Computer boffins have devised a potential hardware-based Trojan attack on neural network models that could be used to alter system output without detection. Adversarial attacks on neural networks and related deep learning systems have received considerable attention in recent years due to the growing use of AI-oriented systems. The researchers – doctoral student Joseph Clements and assistant professor of electrical and computer engineering Yingjie Lao at Clemson University in the US – say that they've come up with a novel threat model by which an attacker could maliciously modify hardware in the supply chain to interfere with the output of machine learning models run on the device. Attacks that focus on the supply chain appear to be fairly uncommon. In 2014, reports surfaced that the US National Security Agency participates in supply chain interdiction to intercept hardware in transit and insert backdoors.

We're using a pre-trained model in Apache MXNet Gluon Python 3 code to classify a webcam image captured and processed with OpenCV. In our Python script, we capture the image to disk and capture JSON metadata about the percentage, probabilities, and device information. This JSON data is then sent via MQTT to a broker. Apache NiFi processes the JSON data. A side effect of the process is that it produces a SQL DDL to create a new table for this schema.

An image is worth a thousand words says the old wisdom quote that captures the importance of visual analysis in the learning process of human species. Every time we are presented with a visual scene, our brains make thousands of inferences about the objects in it and their contextual nature. For instance, if we see a person sitting down, we will infer that there is a chair underneath him. Visual inferences works even when we can't see the object. For instance, if we see a closet in a bedroom, we would assume there are clothing items inside even when we can't see them.

AI knows when you're going to die. But unlike in sci-fi movies, that information could end up saving lives. A new paper published in Nature suggests that feeding electronic health record data to a deep learning model could substantially improve the accuracy of projected outcomes. In trials using data from two US hospitals, researchers were able to show that these algorithms could predict a patient's length of stay and time of discharge, but also the time of death. The neural network described in the study uses an immense amount of data, such as a patient's vitals and medical history, to make its predictions.

Imagine if your car could pull itself over when you're drowsy or nauseous, or adjust the temperature and music when gridlock is stressing you out. Maybe it could even refuse to start if it knows you're intoxicated. With advanced ADAS systems already in place and the days of autonomous vehicles on the horizon, a lot of work is being done around sensing and machine learning to help vehicles better understand the roads and the world around them. But Boston-based startup Affectiva thinks more needs to be done around the internal world of the car--specifically the emotional state of the driver. Affectiva has built its business model around creating "emotional AI," algorithms capable of recognizing human emotional states.

Microsoft has become the latest company dragged into the tech industry's ethical reckoning over the use of its products by government agencies. On Sunday, critics noted a blog post from January in which Microsoft touted its work with US Immigration and Customs Enforcement (ICE). The post celebrated a government certification that allowed Microsoft Azure, the company's cloud-computing platform, to handle sensitive unclassified information for ICE. The sales-driven blog post outlined ways that ICE might use Azure Government, including enabling ICE employees to "utilize deep learning capabilities to accelerate facial recognition and identification," Tom Keane, a general manager at Microsoft wrote. "The agency is currently implementing transformative technologies for homeland security and public safety, and we're proud to support this work with our mission-critical cloud," the post added.

Artificial Intelligence and Deep Learning is the next thing in HFT https://lnkd.in/eycwTgQ Key TakeawaysStreamdata.io was originally built for banks and brokers, but more recently hedge funds have begun using the service.Whilst Hedge Funds like Renaissance Technologies have been using mathematical appr

The restricted Boltzmann machine is a network of stochastic units with undirected interactions between pairs of visible and hidden units. This model was popularized as a building block of deep learning architectures and has continued to play an important role in applied and theoretical machine learning. Restricted Boltzmann machines carry a rich structure, with connections to geometry, applied algebra, probability, statistics, machine learning, and other areas. The analysis of these models is attractive in its own right and also as a platform to combine and generalize mathematical tools for graphical models with hidden variables. This article gives an introduction to the mathematical analysis of restricted Boltzmann machines, reviews recent results on the geometry of the sets of probability distributions representable by these models, and suggests a few directions for further investigation.

Many Machine Reading and Natural Language Understanding tasks require reading supporting text in order to answer questions. For example, in Question Answering, the supporting text can be newswire or Wikipedia articles; in Natural Language Inference, premises can be seen as the supporting text and hypotheses as questions. Providing a set of useful primitives operating in a single framework of related tasks would allow for expressive modelling, and easier model comparison and replication. To that end, we present Jack the Reader (Jack), a framework for Machine Reading that allows for quick model prototyping by component reuse, evaluation of new models on existing datasets as well as integrating new datasets and applying them on a growing set of implemented baseline models. Jack is currently supporting (but not limited to) three tasks: Question Answering, Natural Language Inference, and Link Prediction. It is developed with the aim of increasing research efficiency and code reuse.

With the recent success of embeddings in natural language processing, research has been conducted into applying similar methods to code analysis. Most works attempt to process the code directly or use a syntactic tree representation, treating it like sentences written in a natural language. However, none of the existing methods are sufficient to comprehend program semantics robustly, due to structural features such as function calls, branching, and interchangeable order of statements. In this paper, we propose a novel processing technique to learn code semantics, and apply it to a variety of program analysis tasks. In particular, we stipulate that a robust distributional hypothesis of code applies to both human- and machine-generated programs. Following this hypothesis, we define an embedding space, inst2vec, based on an Intermediate Representation (IR) of the code that is independent of the source programming language. We provide a novel definition of contextual flow for this IR, leveraging both the underlying data- and control-flow of the program. We then analyze the embeddings qualitatively using analogies and clustering, and evaluate the learned representation on three different high-level tasks. We show that with a single RNN architecture and pre-trained fixed embeddings, inst2vec outperforms specialized approaches for performance prediction (compute device mapping, optimal thread coarsening); and algorithm classification from raw code (104 classes), where we set a new state-of-the-art.