Deep neural networks with their large number of parameters are highly flexible learning systems. The high flexibility in such networks brings with some serious problems such as overfitting, and regularization is used to address this problem. A currently popular and effective regularization technique for controlling the overfitting is dropout. Often, large data collections required for neural networks contain sensitive information such as the medical histories of patients, and the privacy of the training data should be protected. In this paper, we modify the recently proposed variational dropout technique which provided an elegant Bayesian interpretation to dropout, and show that the intrinsic noise in the variational dropout can be exploited to obtain a degree of differential privacy. The iterative nature of training neural networks presents a challenge for privacy-preserving estimation since multiple iterations increase the amount of noise added. We overcome this by using a relaxed notion of differential privacy, called concentrated differential privacy, which provides tighter estimates on the overall privacy loss. We demonstrate the accuracy of our privacy-preserving variational dropout algorithm on benchmark datasets.

Knowledge representation is a long-history topic in AI, which is very important. A variety of models have been proposed for knowledge graph embedding, which projects symbolic entities and relations into continuous vector space. However, most related methods merely focus on the data-fitting of knowledge graph, and ignore the interpretable semantic expression. Thus, traditional embedding methods are not friendly for applications that require semantic analysis, such as question answering and entity retrieval. To this end, this paper proposes a semantic representation method for knowledge graph \textbf{(KSR)}, which imposes a two-level hierarchical generative process that globally extracts many aspects and then locally assigns a specific category in each aspect for every triple. Since both aspects and categories are semantics-relevant, the collection of categories in each aspect is treated as the semantic representation of this triple. Extensive experiments show that our model outperforms other state-of-the-art baselines substantially.

Check out the session "Get Your Hard Hat: Intelligent Industrial Systems with Deep Reinforcement Learning" at the AI Conference in Beijing, April 10-13, 2018. Best price ends January 26. The flurry of headlines surrounding AlphaGo Zero (the most recent version of DeepMind's AI system for playing Go) means interest in reinforcement learning (RL) is bound to increase. Next to deep learning, RL is among the most followed topics in AI. For most companies, RL is something to investigate and evaluate but few organizations have identified use cases where RL may play a role. As we enter 2018, I want to briefly describe areas where RL has been applied.

Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We'll also be posting a weekly calendar of upcoming robotics events for the next two months; here's what we have so far (send us your events!): Let us know if you have suggestions for next week, and enjoy today's videos. This is the FX-2 Giant Human Riding Robot from KAIST HuboLab and Rainbow Robotics, and that's all I know about it. Yuichiro Katsumoto, a "gadget creator" based in Singapore, wrote in to share this video of the robotic kinetic typography that he's been working on for the last year.

Rapid advances in AI have spawned a number of recent initiatives that aim to convince engineers, programmers, and others to prioritize ethical considerations in their work--but almost all of them have originated in rich Western countries. An effort from the huge engineering association IEEE is now trying to change that, with its own AI ethics proposal that it says will be a global, multilingual collaboration. In the past two years alone, a raft of new efforts to explore ethics in AI have launched, including the Elon Musk–backed nonprofit OpenAI, the corporate alliance Partnership on AI, Carnegie Mellon University's AI ethics research center, and the Ethics & Society research unit at Google's AI subsidiary DeepMind. But most of these projects are based in the U.S. or U.K., are led by a small group of researchers, and issue updates only in English, which could limit their ability to foster AI that benefits all of humanity, not just those in developed countries. Since 2016, a group called the IEEE Global Initiative for Ethical Considerations in Artificial Intelligence and Autonomous Systems has been writing a document called "Ethically Aligned Design" that recommends societal and policy guidelines for technologies such as chatbots and home robots. This week, the group unveiled an updated version of the document that integrates feedback from people in East Asia, Latin America, the Middle East, and other regions.

Whether financial institutions are looking for improved customer service, risk management, fraud prevention, investment prediction or cybersecurity, the scopes of machine learning and artificial intelligence are limitless. In the modern era of the digital economy, technological advancements are no longer a luxury for the organizations, but a necessity to outsmart their competitors and business growth. With the technological advancements in the recent times, the impact of Machine Learning (ML) and Artificial Intelligence (AI) are very critical than ever before. Previously, we discussed the scopes of big data and data science in banking and financial services. In this article will explain in detail about ML and AI, and their scopes in banking and financial services. Apparently, in order to be successful and making an impact, the banks and financial institutions need to make machine learning and artificial intelligence an expansion of their big data and data analytics approach.

SINGAPORE - Start-ups working on artificial intelligence (AI) will now have a centre in Suzhou to help them venture into the Chinese market. Besides offering one-stop services for the start-ups, the centre aims to incubate 15 innovative companies in its first five years and promote National University of Singapore (NUS) technologies. The NUS Artificial Intelligence Innovation and Commercialisation Centre is a $20 million collaboration between NUS and Suzhou Industrial Park Administrative Committee. It will start operating from January 2018, focusing on advancing "AI research, innovation, application and commercialisation across various areas, including healthcare, financial technology and smart city", NUS said in a statement. "It will provide technology incubation support and solutions to start-ups and small businesses, helping them solve problems with AI as well as accelerate the adoption of AI to promote and form an AI ecosystem and industrial chain in the Suzhou area." The centre will also provide NUS students with opportunities for internship stints as well as lectures and symposiums to connect professionals in the artificial intelligence field.

"Why incur the expense of generating and collecting all of this IoT data if you're not going to monetize it?" Organizations are racing to embrace the Internet of Things (IoT) as the pundits create "visions of sugar-plums dancing in their heads." McKinsey Global Institute released their study "The Internet of Things: Mapping the Value beyond the Hype" in June 2015 that highlighted the staggering financial value that IoT could create! The folks at Wikibon provided a perspective on the sources of "IoT monetization" in their recent research titled "Harvesting Value at the Edge" written by the always delightful and provocative Neil Raden. IoT, though a useful application of available technology, and well-defined at the hardware and network levels, the heart of IoT, that part that yields the real value, is edge analytics.

At the end of 2017, there will be 8.4 billion connected things in use worldwide up 31 percent from 2016, and this figure is expected to reach 20.4 billion by 2020. When Internet of Things (IoT) as an industry took off in India, it spawned a host of startups selling edge devices that could gather and crunch data from corporate customers. These startups ran into one fundamental problem, which was data lifting. The data was so voluminous that these startups took so much time to organise them that they ran out of money to keep the companies afloat. In the end, their services were just organising data for customers with very little insights.

Alisa, a virtual reality assistant developed by Russia's tech giant Yandex, has been nominated to become the next president of Russia by thousands of supporters across the country. The favorite for elections scheduled for March 2018, President Vladimir Putin, announced his candidacy on Wednesday. Putin is seeking a fourth term in office which would extend his tenure into 2024. Other contenders for the presidency will likely include familiar political figures from the pro-Kremlin Communist Party and the nationalist Liberal Democratic Party, as well as new faces such as the former reality TV star Ksenia Sobchak and the business-oriented Party of Growth leader Boris Titov. It appears they will now be challenged by Alisa's progressive promise to bring "the political system of the future, built exclusively on rational decisions made on the basis of clear algorithms," the Lenta news website reported.