Natural language processing (NLP) can be done using either top-down (theory driven) and bottom-up (data driven) approaches, which we call mechanistic and phenomenological respectively. The approaches are frequently considered to stand in opposition to each other. Examining some recent approaches in deep learning we argue that deep neural networks incorporate both perspectives and, furthermore, that leveraging this aspect of deep learning may help in solving complex problems within language technology, such as modelling language and perception in the domain of spatial cognition.

Deep reinforcement learning, and especially the Asynchronous Advantage Actor-Critic algorithm, has been successfully used to achieve super-human performance in a variety of video games. Starcraft II is a new challenge for the reinforcement learning community with the release of pysc2 learning environment proposed by Google Deepmind and Blizzard Entertainment. Despite being a target for several AI developers, few have achieved human level performance. In this project we explain the complexities of this environment and discuss the results from our experiments on the environment. We have compared various architectures and have proved that transfer learning can be an effective paradigm in reinforcement learning research for complex scenarios requiring skill transfer.

Deep Learning is one of the newest trends in Machine Learning and Artificial Intelligence research. It is also one of the most popular scientific research trends now-a-days. Deep learning methods have brought revolutionary advances in computer vision and machine learning. Every now and then, new and new deep learning techniques are being born, outperforming state-of-the-art machine learning and even existing deep learning techniques. In recent years, the world has seen many major breakthroughs in this field. Since deep learning is evolving at a huge speed, its kind of hard to keep track of the regular advances especially for new researchers. In this paper, we are going to briefly discuss about recent advances in Deep Learning for past few years.

Neural models combining representation learning and reasoning in an end-to-end trainable manner are receiving increasing interest. However, their use is severely limited by their computational complexity, which renders them unusable on real world datasets. We focus on the Neural Theorem Prover (NTP) model proposed by Rockt{\"{a}}schel and Riedel (2017), a continuous relaxation of the Prolog backward chaining algorithm where unification between terms is replaced by the similarity between their embedding representations. For answering a given query, this model needs to consider all possible proof paths, and then aggregate results - this quickly becomes infeasible even for small Knowledge Bases (KBs). We observe that we can accurately approximate the inference process in this model by considering only proof paths associated with the highest proof scores. This enables inference and learning on previously impracticable KBs.

This work develops novel strategies for optimal planning with semantic observations using continuous state Partially Observable Markov Decision Processes (CPOMDPs). Two major innovations are presented in relation to Gaussian mixture (GM) CPOMDP policy approximation methods. While existing methods have many theoretically nice properties, they are hampered by the inability to efficiently represent and reason over hybrid continuous-discrete probabilistic models. The first major innovation is the derivation of closed-form variational Bayes GM approximations of Point-Based Value Iteration Bellman policy backups, using softmax models of continuous-discrete semantic observation probabilities. A key benefit of this approach is that dynamic decision-making tasks can be performed with complex non-Gaussian uncertainties, while also exploiting continuous dynamic state space models (thus avoiding cumbersome and costly discretization). The second major innovation is a new clustering-based technique for mixture condensation that scales well to very large GM policy functions and belief functions. Simulation results for a target search and interception task with semantic observations show that the GM policies resulting from these innovations are more effective than those produced by other state of the art GM and Monte Carlo based policy approximations, but require significantly less modeling overhead and runtime cost. Additional results demonstrate the robustness of this approach to model errors.

Taxi destination prediction is a very important task for optimizing the efficiency of electronic dispatching systems, thus allowing relevant advantages for both taxi companies and customers. In fact, during periods of high demand, there should be a taxi whose current ride will end near a requested pick up location from a new customer. If an electronic dispatcher is able to know in advance where all taxi drivers will end their current ride, it will also be able to better allocate its resources, identifying which taxi to assign to each call. Moreover, automatic systems for the taxi mobility monitoring collect data that, integrated with other information sources, can help in understanding daytime human mobility routines. In this paper, we introduce a novel approach for addressing the taxi destination prediction problem, based on Recurrent Neural Networks (RNNs) applied to a regression setting. RNNs are trained based on the individual drivers' history and on geographical information (i.e., points of interest), using only the starting point of each ride (with no knowledge about the whole trajectory). The proposed approach was tested on the dataset of the ECML/PKDD Discovery Challenge 2015 - based on the city of Porto - obtaining better results with respect to the competition winner, whilst using less information, and on Manhattan and San Francisco datasets.

Taking a look back at another week of news from Cupertino, this week's Apple Loop includes FaceID for every new iPhone in 2018, the corners being cut for a cheaper iPhone X, secrets inside the new MacBook Pro, falling iPhone sales in India, the lack of fast chargers for the iPhone, Photoshop for the iPad, and iOS 12's updated USB security. Apple Loop is here to remind you of a few of the very many discussions that have happened around Apple over the last seven days (and you can read my weekly digest of Android news here on Forbes). Thanks to a regulatory listing made by Apple in the Eurasian database, we now know there are three new models of iPhone on the way, with a number of'sub models' in each range. Spotted by Consomac, Apple has chosen to publicly file identifiers for all its new iPhones in the Eurasian database, and it confirms three distinct designs will be coming to market. There are three clear runs here: A19, A20 and A21.

Cisco has announced a major investment into the UK's AI scene. The networking giant has revealed a $100m (£77m) commitment that includes a new partnership with University College London (UCL) aimed at creating one of the world's largest AI research centres. The facility will house between 200 and 250 masters students and researchers looking to take on some of the biggest AI industry challenges, as well as continue Cisco's 30-year relationship with UCL. "It's wonderful to renew our partnership with Cisco and work together to upskill the UK in machine learning and artificial intelligence," said UCL provost Professor Michael Arthur. "I'm particularly looking forward to opening the new AI Centre in the coming months to position us as a sector leader in computer science."

The International Space Station should prepare for the arrival of its first android crew members, Russian state media says. The Roskosmos space agency has approved a preliminary plan to send a pair of humanoid robots called FEDOR into space in August 2019, according to "a source in the space and rocket industry" quoted by the RIA Novosti website. Robots in space have become commonplace for space superpowers: the U.S. has two operational Mars rovers, China has a lunar lander on the moon and more on the way, and Russia has several now-defunct rovers on both the moon and Mars. In 2011, NASA sent Robonaut 2, a 330-pound manually controlled "humanoid" robot, to the ISS to look into how such robots might be used to perform simple, repetitive, or especially dangerous tasks. But while previous robots were shot into space on as cargo, Russia's pair of FEDORs -- the acronym stands for Final Experimental Demonstration Object Research -- will "fly for the first time to the ISS as crew members, and not as cargo in the transport compartment," RIA Novosti wrote, adding that the robots will fly in an otherwise unmanned Soyuz rocket.

Last month, two parliamentary reports on the involvement of the British intelligence services in torture and rendition were released last week. What has been hypothesized by several journalists is now confirmed: that British functionaries--to include soldiers, civil servants and intelligence officers with MI5 and MI6--knew about and participated in a vast array of human rights abuses committed during the capture and interrogation of terrorism suspects. The Guardian's Peter Beaumont writes with great contempt for what has transpired since 2001, "[A]s it is now quite clear, it was all a bloody lie. The answers given to journalists at the Observer over the years, as well as colleagues at The Guardian and those at other news organisations, as they investigated these allegations, were rotten with untruth and evasion." Governments' lying to their citizens about covert wars is hardly new, nor is the pervasive use of kidnapping of terrorism suspects by the CIA to include its many "black sites."