We analyze the global convergence of gradient descent for deep linear residual networks by proposing a new initialization: zero-asymmetric (ZAS) initialization. It is motivated by avoiding stable manifolds of saddle points. We prove that under the ZAS initialization, for an arbitrary target matrix, gradient descent converges to an $\varepsilon$-optimal point in $O(L^3 \log(1/\varepsilon))$ iterations, which scales polynomially with the network depth $L$. Our result and the $\exp(\Omega(L))$ convergence time for the standard initialization (Xavier or near-identity) [Shamir, 2018] together demonstrate the importance of the residual structure and the initialization in the optimization for deep linear neural networks, especially when $L$ is large.

This project challenges the car racing problem from OpenAI gym environment. The problem is very challenging since it requires computer to finish the continuous control task by learning from pixels. To tackle this challenging problem, we explored two approaches including evolutionary algorithm based genetic multi-layer perceptron and double deep Q-learning network. The result shows that the genetic multi-layer perceptron can converge fast but when training many episodes, double deep Q-learning can get better score. We analyze the result and draw a conclusion that for limited hardware resources, using genetic multi-layer perceptron sometimes can be more efficient.

A coalition of agents, or a single agent, has an ethical dilemma between several statements if each joint action of the coalition forces at least one specific statement among them to be true. For example, any action in the trolley dilemma forces one specific group of people to die. In many cases, agents face ethical dilemmas because they are restricted in the amount of the resources they are ready to sacrifice to overcome the dilemma. The paper presents a sound and complete modal logical system that describes properties of dilemmas for a given limit on a sacrifice.

This paper analyzes challenges in cloze-style reading comprehension on multiparty dialogue and suggests two new tasks for more comprehensive predictions of personal entities in daily conversations. We first demonstrate that there are substantial limitations to the evaluation methods of previous work, namely that randomized assignment of samples to training and test data substantially decreases the complexity of cloze-style reading comprehension. According to our analysis, replacing the random data split with a chronological data split reduces test accuracy on previous single-variable passage completion task from 72\% to 34\%, that leaves much more room to improve. Our proposed tasks extend the previous single-variable passage completion task by replacing more character mentions with variables. Several deep learning models are developed to validate these three tasks. A thorough error analysis is provided to understand the challenges and guide the future direction of this research.

We show that consistency of Statistical EL knowledge bases, as defined by Penaloza and Potyka in SUM 2017 [4] is ExpTime-hard. Together with the existing ExpTime upper bound by Baader in FroCos 2017 [1], the result leads to the ExpTime-completeness of the mentioned logic. Our proof goes via a reduction from consistency of EL extended with an atomic negation, which is known to be equivalent to the well-known ExpTime-complete description logic ALC.

Google, Facebook and other internet giants would disclose the algorithms they use to return search results under new legislation proposed by US law makers. The bipartisan Filter Bubble Transparency Act also would require the online companies to offer users an unfiltered search option that delivers results without any algorithmic tinkering. Senator John Thune, a Republican from North Dakota, filed the bill on Friday. The legislation was co-sponsored by Republican senators Jerry Moran of Kansas and Marsha blackburn of Tennessee, as well as Democrats Richard Blumenthal of Connecticut and Mark Warner of Virginia. Senator John Thune, a Republican from North Dakota, filed the bipartisan'Filter Bubble Transparency Act,' which would require internet companies to reveal algorithms used to determine online searches The online firm, owned by Alphabet, like other internet companies relies on algorithms - a highly-specific set of instructions to computers - that track users' behavior and location Thune says the legislation is needed because'people are increasingly impatient with the lack of transparency,' on the internet, reports the Wall Street Journal.

It has been more than a decade in the making, but Waymo's self-driving taxis are officially picking up passengers without a human operator at the wheel. A group of early rider program members in Phoenix, Arizona received a message on this week offering them a free ride with the fully driverless service. Once the passenger is seated and the ride is underway, the car dials Waymo support to address any questions or concerns about the driverless ride – as many riders have never been carted around by a robot. It has been more than a decade in the making, but Waymo's self-driving taxis are officially picking up passengers without a human operator at the wheel. This car is all yours, with no one up front,' the pop-up notification from the Waymo app reads.

This week, the Defense Innovation Board issued a series of recommendations to the Department of Defense on how artificial intelligence should be implemented in future military conflict. The Defense Innovation Board was first created in 2016 to establish a series of best practices on potential collaborations between the US military and Silicon Valley. There are sixteen current board members from a broad number of disciplines, including former Google CEO Eric Schmidt, Facebook executive Marne Levine, Microsoft's Chief Digital Officer Kurt Delbene, astrophysicist Neil deGrasse Tyson, Steve Jobs biographer Walter Isaacson, and LinkedIn co-founded Reid Hoffman. 'Now is the time, at this early stage of the resurgence of interest in AI, to hold serious discussions about norms of AI development and use in a military context--long before there has been an incident.' the report says. The report says that using AI for military actions or decision-making comes with'the duty to take feasible precautions to reduce the risk of harm to the civilian population and other protected persons and objects.'

Personal Note: I'm really honored to share this conversation. I really hope you enjoy listening to it as much as I enjoyed talking to Dr. Marc Lanctot. In this interview, they talk all about Research at DeepMind, Deep Learning Research, AlphaGo. They also talk all about Swift For Tensorflow and OpenSpiel. Dr. Marc Lanctot is a research scientist at Google DeepMind.

The graph represents a network of 3,855 Twitter users whose tweets in the requested range contained "#iot OR "internet of things"", or who were replied to or mentioned in those tweets. The network was obtained from the NodeXL Graph Server on Friday, 01 November 2019 at 21:34 UTC. The requested start date was Friday, 01 November 2019 at 00:01 UTC and the maximum number of days (going backward) was 14. The maximum number of tweets collected was 5,000. The tweets in the network were tweeted over the 1-day, 3-hour, 59-minute period from Wednesday, 30 October 2019 at 20:01 UTC to Friday, 01 November 2019 at 00:01 UTC.