Bilingual word embeddings have been widely used to capture the correspondence of lexical semantics in different human languages. However, the cross-lingual correspondence between sentences and lexicons is less studied, despite that this correspondence can largely benefit many applications, such as cross-lingual semantic search and question answering. To bridge this gap, we propose a neural embedding model that leverages bilingual dictionaries. The proposed model is trained to map the literal word definitions to the cross-lingual target words, for which we explore with different sentence encoding techniques. To enhance the learning process on limited resources, our model adopts several critical learning strategies, including multi-task learning on different bridges of languages, and joint learning of the dictionary model with a bilingual word embedding model. We conduct experiments on two tasks: (i) cross-lingual reverse dictionary retrieval, and (ii) bilingual paraphrase identification. In the former task, we demonstrate that our model is capable of comprehending bilingual concepts based on descriptions, and we also highlight the effectiveness of proposed learning strategies. In the latter one, we show that the proposed model effectively associates sentences in different languages via a shared embedding space, and outperforms existing approaches in identifying bilingual paraphrases.

Humans have been pondering the potential of artificial intelligence for thousands of years. Ancient Greeks believed, for example, that a bronze automaton named Talos protected the island of Crete from maritime adversaries. But AI only moved from the mythical realm to the real world in the last half-century, beginning with legendary computer scientist Alan Turing's foundational 1950 essay asked and provided a framework for answering the provocative question, "Can machines think?" At that time, the United States was in the midst of the Cold War. Congressional representatives decided to invest heavily in artificial intelligence as part of a larger security strategy.

We propose a novel geometric approach for learning bilingual mappings given monolingual embeddings and a bilingual dictionary. Our approach decouples learning the transformation from the source language to the target language into (a) learning rotations for language-specific embeddings to align them to a common space, and (b) learning a similarity metric in the common space to model similarities between the embeddings. We model the bilingual mapping problem as an optimization problem on smooth Riemannian manifolds. We show that our approach outperforms previous approaches on the bilingual lexicon induction and cross-lingual word similarity tasks. We also generalize our framework to represent multiple languages in a common latent space. In particular, the latent space representations for several languages are learned jointly, given bilingual dictionaries for multiple language pairs. We illustrate the effectiveness of joint learning for multiple languages in zero-shot word translation setting.

Recent studies have shown that reinforcement learning (RL) is an effective approach for improving the performance of neural machine translation (NMT) system. However, due to its instability, successfully RL training is challenging, especially in real-world systems where deep models and large datasets are leveraged. In this paper, taking several large-scale translation tasks as testbeds, we conduct a systematic study on how to train better NMT models using reinforcement learning. We provide a comprehensive comparison of several important factors (e.g., baseline reward, reward shaping) in RL training. Furthermore, to fill in the gap that it remains unclear whether RL is still beneficial when monolingual data is used, we propose a new method to leverage RL to further boost the performance of NMT systems trained with source/target monolingual data. By integrating all our findings, we obtain competitive results on WMT14 English- German, WMT17 English-Chinese, and WMT17 Chinese-English translation tasks, especially setting a state-of-the-art performance on WMT17 Chinese-English translation task.

Most of the Neural Machine Translation (NMT) models are based on the sequence-to-sequence (Seq2Seq) model with an encoder-decoder framework equipped with the attention mechanism. However, the conventional attention mechanism treats the decoding at each time step equally with the same matrix, which is problematic since the softness of the attention for different types of words (e.g. content words and function words) should differ. Therefore, we propose a new model with a mechanism called Self-Adaptive Control of Temperature (SACT) to control the softness of attention by means of an attention temperature. Experimental results on the Chinese-English translation and English-Vietnamese translation demonstrate that our model outperforms the baseline models, and the analysis and the case study show that our model can attend to the most relevant elements in the source-side contexts and generate the translation of high quality.

Artificial intelligence (AI) is surpassing human performance in a growing number of domains. However, there is limited evidence of its economic effects. Using data from a digital platform, we study a key application of AI: machine translation. We find that the introduction of a machine translation system has significant increased international trade on this platform, increasing exports by 17.5%. Furthermore, heterogeneous treatment effects are all consistent with a substantial reduction in translation-related search costs. Our results provide causal evidence that language barriers significantly hinder trade and that AI has already begun to improve economic efficiency in at least one domain.

We propose a simple modification to existing neural machine translation (NMT) models that enables using a single universal model to translate between multiple languages while allowing for language specific parameterization, and that can also be used for domain adaptation. Our approach requires no changes to the model architecture of a standard NMT system, but instead introduces a new component, the contextual parameter generator (CPG), that generates the parameters of the system (e.g., weights in a neural network). This parameter generator accepts source and target language embeddings as input, and generates the parameters for the encoder and the decoder, respectively. The rest of the model remains unchanged and is shared across all languages. We show how this simple modification enables the system to use monolingual data for training and also perform zero-shot translation. We further show it is able to surpass state-of-the-art performance for both the IWSLT-15 and IWSLT-17 datasets and that the learned language embeddings are able to uncover interesting relationships between languages.

For more than 200 years, the synthesis of organic molecules has remained one of the most important tasks in organic chemistry. The work of chemists has scientific and commercial implications that range from the production of Aspirin to that of Nylon. Yet, little has been done to change age-old practices dramatically and allow a new era of productivity based on pioneering artificial intelligence (AI) science and technologies. The challenge for organic chemists in fields such as chemistry, materials science, oil and gas, and life sciences is that there are hundreds of thousands of reactions and, while it is manageable to remember a few dozen in a narrow specialist's field, it's impossible to be an expert generalist. To address this, we asked ourselves, can we use deep learning and artificial intelligence to predict reactions of organic compounds?

Translating between human languages is something which artificial intelligence – specifically machine learning – has proven to be very competent at. So much so that the CEO of one of the world's largest employers of human translators has warned that many of them should be facing up to the stark reality of losing their job to a machine. One Hour Translation CEO Ofer Shoshan told me that within one to three years, neural machine technology (NMT) translators will carry out more than 50% of the work handled by the $40 billion market. His words stand in stark contrast to the often-repeated maxim that, in the near future at least, artificial intelligence will primarily augment, rather than replace, human professionals. Shoshan told me that the quality of machine translation has improved by leaps and bounds in recent years, to the point where half a million human translators and 21,000 agencies could soon find themselves out of work.

Besides the Dutch-English and Indonesian-English translation systems we offer through our client apps and connectors, we have built neural machine systems for the Turkish-English, Latvian-English and Spanish-English language pairs. Whether you are a professional translator, translation company or a business user, we can put artificial intelligence and deep learning to work for you. If you already have parallel texts (texts with source language & translation) we can use these as a basis for building a customised system for you. If you don't have such material we can usually build a useful baseline system from freely available public resources. Your own neural machine translation system can then be placed on a secure server "in the cloud" or on a dedicated server located in your offices and accessible only on your corporate network.