We present a novel approach for training deep neural networks in a Bayesian way. Classical, i.e. non-Bayesian, deep learning has two major drawbacks both originating from the fact that network parameters are considered to be deterministic. First, model uncertainty cannot be measured thus limiting the use of deep learning in many fields of application and second, training of deep neural networks is often hampered by overfitting. The proposed approach uses variational inference to approximate the intractable a posteriori distribution on basis of a normal prior. The variational density is designed in such a way that the a posteriori uncertainty of the network parameters is represented per network layer and depending on the estimated parameter expectation values. This way, only a few additional parameters need to be optimized compared to a non-Bayesian network. We apply this Bayesian approach to train and test the LeNet architecture on the MNIST dataset. Compared to classical deep learning, the test error is reduced by 15%. In addition, the trained model contains information about the parameter uncertainty in each layer. We show that this information can be used to calculate credible intervals for the prediction and to optimize the network architecture for a given training data set.

The recent advent of automated neural network architecture search led to several methods that outperform state-of-the-art human-designed architectures. However, these approaches are computationally expensive, in extreme cases consuming GPU years. We propose two novel methods which aim to expedite this optimization problem by transferring knowledge acquired from previous tasks to new ones. First, we propose a novel neural architecture selection method which employs this knowledge to identify strong and weak characteristics of neural architectures across datasets. Thus, these characteristics do not need to be rediscovered in every search, a strong weakness of current state-of-the-art searches. Second, we propose a method for learning curve extrapolation to determine if a training process can be terminated early. In contrast to existing work, we propose to learn from learning curves of architectures trained on other datasets to improve the prediction accuracy for novel datasets. On five different image classification benchmarks, we empirically demonstrate that both of our orthogonal contributions independently lead to an acceleration, without any significant loss in accuracy.

Explainability and effectiveness are two key aspects for building recommender systems. Prior efforts mostly focus on incorporating side information to achieve better recommendation performance. However, these methods have some weaknesses: (1) prediction of neural network-based embedding methods are hard to explain and debug; (2) symbolic, graph-based approaches (e.g., meta path-based models) require manual efforts and domain knowledge to define patterns and rules, and ignore the item association types (e.g. substitutable and complementary). In this paper, we propose a novel joint learning framework to integrate \textit{induction of explainable rules from knowledge graph} with \textit{construction of a rule-guided neural recommendation model}. The framework encourages two modules to complement each other in generating effective and explainable recommendation: 1) inductive rules, mined from item-centric knowledge graphs, summarize common multi-hop relational patterns for inferring different item associations and provide human-readable explanation for model prediction; 2) recommendation module can be augmented by induced rules and thus have better generalization ability dealing with the cold-start issue. Extensive experiments\footnote{Code and data can be found at: \url{https://github.com/THUIR/RuleRec}} show that our proposed method has achieved significant improvements in item recommendation over baselines on real-world datasets. Our model demonstrates robust performance over "noisy" item knowledge graphs, generated by linking item names to related entities.

Illegal wildlife poaching threatens ecosystems and drives endangered species toward extinction. However, efforts for wildlife monitoring and protection in conservation areas are constrained by the limited resources of law enforcement agencies. To aid in wildlife protection, PAWS is an ML pipeline that has been developed as an end-to-end, data-driven approach to combat illegal poaching. PAWS assists park managers by identifying areas at high risk of poaching throughout protected areas based on real-world data and generating optimal patrol routes for deployment in the field. In this paper, we address significant challenges including extreme class imbalance (up to 1:200), bias, and uncertainty in wildlife poaching data to enhance PAWS and apply its methodology to several national parks with diverse characteristics. (i) We use Gaussian processes to quantify predictive uncertainty, which we exploit to increase the robustness of our prescribed patrols. We evaluate our approach on real-world historic poaching data from Murchison Falls and Queen Elizabeth National Parks in Uganda and, for the first time, Srepok Wildlife Sanctuary in Cambodia. (ii) We present the results of large-scale field tests conducted in Murchison Falls and Srepok Wildlife Sanctuary which confirm that the predictive power of PAWS extends promisingly to multiple parks. This paper is part of an effort to expand PAWS to 600 parks around the world through integration with SMART conservation software.

In this paper, we investigate deep learning (DL)-enabled signal demodulation methods and establish the first open dataset of real modulated signals for wireless communication systems. Specifically, we propose a flexible communication prototype platform for measuring real modulation dataset. Then, based on the measured dataset, two DL-based demodulators, called deep belief network (DBN)-support vector machine (SVM) demodulator and adaptive boosting (AdaBoost) based demodulator, are proposed. The proposed DBN-SVM based demodulator exploits the advantages of both DBN and SVM, i.e., the advantage of DBN as a feature extractor and SVM as a feature classifier. In DBN-SVM based demodulator, the received signals are normalized before being fed to the DBN network. Furthermore, an AdaBoost based demodulator is developed, which employs the $k$-Nearest Neighbor (KNN) as a weak classifier to form a strong combined classifier. Finally, experimental results indicate that the proposed DBN-SVM based demodulator and AdaBoost based demodulator are superior to the single classification method using DBN, SVM, and maximum likelihood (MLD) based demodulator.

This paper studies a new application of deep learning (DL) for optimizing constellations in two-way relaying with physical-layer network coding (PNC), where deep neural network (DNN)-based modulation and demodulation are employed at each terminal and relay node. We train DNNs such that the cross entropy loss is directly minimized, and thus it maximizes the likelihood, rather than considering the Euclidean distance of the constellations. The proposed scheme can be extended to higher level constellations with slight modification of the DNN structure. Simulation results demonstrate a significant performance gain in terms of the achievable sum rate over conventional relaying schemes. Furthermore, since our DNN demodulator directly outputs bit-wise probabilities, it is straightforward to concatenate with soft-decision channel decoding.

Distant supervision for relation extraction heavily suffers from the wrong labeling problem. To alleviate this issue in news data with the timestamp, we take a new factor time into consideration and propose a novel time-aware distant supervision framework (Time-DS). Time-DS is composed of a time series instance-popularity and two strategies. Instance-popularity is to encode the strong relevance of time and true relation mention. Therefore, instance-popularity would be an effective clue to reduce the noises generated through distant supervision labeling. The two strategies, i.e., hard filter and curriculum learning are both ways to implement instance-popularity for better relation extraction in the manner of Time-DS. The curriculum learning is a more sophisticated and flexible way to exploit instance-popularity to eliminate the bad effects of noises, thus get better relation extraction performance. Experiments on our collected multi-source news corpus show that Time-DS achieves significant improvements for relation extraction.

Zone7 announced last month it had raised $2.5 million in seed funding from a Resolute Ventures-led group, which also included Amicus Capital Partners, PLG Ventures, UpWest Labs, Dave Pell, PLG Ventures and UpWest, as well as former and current athletes, including the National Basketball Association's Kristaps Porzingis. Zone7, a 10-person startup, said Major League Baseball, professional soccer teams and several college programs are using its system, although the company declined to name the teams. However, the Maccabi Tel Aviv Football Club, an Israeli pro team, has tested the technology, according to Jordi Cruyff, who was the team's sports director from 2012 to 2017. Mr. Cruyff is also a former player and one-time member of some of the biggest teams in the sport, including FC Barcelona and Manchester United Football Club. With Maccabi, Mr. Cruyff said he would receive daily alerts from Zone7 that certain players were at high risk.

Venture capital firm GFR Fund announced it has raised $20 million to launch its second fund, GFR Fund II, with a focus on technology startups disrupting the digital entertainment sector. The new fund will invest in virtual reality, augmented reality, esports, livestreaming, artificial intelligence, and blockchain. Multiple strategic investors from Asia are backing it, and it is an affiliate of leading mobile gaming company Gree, based in Japan. San Francisco-based GFR Fund hopes to distinguish itself with its deep connections in Silicon Valley and business opportunities in Asian markets. "We are looking beyond traditional game developers to invest in disruptive entertainment technologies," said Teppei Tsutsui, managing director of GFR Fund, in a statement.

Artificial intelligence (AI) is the 21st century space race where India lags far behind leaders like China and the US. However, there is one area where the country, with second largest number of poor, can lead the world. It can use AI to solve problems for the underserved billions. That's exactly what Wadhwani Institute of AI does. Launched last February by prime minister Narendra Modi, backed by NRI entrepreneurs (Rs 200 crore grant) - Wadhwani brothers Sunil and Romesh – WIAI is using AI to serve the bottom of the pyramid.