When people think of Artificial Intelligence (AI), typically they think of human-like thinking robots or futuristic scenes of self-aware computers taking over the world. But AI goes well beyond the automatons of science fiction into the real world of advanced computer science. In this session, we explore the new Amazon AI cloud-native deep learning technologies to add the power of artificial intelligence to your apps for; natural language understanding (NLU), automatic speech recognition (ASR), visual search & Image recognition, and text-to-speech (TTS). Amazon Lex enables building conversational interfaces into any app using voice and text that is powered by the same technology as Alexa used by Amazon Echo. Making it easy to add visual search and image classification to your applications by object, scenes, and faces detection, comparison, and search.

Recently, deep learning has been successfully applied to the problem of hashing, yielding remarkable performance compared to traditional methods with hand-crafted features. However, most of existing deep hashing methods are designed for the supervised scenario and require a large number of labeled data. In this paper, we propose a novel semi-supervised hashing method for image retrieval, named Deep Hashing with a Bipartite Graph (DHBG), to simultaneously learn embeddings, features and hash codes. More specifically, we construct a bipartite graph to discover the underlying structure of data, based on which an embedding is generated for each instance. Then, we feed raw pixels as well as embeddings to a deep neural network, and concatenate the resulting features to determine the hash code.

DeepSense is a deep learning framework that runs on mobile devices, and can be used for regression and classification tasks based on data coming from mobile sensors (e.g., motion sensors). An example of a classification task is heterogeneous human activity recognition (HHAR) – detecting which activity someone might be engaged in (walking, biking, standing, and so on) based on motion sensor measurements. Another example is biometric motion analysis where a user must be identified from their gait. An example of a regression task is tracking the location of a car using acceleration measurements to infer position. Compared to the state-of-art, DeepSense provides an estimator with far smaller tracking error on the car tracking problem, and outperforms state-of-the-art algorithms on the HHAR and biometric user identification tasks by a large margin.

Deep neural networks have many, many learnable parameters that are used to make inferences. Often, this poses a problem in two ways: Sometimes, the model does not make very accurate predictions. It also takes a long time to train them. The papers can be found here (Snapshot ensembles) and here (FreezeOut). This article assumes some familiarity with neural networks, including aspects like SGD, minima, optimisation, etc. Editor: this post describes Snapshot ensembles, and here is the second part which explains Freezout.

While it's easier than ever to define and train deep neural networks (DNNs), understanding the learning process remains somewhat opaque. Monitoring the loss or classification error during training won't always prevent your model from learning the wrong thing or learning a proxy for your intended classification task. Once upon a time, the US Army wanted to use neural networks to automatically detect camouflaged enemy tanks. Wisely, the researchers had originally taken 200 photos, 100 photos of tanks and 100 photos of trees. They had used only 50 of each for the training set.

Adoption of artificial intelligence in different fields is growing at a rapid pace. AI-based systems are going way beyond the usual expectations from machines, as they can rival, even better, human capabilities in certain areas. AI can now outwit and outperform humans in various comprehension and image-recognition tasks. Apart from a robot's ability to survive deadly environments like deep space, deep learning has been widely used to teach AI-based system fine motor skills for doing tasks such as removing a nail and placing caps on bottles. The IoT is imminent – and so are the security challenges it will inevitably bring.

Facebook announced today that it has started using neural network systems to carry out more than 4.5 billion translations that occur each day on the backend of the social network. Translations carried out with recurrent neural networks (RNNs) were able to scale with the use of Caffe2, a deep learning framework open-sourced by Facebook in April. The Caffe2 team today also announced that in part due to work done around translation, the framework is now able to work with recurrent neural networks. "Using Caffe2, we significantly improved the efficiency and quality of machine translation systems at Facebook. We got an efficiency boost of 2.5x, which allows us to deploy neural machine translation models into production," the Caffe2 team said in a blog post.

This post is a modified excerpt from one of my recent publications on machine learning. Everyone is curious and the jargon doesn't get off our backs. I attempted a comparison between the prevalent terminologies that exist today and how each of these are similar or dissimilar to machine learning. Please remember, I haven't yet brought in the Cognitive computing to the mix. My next post will cover more on cognition and how it is different from other areas of learning / intelligence.

Angela Zutavern is Vice President at Booz Allen Hamilton and Board of Directors ICE Foundation. Zutavern has pioneered the application of machine intelligence to organizational leadership and strategy. She is an inventor of the machine intelligence and data science strategies that are now helping business and government organizations make better decisions and gain competitive advantages. Zutavern led Booz Allen's most advanced data science research and development efforts, including the areas of deep learning and quantum machine learning. She has worked with clients in every major U.S. government cabinet-level department as well as in sub-level agencies.

Incorporating game mechanics into daily tasks has proven to be an effective way to motivate workers. As it turns out, gamification techniques don't just work on us. Google DeepMind is applying the tactic to machine learning. Looking to establish accountability across disparate project teams? Trying to automate processes or allow for lean methodology support?