We tackle the problem of penalty selection of regularization on the basis of the minimum description length (MDL) principle. In particular, we consider that the design space of the penalty function is high-dimensional. In this situation, the luckiness-normalized-maximum-likelihood(LNML)-minimization approach is favorable, because LNML quantifies the goodness of regularized models with any forms of penalty functions in view of the minimum description length principle, and guides us to a good penalty function through the high-dimensional space. However, the minimization of LNML entails two major challenges: 1) the computation of the normalizing factor of LNML and 2) its minimization in high-dimensional spaces. In this paper, we present a novel regularization selection method (MDL-RS), in which a tight upper bound of LNML (uLNML) is minimized with local convergence guarantee. Our main contribution is the derivation of uLNML, which is a uniform-gap upper bound of LNML in an analytic expression. This solves the above challenges in an approximate manner because it allows us to accurately approximate LNML and then efficiently minimize it. The experimental results show that MDL-RS improves the generalization performance of regularized estimates specifically when the model has redundant parameters.

Building intelligent agents that can communicate with and learn from humans in natural language is of great value. Supervised language learning is limited by the ability of capturing mainly the statistics of training data, and is hardly adaptive to new scenarios or flexible for acquiring new knowledge without inefficient retraining or catastrophic forgetting. We highlight the perspective that conversational interaction serves as a natural interface both for language learning and for novel knowledge acquisition and propose a joint imitation and reinforcement approach for grounded language learning through an interactive conversational game. The agent trained with this approach is able to actively acquire information by asking questions about novel objects and use the just-learned knowledge in subsequent conversations in a one-shot fashion. Results compared with other methods verified the effectiveness of the proposed approach.

This paper describes the methodology followed to build a neural machine translation system in the biomedical domain for the English-Catalan language pair. This task can be considered a low-resourced task from the point of view of the domain and the language pair. To face this task, this paper reports experiments on a cascade pivot strategy through Spanish for the neural machine translation using the English-Spanish SCIELO and Spanish-Catalan El Peri\'odico database. To test the final performance of the system, we have created a new test data set for English-Catalan in the biomedical domain which is freely available on request.

Recent advances in artificial intelligence (AI) and machine learning have created a general perception that AI could be used to solve complex problems, and in some situations over-hyped as a tool that can be so easily used. Unfortunately, the barrier to realization of mass adoption of AI on various business domains is too high because most domain experts have no background in AI. Developing AI applications involves multiple phases, namely data preparation, application modeling, and product deployment. The effort of AI research has been spent mostly on new AI models (in the model training stage) to improve the performance of benchmark tasks such as image recognition. Many other factors such as usability, efficiency and security of AI have not been well addressed, and therefore form a barrier to democratizing AI. Further, for many real world applications such as healthcare and autonomous driving, learning via huge amounts of possibility exploration is not feasible since humans are involved. In many complex applications such as healthcare, subject matter experts (e.g. Clinicians) are the ones who appreciate the importance of features that affect health, and their knowledge together with existing knowledge bases are critical to the end results. In this paper, we take a new perspective on developing AI solutions, and present a solution for making AI usable. We hope that this resolution will enable all subject matter experts (eg. Clinicians) to exploit AI like data scientists.

A popular recent approach to answering open-domain questions is to first search for question-related passages and then apply reading comprehension models to extract answers. Existing methods usually extract answers from single passages independently. But some questions require a combination of evidence from across different sources to answer correctly. In this paper, we propose two models which make use of multiple passages to generate their answers. Both use an answer-reranking approach which reorders the answer candidates generated by an existing state-of-the-art QA model. We propose two methods, namely, strength-based re-ranking and coverage-based re-ranking, to make use of the aggregated evidence from different passages to better determine the answer. Our models have achieved state-of-the-art results on three public open-domain QA datasets: Quasar-T, SearchQA and the open-domain version of TriviaQA, with about 8 percentage points of improvement over the former two datasets.

The head of the U.S. military's Special Operations Command said Wednesday that Air Force gunships, needed to provide close air support for American commandos and U.S.-backed rebel fighters in Syria, were being "jammed" by "adversaries." Calling the electronic warfare environment in Syria "the most aggressive" on earth, Air Force Gen. Tony Thomas told an intelligence conference in Tampa that adversaries "are testing us every day, knocking our communications down, disabling our AC-130s, etc." Thomas' remarks, which were first reported by the website The Drive, come on the heels of reports that Russian forces are jamming U.S. surveillance drones flying over the war-torn nation. An Air Force AC-130 gunship was among the U.S. military aircraft used to kill dozens of Russian mercenaries in Syria in early February. The Pentagon said the mercenaries attacked an outpost manned by American commandos and U.S.-backed fighters of the Syrian Democratic Forces (SDF), comprising Syrian Kurdish and Arab fighters. Wednesday was not the first time General Thomas has been so forthcoming about Syria in a public setting.

Artificial intelligence could destabilize the delicate balance of nuclear deterrence, inching the world closer to catastrophe, according to a working group of experts convened by RAND. New smarter, faster intelligence analysis from AI agents, combined with more sensor and open-source data, could convince countries that their nuclear capability is increasingly vulnerable. That may cause them to take more drastic steps to keep up with the U.S. Another worrying scenario: commanders could make decisions to launch strikes based on advice from AI assistants that have been fed wrong information. Last May and June, RAND convened a series of workshops, bringing together experts from nuclear security, artificial intelligence, government, and industry. The workshops produced a report, released on Tuesday, that underlines how AI promises to rapidly improve Country A's ability to target Country B's nuclear weapons.

Hong Kong-based shipping company Orient Overseas Container Line (OOCL) has teamed up with Microsoft's research arm in Asia to advance the application of Artificial Intelligence research in the shipping industry. The collaboration will look for ways to use AI to improve shipping network operations and achieve efficiencies within OOCL's business. The project is expected to nurture over 200 AI developers over the next 12 months. OOCL sees AI as key to the it's digital transformation. The company already uses machine learning in some its operations and has as a talent base of over 1,000 developers located in San Jose, Hong Kong, Zhuhai, Shanghai and Manila.

Brussels has called for a €20bn (£14bn) cash injection for artificial intelligence research, while pouring cold water over controversial plans to give robots human rights. The European commission wants governments and private companies to boost research and innovation spending on AI, amid rising concern Europe is losing ground to the US and China, where most leading AI firms are based. Health, transport and agriculture are among the areas the commission would like researchers to prioritise. But the commission distanced itself from proposals to give the most advanced robots the legal status of personhood. "I don't think it will happen," Andrus Ansip, a commission vice-president in charge of digital single-market policy told journalists.

Long, narrowly spaced rows of shelves fill a multi-storey building about the size of two sports fields. The shelves are lined with open containers of food and water. It is warm, humid and dark all year round, with freedom to roam to find food and reproduce. Fully sealed like a prison, it has strict limitations on access to visitors. From birth to death, inhabitants never see the sun.