With the development of deep neural networks, the size of network models becomes larger and larger. Model compression has become an urgent need for deploying these network models to mobile or embedded devices. Model quantization is a representative model compression technique. Although a lot of quantization methods have been proposed, many of them suffer from a high quantization error caused by a long-tail distribution of network weights. In this paper, we propose a novel quantization method, called weight normalization based quantization (WNQ), for model compression. WNQ adopts weight normalization to avoid the long-tail distribution of network weights and subsequently reduces the quantization error. Experiments on CIFAR-100 and ImageNet show that WNQ can outperform other baselines to achieve state-of-the-art performance.

Machine learning plays an increasing role in intelligent tutoring systems as both the amount of data available and specialization among students grow. Nowadays, these systems are frequently deployed on mobile applications. Users on such mobile education platforms are dynamic, frequently being added, accessing the application with varying levels of focus, and changing while using the service. The education material itself, on the other hand, is often static and is an exhaustible resource whose use in tasks such as problem recommendation must be optimized. The ability to update user models with respect to educational material in real-time is thus essential; however, existing approaches require time-consuming re-training of user features whenever new data is added. In this paper, we introduce a neural pedagogical agent for real-time user modeling in the task of predicting user response correctness, a central task for mobile education applications. Our model, inspired by work in natural language processing on sequence modeling and machine translation, updates user features in real-time via bidirectional recurrent neural networks with an attention mechanism over embedded question-response pairs. We experiment on the mobile education application SantaTOEIC, which has 559k users, 66M response data points as well as a set of 10k study problems each expert-annotated with topic tags and gathered since 2016. Our model outperforms existing approaches over several metrics in predicting user response correctness, notably out-performing other methods on new users without large question-response histories. Additionally, our attention mechanism and annotated tag set allow us to create an interpretable education platform, with a smart review system that addresses the aforementioned issue of varied user attention and problem exhaustion.

Knowledge graph embeddings rank among the most successful methods for link prediction in knowledge graphs, i.e., the task of completing an incomplete collection of relational facts. A downside of these models is their strong sensitivity to model hyperparameters, in particular regularizers, which have to be extensively tuned to reach good performance [Kadlec et al., 2017]. We propose an efficient method for large scale hyperparameter tuning by interpreting these models in a probabilistic framework. After a model augmentation that introduces per-entity hyperparameters, we use a variational expectation-maximization approach to tune thousands of such hyperparameters with minimal additional cost. Our approach is agnostic to details of the model and results in a new state of the art in link prediction on standard benchmark data.

We introduce a modular system that can be deployed on any Kubernetes cluster for question answering via REST API. This system, called Katecheo, includes four configurable modules that collectively enable identification of questions, classification of those questions into topics, a search of knowledge base articles, and reading comprehension. We demonstrate the system using publicly available, pre-trained models and knowledge base articles extracted from Stack Exchange sites. However, users can extend the system to any number of topics, or domains, without the need to modify any of the model serving code. All components of the system are open source and available under a permissive Apache 2 License.

There is renewed interest in simulating language emergence among deep neural agents that communicate to jointly solve a task, spurred by the practical aim to develop language-enabled interactive AIs, as well as by theoretical questions about the evolution of human language. However, optimizing deep architectures connected by a discrete communication channel (such as that in which language emerges) is technically challenging. We introduce EGG, a toolkit that greatly simplifies the implementation of emergent-language communication games. EGG's modular design provides a set of building blocks that the user can combine to create new games, easily navigating the optimization and architecture space. We hope that the tool will lower the technical barrier, and encourage researchers from various backgrounds to do original work in this exciting area.

In 1863 five members of the Chōshū han in Japan made a secret journey to University College London in the UK to study. At the time of their departure, travel overseas was illegal in Japan, nonetheless all five students made an impact on the University that is commemorated to this day, and returned to establish institutions that augured a new era in their homeland, including the National Mint, the Japanese railways and the first Prime Minister. In the same spirit of international collaborations fostering pioneering innovations, materials and data scientists met at the Japanese Embassy in London on Friday 21st June during the "Season of Culture" to discuss "Global Trends in Research on Data-driven Discovery in Materials Science". The event was the 10th scholarly colloquium organized by the journal Science and Technology of Advanced Materials (STAM). Developments in data present an interesting example in science diplomacy where science and technology may facilitate a diplomatic agenda that in turn serves the interests of science.

The world of computing, from chips to software to systems, is going to change dramatically in coming years as a result of the spread of machine learning. We may still refer to these computers as "Universal Turing Machines," as we have for eighty years or more. But in practice they will be different from the way they have been built and used up to now. Such a change is of interest both to anyone who cares about what computers do, and to anyone who's interested in machine learning in all its forms. In February, Facebook's head of A.I. research, Yann LeCun, gave a talk at the International Solid State Circuits Conference in San Francisco, one of the longest running computer chip conferences in the world.

These "breakfasts" are intended to be a convivial place for exchange with the invited expert, but also between the Council of Europe community and Strasbourg academics interested in this subject. The first meeting will take place on Thursday 27 June 2019 (8h30 - 10h30) at the Maison interuniversitaire des sciences de l'Homme (MISHA), 5 allée du Général Rouvillois (Strasbourg) - Tram E / Stop Observatoire. Frédéric Wickert, an entrepreneur and expert with the Council of Europe, will take part in this first exercise on the theme: the different faces of AI.

Organised around the three main pillars that constitute the Council of Europe core values, human rights, democracy, and the rule of law, panel discussions addressed the challenges and opportunities of AI development for individuals, for societies, and for the viability of our legal and institutional frameworks, and explored options for ensuring that effective mechanisms of democratic oversight are in place.

Nuro, the self-driving delivery startup, is teaming up with Domino's to launch a pilot for driverless pizza delivery in Houston, Texas, the companies announced Monday. Starting later this year, Domino's will use Nuro's driverless fleet of custom-built robot cars to deliver pizza to select Houston residents who place orders online. Nuro, which was founded by two ex-members of Google's pioneering self-driving team, has been using its fleet of R1 robot cars to deliver groceries to residents of Scottsdale, Arizona, and more recently, Houston. If the pilot with Domino's goes well, it's safe to assume Nuro will look to expand it to other markets as well. Nuro has been ramping up its activities in recent months since receiving a $1 billion investment from Japanese tech company SoftBank.