The project came into inception in Q4 2017, and a Proof of Concept based around a JavaScript deep learning framework was developed in Q1 of 2018. Since then the team have continued to structure the framework and have also engaged in building both their client and developer communities. They are also currently engaged in building a public repository to encourage open source development with developers based around the world. They are also working on implementing basic calculus and statistical functions that will allow their platform to support more advanced and complex calculations. As a result, Raven Protocol is still a work in progress with their private beta launch set for Q1 of 2020, and being followed up with the public beta launch in Q2 of 2020.

Artificial Intelligence (AI) is working its way into almost every industry you can think of – including video games, healthcare, autonomous vehicles, cybersecurity, retail, and banking. With the growth of AI came the introduction of other terms such as "Machine Learning" and "Deep Learning". You may have heard them before but knowing how they relate to AI and how they are different can be confusing. Thanks to astonishing advancements in artificial intelligence (AI) and its sub-segments machine learning and deep learning, companies are achieving new levels of efficiency in data analysis that impact their entire business. The term, "artificial intelligence" was first created in 1956, but has become more popular today.

New Cleveland Clinic-led research shows that artificial intelligence (AI) can use medical scans and health records to personalize the dose of radiation therapy used to treat cancer patients. Published today in The Lancet Digital Health, the research team developed an AI framework based on patient computerized tomography (CT) scans and electronic health records. This new AI framework is the first to use medical scans to inform radiation dosage, moving the field forward from using generic dose prescriptions to more individualized treatments. Currently, radiation therapy is delivered uniformly. The dose delivered does not reflect differences in individual tumor characteristics or patient-specific factors that may affect treatment success.

An AI "designer" won big at this year's China International Fashion Design Innovation Competition (中国国际服装设计创新大赛), beating out many of its human counterparts before a panel of 50 judges, and causing some to speculate about the impact of AI on creativity. DeepVogue, an AI design system created by DeepBlue Technology (深兰科技), was the only "non-human" participant among 16 teams from around the country. The system went on to win the runner-up prize overall, as well as its "People's Choice Award." According to DeepBlue representatives, the technology requires a great deal of input from human designers, who can import images, themes and keywords into the DeepVogue system. The system then uses "deep learning"– essentially extensive studying by a machine of a database of information -- to produce original designs.

We identify the 6 tools in the modern open-source Data Science ecosystem, examine the Python vs R question, and determine which tools are used the most with Deep Learning and Big Data.

As the co-founder of a machine learning startup, many people have asked me to explain the difference between Machine Learning, Artificial intelligence and Deep Learning. Usually, I provide a reasonably concise explanation of the three areas. I then point them to this great article by Michael Copeland on the Nvidia Blog. However, that article is probably a bit more in-depth than people are expecting; most people want at maximum a 1–2-minute answer. So, I thought it was about time I have a go at writing a sub-2-minute article on the topic.

Interaction function (IFC), which captures interactions among items and users, is of great importance in collaborative filtering (CF). The inner product is the most popular IFC due to its success in low-rank matrix factorization. However, interactions in real-world applications can be highly complex. Many other operations (such as plus and concatenation) have also been proposed, and can possibly offer better performance than the inner product. In this paper, motivated by the success of automated machine learning, we propose to search for proper interaction functions (SIF) for CF tasks. We first design an expressive search space for SIF by reviewing and generalizing existing CF approaches. We then propose to represent the search space as a structured multi-layer perceptron, and design a stochastic gradient descent algorithm which can simultaneously update both architectures and learning parameters. Experimental results demonstrate that the proposed method can be much more efficient than popular AutoML approaches, and also obtain much better prediction performance than state-of-the-art CF approaches.

Hierarchical probabilistic models are able to use a large number of parameters to create a model with a high representation power. However, it is well known that increasing the number of parameters also increases the complexity of the model which leads to a bias-variance trade-off. Although it is a classical problem, the bias-variance trade-off between hidden layers and higher-order interactions have not been well studied. In our study, we propose an efficient inference algorithm for the log-linear formulation of the higher-order Boltzmann machine using a combination of Gibbs sampling and annealed importance sampling. We then perform a bias-variance decomposition to study the differences in hidden layers and higher-order interactions. Our results have shown that using hidden layers and higher-order interactions have a comparable error with a similar order of magnitude and using higher-order interactions produce less variance for smaller sample size.

Machine learning (ML) models trained by differentially private stochastic gradient descent (DP-SGD) has much lower utility than the non-private ones. To mitigate this degradation, we propose a DP Laplacian smoothing SGD (DP-LSSGD) for privacy-preserving ML. At the core of DP-LSSGD is the Laplace smoothing operator, which smooths out the Gaussian noise vector used in the Gaussian mechanism. Under the same amount of noise used in the Gaussian mechanism, DP-LSSGD attains the same differential privacy guarantee, but a strictly better utility guarantee, excluding an intrinsic term which is usually dominated by the other terms, for convex optimization than DP-SGD by a factor which is much less than one. In practice, DP-LSSGD makes training both convex and nonconvex ML models more efficient and enables the trained models to generalize better. For ResNet20, under the same strong differential privacy guarantee, DP-LSSGD can lift the testing accuracy of the trained private model by more than $8$\% compared with DP-SGD. The proposed algorithm is simple to implement and the extra computational complexity and memory overhead compared with DP-SGD are negligible. DP-LSSGD is applicable to train a large variety of ML models, including deep neural nets. The code is available at \url{https://github.com/BaoWangMath/DP-LSSGD}.

Corner cases are the main bottlenecks when applying Artificial Intelligence (AI) systems to safety-critical applications. An AI system should be intelligent enough to detect such situations so that system developers can prepare for subsequent planning. In this paper, we propose semi-supervised anomaly detection considering the imbalance of normal situations. In particular, driving data consists of multiple positive/normal situations (e.g., right turn, going straight), some of which (e.g., U-turn) could be as rare as anomalous situations. Existing machine learning based anomaly detection approaches do not fare sufficiently well when applied to such imbalanced data. In this paper, we present a novel multi-task learning based approach that leverages domain-knowledge (maneuver labels) for anomaly detection in driving data. We evaluate the proposed approach both quantitatively and qualitatively on 150 hours of real-world driving data and show improved performance over baseline approaches.