Voice Assistants like Alexa and Siri definitely have a place in the classroom. As the market for voice assistants continues to grow, more and more applications will be built for voice to supplement the classroom experience. Even in their current state, voice assistants can provide immense value to classrooms. Take a simple use case, such as a teacher setting a reminder to discuss higher-level lesson points the following week. Often times, teachers may not find the time or even remember to review difficult material, so using voice assistants to set reminders in real time can greatly enhance the classroom processes and therefore help students to learn more and continue to build their knowledge base.

Train and certify yourself as a Cogito and Cognitive Computing Expert. Join the community of professionals able to face today's challenge of dealing with large volumes of unstructured data. It's no longer simply a question of computing speed and power: foremost technologies must be intelligent and professionals must know how to use them successfully. Expert System's Certification will help you build a sense of trust with your customers because they'll know that you have the necessary skills to perform the job.

The evolution of AI has been a rich tale of exploration since its origins in the 1950's, with the last decade providing an especially dramatic chapter of breakthrough innovations. But I believe the real story is what comes next -- when the disruption stabilizes and machine learning transitions from a staple of Silicon Valley headlines to an everyday technology. It'll be a far longer chapter -- perhaps decades -- in which developers all over the world use a mature set of tools to transform their industries. In 2019, we find ourselves at the start of this new chapter. AI has undergone a remarkable refinement in recent years, as barriers to entry have fallen and a wide range of products, services, resources, and best practices have emerged. As our focus shifts -- finally -- from AI itself to the impact that AI can have on your business, the question is no longer how this technology works, but what it can do for you.

It won't be wrong to say that AI is one of the hottest topics right now, tech companies are working round the clock to get best AI for their work. Companies like Google, Apple, Baidu, etc. are investing billions of dollars in their AI programs. But before we dive into maths, algorithms and programming languages, let's have some background knowledge of AI. We'll try our best to make this article simple as possible. Well, when someone talks about AI most people think about movies like Chappie, Terminator, and Lucy, etc.

The rise of technology within the education sector over the last few decades has been astounding. This is certainly the case if we consider that teaching with technology has become pervasive in almost every classroom environment. Within today's classroom, for example, we find ourselves surrounded by devices such as smart boards, AV, computers, laptops, tablets and phones, to name but a few technologies which are now being integrated into teaching. We have also seen the rise of the virtual learning environment and blended learning, alongside a significant rise in online education. This has allowed distance learning to take new forms and shapes and to reach greater audiences around the world.

Recently, several studies have proven the global convergence and generalization abilities of the gradient descent method for two-layer ReLU networks by making a positivity assumption of the Gram-matrix of the neural tangent kernel. However, the performance of gradient descent on classification problems has not been well studied, and further investigation of the problem structure is possible. In this work, we present a partially stronger but reasonable assumption for binary classification problems compared to the positivity assumption of the Gram-matrix, where a data distribution can be perfectly classifiable by a tangent model, and we provide a refined generalization analysis of the gradient descent method for two-layer networks with smooth activations. A remarkable point of this study is that our generalization bound has much better dependence on the network width compared to existing results. As a result, our theory significantly enlarges a class of over-parameterized networks having provable generalization ability, with respect to network width, while most studies require much higher over-parameterization.

Multi-hop Reading Comprehension (RC) requires reasoning and aggregation across several paragraphs. We propose a system for multi-hop RC that decomposes a compositional question into simpler sub-questions that can be answered by off-the-shelf single-hop RC models. Since annotations for such decomposition are expensive, we recast sub-question generation as a span prediction problem and show that our method, trained using only 400 labeled examples, generates sub-questions that are as effective as human-authored sub-questions. We also introduce a new global rescoring approach that considers each decomposition (i.e. the sub-questions and their answers) to select the best final answer, greatly improving overall performance. Our experiments on HotpotQA show that this approach achieves the state-of-the-art results, while providing explainable evidence for its decision making in the form of sub-questions.

Multi-hop reading comprehension (RC) questions are challenging because they require reading and reasoning over multiple paragraphs. We argue that it can be difficult to construct large multi-hop RC datasets. For example, even highly compositional questions can be answered with a single hop if they target specific entity types, or the facts needed to answer them are redundant. Our analysis is centered on HotpotQA, where we show that single-hop reasoning can solve much more of the dataset than previously thought. We introduce a single-hop BERT-based RC model that achieves 67 F1---comparable to state-of-the-art multi-hop models. We also design an evaluation setting where humans are not shown all of the necessary paragraphs for the intended multi-hop reasoning but can still answer over 80% of questions. Together with detailed error analysis, these results suggest there should be an increasing focus on the role of evidence in multi-hop reasoning and possibly even a shift towards information retrieval style evaluations with large and diverse evidence collections.

In the last two decades, automatic extractive text summarization on lectures has demonstrated to be a useful tool for collecting key phrases and sentences that best represent the content. However, many current approaches utilize dated approaches, producing sub-par outputs or requiring several hours of manual tuning to produce meaningful results. Recently, new machine learning architectures have provided mechanisms for extractive summarization through the clustering of output embeddings from deep learning models. This paper reports on the project called Lecture Summarization Service, a python based RESTful service that utilizes the BERT model for text embeddings and KMeans clustering to identify sentences closes to the centroid for summary selection. The purpose of the service was to provide students a utility that could summarize lecture content, based on their desired number of sentences. On top of the summary work, the service also includes lecture and summary management, storing content on the cloud which can be used for collaboration. While the results of utilizing BERT for extractive summarization were promising, there were still areas where the model struggled, providing feature research opportunities for further improvement.

Neural Networks (NNs) have been found vulnerable to a class of imperceptible attacks, called adversarial examples, which arbitrarily alter the output of the network. These attacks have called the validity of NNs into question, particularly on sensitive problems such as medical imaging or fraud detection. We further argue that the fields of explainable AI and Human-In-The-Loop (HITL) algorithms are impacted by adversarial attacks, as attacks result in perturbations outside of the salient regions highlighted by state-of-the-art techniques such as LIME or Grad-CAM. This work accomplishes three things which greatly reduce the impact of adversarial examples, and pave the way for future HITL workflows: we propose a novel regularization technique inspired by the Lipschitz constraint which greatly improves an NN's resistance to adversarial examples; we propose a collection of novel network and training changes to complement the proposed regularization technique, including a Half-Huber activation function and an integrator-based controller for regularization strength; and we demonstrate that networks trained with this technique may be deliberately attacked to generate rich explanations. Our techniques led to networks more robust than the previous state of the art: using the Accuracy-Robustness Area (ARA), our most robust ImageNet classification network scored 42.2% top-1 accuracy on unmodified images and demonstrated an attack ARA of 0.0053, an ARA 2.4x greater than the previous state-of-the-art at the same level of accuracy on clean data, achieved with a network one-third the size. A far-reaching benefit of this technique is its ability to intuitively demonstrate decision boundaries to a human observer, allowing for improved debugging of NN decisions, and providing a means for improving the underlying model.