain in processing data and creating patterns for use in decision making. Deep learning is a subfield of machine learning concerned with algorithms inspired by the structure and function of the brain called artificial neural network (ANN).

It's no secret that a lot of people, including some very influential personalities, harbor the fear that AI or artificial intelligence could one day surpass human intelligence and take over the world. But while such a threat may not be as far-fetched, right now, there's one big stumbling block that's keeping robots and machines from even scratching the surface of what it is like to be human. And that's the ability to reason and think logically. Researchers have been trying to figure out how to incorporate this ability into computers. Now Google's DeepMind might be moving ahead in the race as it has developed an algorithm that enables AI to handle the most basic form of reasoning, that is, relational reasoning.